JP2012515407A - Retrieve and display information from unstructured electronic document collections - Google Patents

Abstract

  A method, system, and apparatus is implemented that includes a computer program stored on a computer storage medium for retrieving and displaying information from an unstructured electronic document collection. One aspect may be embodied by a machine-implemented method that includes an action to receive a machine-readable search query from a user and an action to respond to the search query with instructions for presenting the user with a structured presentation of instances associated with the search query. The visual presentation of a structured presentation represents the association between an instance and the values that characterize the attributes of the instance by an array of instance identifiers and values.

Description

  The present specification relates to a method for extracting and displaying information from an unstructured electronic document collection.

  An electronic document is a collection of machine-readable data. Electronic documents are typically individual files and are formatted according to a predefined format (eg, PDF, TIFF, HTML, MS Word, PCL, PostScript, or the like). Electronic documents can be stored electronically and distributed electronically. In some cases, the electronic document includes audio content, visual content, and other information, as well as links to text and other electronic documents.

  Electronic documents can be collected in an electronic document collection. Electronic document collections can be either unstructured or structured. The format of documents included in an unstructured electronic document collection is not constrained to conform to a predetermined structure and can evolve in an unpredictable manner in many cases. In other words, the format of individual documents within an unstructured electronic document collection is neither restrictive nor persistent throughout the document collection. Furthermore, in an unstructured electronic document collection, there is no mechanism for ensuring that a new document follows a certain format, or that changes to the format are applied to an existing document. Therefore, documents in an unstructured electronic document collection cannot be expected to share a common structure that can be used when extracting information. Examples of unstructured electronic document collections include documents available on the Internet, resume collections, magazine article collections, and news article collections. Documents in some unstructured electronic document collections are prohibited from including links to other documents inside the collection and other documents outside the collection.

  In contrast, documents in a structured electronic document collection generally follow a format that can be restrictive and permanent. The format applied to documents in structured electronic document collections is restrictive in that a common format applies to all documents in those collections, even if the applied format is not entirely suitable. Good. These formats may be permanent in that a significant involvement in a particular format by the party building the structured electronic document collection is generally required. In addition, users of these collections (especially programs that use the documents in the collection) rely on documents having the expected format. As a result, format changes can be difficult to implement. Structured electronic document collections are ideal for applications that help categorize information content simply and stably. Thus, the documents in a structured electronic document collection generally share a common structure that can be used in extracting information. An example of a structured electronic document collection, organized through a database management system (DBMS) according to a hierarchical data model and a relational data model, created by a single entity to present and consistently present a database to be viewed A collection of electronic documents. For example, a collection of web pages provided by an online bookstore to present information about individual books can form a structured electronic document collection. As another example, a collection of web pages created by server-side script and viewed through an application server can form a structured electronic document collection. Thus, each of the one or more structured electronic document collections may be a subset of the unstructured electronic document collection.

  In this specification, a technique related to the stage of extracting and displaying information from an unstructured electronic document collection, for example, an electronic document available on the Internet will be described. The electronic document collection may be an unstructured collection, but the information content of the unstructured electronic document collection can be displayed in a structured presentation. In particular, the information content of an unstructured electronic document collection can be used not only to determine attribute values, but also to identify, select, and name attributes and instances in structured presentations. Using such structured presentations, information can be presented to the user in a consistent manner despite the diversity of information sources. Examples of structured presentations include tables and other collections of records.

  In general, one aspect of the subject matter described herein is to respond to a search query with an action to receive a machine-readable search query from a user and instructions for presenting the user with a structured presentation of instances associated with the search query. And an action to be implemented. The visual presentation of a structured presentation represents the association between an instance and the values that characterize the attributes of the instance by an array of instance identifiers and values. Instance identifiers and values are derived from two or more documents in an unstructured collection of electronic documents. An electronic document collection is an unstructured collection in that the format of electronic documents within the electronic document collection is neither restrictive nor permanent.

  This and other aspects can include one or more of the following features. Responding to the search query includes identifying a first collection of electronic documents in an unstructured collection related to the instance; extracting an instance attribute value from the first collection of electronic documents; Writing values extracted from two or more electronic documents to a structured presentation. Responsive to the search query includes extracting a first value of the first attribute of the first instance from the first electronic document and second attribute of the first instance from the second electronic document. Extracting a second value and associating the first value and the second value with the first instance as a single in the structured presentation may be included. The first attribute may be different from the second attribute, and the first electronic document may be different from the second electronic document. In response to the search query, the first value of the first instance attribute is extracted from the first electronic document, and the second value of the second instance attribute is extracted from the first electronic document. And associating the first value with the first instance in the first record and associating the second value with the second instance in the second record. The first instance may be different from the second instance. A structured presentation includes a table, and a record can include a row or column of the table. A structured presentation includes a collection of cards, and a record may be an individual card within that collection. The method can also include receiving a trigger to add a new instance to the structured presentation and proposing a new instance to add to the structured presentation in response to the trigger. The method can also include receiving a constraint specification from the user, and proposing a new instance includes proposing a new instance that satisfies the user-specified constraint condition. The method can include receiving a trigger to add a new attribute to the structured presentation and adding a new attribute to the structured presentation in response to the trigger. The method may also include receiving a user designation of the new attribute characteristic and writing an attribute value to the structured presentation based on the user-specified characteristic. An unstructured electronic document collection can include electronic documents available on the Internet. A structured presentation can be physically presented on the display screen, including physically transforming one or more elements of the display screen.

  Other embodiments of this aspect include corresponding systems, apparatus, and computer programs recorded on computer storage devices that are each configured to perform the operations of these methods.

  Other aspects of the subject matter described herein comprise one or more machine-readable data storage media that store instructions operable to cause one or more data processors to perform operations. It can be embodied by a device. These operations include receiving descriptive data describing a pre-existing structured presentation, retrieving a first instance identifier from a first website, and a first from a second website. Deriving the first value of the first attribute of the instance and adding the first instance identifier and new value to the pre-existing structured presentation to form a new record in the new structured presentation And outputting instructions for visually presenting the new structured presentation. Visual presentation of pre-existing structured presentation visually presents information in a system layout that matches the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  This and other aspects can include one or more of the following features. Deriving the identifier of the first instance from the first website may include comparing pre-existing structured presentation characteristics with pre-existing structured presentation content. These operations may also include receiving a second instance identifier from the user. The new structured presentation can include a second new record that presents the second instance in relation to the second value of the first attribute of the second instance. These operations can include receiving a second value from the user. A collection of candidate values can be presented to the user and a second value selection can be received from the user. The collection of candidate values includes the second value. A collection of candidate values for the first attribute of the second instance can be identified, and for each candidate value, the certainty that the candidate value is correct can be determined. These operations can include proposing a collection of new instances to be added to the structured presentation. A collection of new instances is created by comparing pre-existing structured presentation characteristics with the content of the first website and the second website, and / or machine-readable search queries on the first website. And can be proposed by comparing with the content of the second website. Deriving the first value from the second website may include identifying that the second website includes reviews, extracting an identifier directly from the first website, or first web Extracting the identifier from a machine readable database containing information extracted from the site may be included. A pre-existing structured presentation includes a table, and a record can include a row or column of that table. A pre-existing structured presentation includes a collection of cards, and a record may be an individual card within that collection. These operations can include visually displaying the new structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Other embodiments of this aspect include corresponding systems, devices, and methods.

  In other aspects, the system comprises a client device and one or more computers programmed to interact with the client device and perform operations. These operations include receiving descriptive data describing a pre-existing structured presentation, retrieving a first instance identifier from a first website, and a first from a second website. Deriving the first value of the first attribute of the instance and adding the first instance identifier and new value to the pre-existing structured presentation to form a new record in the new structured presentation And outputting to the client device instructions for visually presenting the new structured presentation. Visual presentation of pre-existing structured presentation visually presents information in a system layout that matches the structured design. A structured presentation includes a collection of records, each of which represents an association between an instance and a value characterizing the attributes of the instance by an instance identifier and an array of values in the visual presentation of the structured presentation.

  This and other aspects can include one or more of the following features. The one or more computers may include a server operable to interact with the client device through the data communication network, and the client device is operable to interact with the server as a client. .

  Other embodiments of this aspect include corresponding systems, devices, and methods.

  In other aspects, the system comprises a client device and one or more computers programmed to interact with the client device and perform operations. These operations include receiving a machine readable search query from a client device and responding to the search query by sending instructions to the client device to present a structured presentation of instances associated with the search query. . The visual presentation of a structured presentation represents the association between an instance and the values that characterize the attributes of the instance by an array of instance identifiers and values. Instance identifiers and values are derived from two or more documents in an unstructured collection of electronic documents. An electronic document collection is an unstructured collection in that the format of electronic documents within the electronic document collection is neither restrictive nor permanent.

  This and other aspects can include one or more of the following features. The one or more computers may include a server operable to interact with the client device through the data communication network, and the client device is operable to interact with the server as a client. .

  Other embodiments of this aspect include corresponding systems, devices, and methods.

  Another aspect of the subject matter described herein is the action of receiving descriptive data describing a pre-existing structured presentation and the characteristics of the pre-existing structured presentation. An action that identifies an electronic document that identifies a new instance associated with a pre-existing structured presentation compared to the content of the electronic document in the structured collection, and a structure that pre-exists the identifier of the new instance It can be embodied by a machine-implemented method including an action for forming an extended structured presentation in addition to a structured presentation and an action for outputting a command for presenting the extended structured presentation. Visual presentation of pre-existing structured presentation visually presents information in a system layout that matches the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  This and other aspects can include one or more of the following features. Adding a new instance identifier includes forming a collection of instance suggestions, providing the instance proposal collection to the user, and receiving a user selection of the new instance. However, the new instance is included in the collection of instance proposals. Comparing the characteristics of a pre-existing structured presentation with the content of the electronic document includes documents in the electronic document collection that contain structured components related to the instances identified in the pre-existing structured presentation. Can be included. Configuring a collection of instance proposals includes identifying an instance document identified in a pre-existing structured presentation and identifying a first document in an electronic document collection arranged according to a template; Identifying a second document arranged according to the template but related to the second instance and including the second instance in the instance proposal collection may be included. Comparing the characteristics of a pre-existing structured presentation with the content of an electronic document is the process of comparing documents in an electronic document collection that contain information about one or more instances in the pre-existing structured presentation. Characterized in the pre-existing structured presentation, the identifying stage, identifying new instances in the stored collection of instance-attribute associations, pre-existing structured presentation characteristics Comparing with existing attributes, comparing attributes used to characterize instances in a pre-existing structured presentation with the content of the electronic document, instances in pre-existing structured presentations Used to characterize Comparing one or more of comparing the value of the attribute with the content of the electronic document and comparing the category of instances, including instances in a pre-existing structured presentation, with the content of the electronic document be able to. The collection of electronic documents can include electronic documents available on the Internet. The electronic document can include a web page. An extended structured presentation can include a table or a collection of cards. The method can include visually displaying the expanded structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Other embodiments of this aspect include corresponding systems, apparatus, and computer programs recorded on computer storage devices that are each configured to perform the operations of these methods.

  Other aspects of the subject matter described herein comprise one or more machine-readable data storage media that store instructions operable to cause one or more data processors to perform operations. It can be embodied by a device. These operations consist of constructing a collection of instance proposals based on the content of two or more documents in an unstructured electronic document collection, providing the instance proposal collection to the user, and collecting the instance proposal collection. Receiving a user selection of the first instance of the first instance and adding an identifier of the first instance proposal to the structured presentation. An electronic document collection is an unstructured collection in that the format of electronic documents within the electronic document collection is neither restrictive nor permanent. Visual presentation of structured presentations presents information visually in an organized manner. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  This and other aspects can include one or more of the following features. The stage of constructing a collection of instance proposals is identified in the preexisting structured presentation, comparing the characteristics of the preexisting structured presentation with the content of the electronic documents in the electronic document collection. Identifying a document in an electronic document collection that contains structured components related to the existing instance, in the electronic document collection related to the instance identified in the pre-existing structured presentation and arranged according to the template Identifying the first document of the first, identifying the second document arranged according to the template but related to the second instance, including the second instance in the instance proposal collection, in advance Existing structured presentation Identifying documents in an electronic document collection that include identifiers of one or more instances in, identifying additional attributes used to characterize instances in pre-existing structured presentations; Comparing the value of an attribute used to characterize an instance in a pre-existing structured presentation with the value of the instance proposal, and the category of the instance containing the instance in the pre-existing structured presentation Identifying, using the instances in that category of instances to configure a collection of instance proposals, identifying instance proposals in the collection that stores the associations between instances and attributes, and pre-existing Structured presentations One or more of comparing the attributes characterized in the application to the content of the documents in the unstructured electronic document collection. A collection of electronic documents can include documents available on the Internet. The electronic document can include a web page. A structured presentation can include a table or a collection of cards.

  Other embodiments of this aspect include corresponding systems, devices, and methods.

  Other aspects of the inventive subject matter described herein can be embodied by a system comprising a client device and one or more computers programmed to interact with the client device and perform operations. These operations involve receiving descriptive data describing pre-existing structured presentations, and pre-existing structured presentation characteristics with the contents of electronic documents in an unstructured collection of electronic documents. Identifying an electronic document that identifies new instances associated with a pre-existing structured presentation by comparison, and adding an identifier of the new instance to the pre-existing structured presentation to extend the structured presentation And outputting instructions for presenting the extended structured presentation on a display device that is coupled in data communication with the client device. Visual presentation of pre-existing structured presentation visually presents information in a system layout that matches the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  Other embodiments of this aspect include corresponding computer program products, apparatus, and methods.

  Other aspects of the inventive subject matter described herein can be embodied by a system comprising a client device and one or more computers programmed to interact with the client device and perform operations. These operations consist of configuring a collection of instance proposals based on the content of two or more documents in an unstructured electronic document collection and providing the instance proposal collection to the user using a client device. Receiving a user selection of the first instance in the collection of instance proposals, and the first instance proposal in a structured presentation presented on a display device coupled to perform data communication with the client device. Adding the identifier, and the visual presentation of the structured presentation visually presents the information in an organized manner. An electronic document collection is an unstructured collection in that the format of electronic documents within the electronic document collection is neither restrictive nor permanent. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  This and other aspects can include one or more of the following features. The one or more computers can include a server operable to interact with client devices through a data communication network. The client device may be operable to interact with the server as a client. The client device may include a personal computer that runs a web browser. The personal computer can include a display device.

  Other embodiments of this aspect include corresponding computer program products, apparatus, and methods.

  Another aspect of the subject matter described herein is the action of receiving descriptive data describing a pre-existing structured presentation and the characteristics of the pre-existing structured presentation. An action that identifies an electronic document that identifies a new attribute associated with a pre-existing structured presentation compared to the content of the electronic document in the structured collection, and a structure that pre-exists with an identifier for the new attribute It can be embodied by a machine-implemented method including an action for forming an extended structured presentation in addition to a structured presentation and an action for outputting a command for presenting the extended structured presentation. Visual presentation of pre-existing structured presentation visually presents information in a system layout that matches the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  This and other aspects can include one or more of the following features. Adding a new attribute identifier may include configuring a collection of attribute suggestions, providing the user with an attribute suggestion collection, and receiving a user selection of the new attribute. New attributes may be in the instance proposal collection. Comparing the characteristics of a pre-existing structured presentation with the content of the electronic document includes documents in the electronic document collection that contain structured components related to the instances identified in the pre-existing structured presentation. Can be included. Configuring the attribute proposal collection involves identifying a first document in the electronic document collection that is associated with the instance identified in the pre-existing structured presentation and arranged according to the template; Adding an attribute used in the first document to characterize an instance in the proposal collection. Comparing the characteristics of a pre-existing structured presentation with the content of an electronic document is the process of comparing documents in an electronic document collection that contain information about one or more instances in the pre-existing structured presentation. Identifying, identifying a new attribute in a stored collection of instance-attribute associations, comparing an instance characterized in a pre-existing structured presentation with the content of the electronic document Identifying additional instances related to the instances identified in the pre-existing structured presentation, attributes or attributes used to characterize the instances in the pre-existing structured presentation The value of the electronic document One or more of comparing with the content, comparing a category of instances, including instances in a pre-existing structured presentation, with the content of the electronic document may be included. The collection of electronic documents can include electronic documents available on the Internet, and the electronic documents can include web pages. An extended structured presentation can include a table or a collection of cards. The method can include visually presenting the augmented structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Other embodiments of this aspect include corresponding systems, devices, and computer program products.

  Other aspects of the subject matter described herein comprise one or more machine-readable data storage media that store instructions operable to cause one or more data processors to perform operations. It can be embodied by a device. These operations consist of constructing a collection of attribute proposals based on the contents of two or more documents in an unstructured electronic document collection, providing the user with an attribute proposal collection, and collecting attribute proposals. Receiving a user selection of the first attribute of the first attribute and adding an identifier of the first attribute proposal to the structured presentation. An electronic document collection is an unstructured collection in that the format of electronic documents within the electronic document collection is neither restrictive nor permanent. Visual presentation of structured presentations presents information visually in an organized manner. A structured presentation represents the association between an instance and the values that characterize the attributes of the instance in the presentation of the structured presentation by an array of instance identifiers and values.

  This and other aspects can include one or more of the following features. The stage of constructing the collection of attribute proposals is identified in the pre-existing structured presentation, comparing the characteristics of the pre-existing structured presentation with the content of the electronic documents in the electronic document collection. Identifying a document in an electronic document collection that contains a structured component that characterizes the existing instance, associated with the instance identified in the pre-existing structured presentation, and arranged in a template An electronic that contains information about one or more instances in a pre-existing structured presentation, identifying the first document, including the attributes used to characterize the instances in the attribute proposal collection Statements in the document collection It may include one or more of identifying a. The step of comparing pre-existing structured presentation characteristics with the content of the electronic document is pre-existing, comparing the instance identified in the pre-existing structured presentation with the content of the electronic document. One or more of comparing an attribute or attribute value used to characterize an instance in a structured presentation to the content of the electronic document may be included. The stage of constructing a collection of attribute proposals is to identify an instance category that includes an instance in a pre-existing structured presentation, and attributes from the attributes used to characterize the instances in that category of instances. Composing a collection of proposals. A collection of attribute proposals can also be constructed by identifying attribute proposals in the collection in which the association of instances and attributes is stored. The collection of electronic documents can include electronic documents available on the Internet, and the electronic documents can include web pages. A structured presentation can include a table or a collection of cards. These operations may also include visually presenting the structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Other embodiments of this aspect include corresponding systems, devices, and computer program products.

  Another aspect of the subject matter described herein is by a system comprising a client device comprising a display screen and one or more computers programmed to interact with the client device and perform operations. Can be embodied. These operations involve receiving descriptive data describing pre-existing structured presentations, and pre-existing structured presentation characteristics with the contents of electronic documents in an unstructured collection of electronic documents. Identifying an electronic document that identifies new attributes associated with a pre-existing structured presentation by comparison, and adding a new attribute identifier to the pre-existing structured presentation to expand the structured presentation And outputting instructions for presenting the extended structured presentation on the display screen. Visual presentation of pre-existing structured presentation visually presents information in a system layout that matches the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  Other embodiments of this aspect include corresponding computer program products, apparatus, and methods.

  Another aspect of the subject matter described herein is by a system comprising a client device comprising a display screen and one or more computers programmed to interact with the client device and perform operations. Can be embodied. These operations consist of configuring a collection of attribute proposals based on the contents of two or more documents in the unstructured electronic document collection, providing the attribute proposal collection to the client device, and from the client device Receiving a selection of a first attribute in the collection of attribute proposals and adding an identifier of the first attribute proposal to the structured presentation presented on the display screen. An electronic document collection is an unstructured collection in that the format of electronic documents within the electronic document collection is neither restrictive nor permanent. Visual presentation of structured presentations presents information visually in an organized manner. A structured presentation represents the association between an instance and the values that characterize the attributes of the instance in the presentation of the structured presentation by an array of instance identifiers and values.

  Other embodiments of this aspect include corresponding computer program products, apparatus, and methods.

  Another aspect of the subject matter described herein is the action of receiving descriptive data describing a pre-existing structured presentation and the characteristics of the pre-existing structured presentation. An action that identifies an electronic document that identifies a new value associated with a pre-existing structured presentation compared to the content of the electronic document in the structured collection, and a structured presentation that pre-exists the new value In addition, the machine-implemented method may include an action for forming a new structured presentation and an action for outputting a command for visually presenting the new structured presentation. Visual presentation of pre-existing structured presentation visually presents information in a system layout that matches the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  This and other aspects can include one or more of the following features. The identifier of the first instance that appears in the structured presentation in the first electronic document can be identified and a new value can be extracted from the first electronic document. Adding a new value includes identifying a collection of values for the first attribute of the first instance, and substituting a subset of one or more of the identified values for the first instance. Determining that one attribute is appropriately characterized. Determining a subset of values as appropriate involves dividing the values in the collection into groups, selecting a subset based at least in part on the number of values in the subset, and providing user-specified constraints. Selecting a subset based at least in part on the values in the satisfying subset, selecting a subset based at least in part on the values in the subset being extracted from the high quality document, and pre-setting the values in the subset Selecting a subset based at least in part on steps derived from documents associated with other instances in an existing structured presentation; others in a structured presentation in which values in the subset already exist Based at least in part on the steps derived from the document associated with the attribute It may include one or more of the steps of selecting a subset. The collection of electronic documents may be the Internet, and the electronic documents may be web pages. The pre-existing structured presentation can include a table or a collection of cards. The method can include visually presenting the new structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Other embodiments of this aspect include corresponding systems, devices, and computer program products.

  Other aspects of the subject matter described herein comprise one or more machine-readable data storage media that store instructions operable to cause one or more data processors to perform operations. It can be embodied by a device. These operations receive descriptive data describing a first instance, a second instance, and a first attribute, and the first instance of two or more documents in the unstructured electronic document collection Extracting a first collection of first attribute values and extracting a second collection of first attribute values of a second instance from two or more documents of an unstructured electronic document collection Determining a first subset of the first collection of values as appropriately characterizing the first attribute of the first instance, and a second subset of the second collection of values, A structured presentation including determining a first attribute of the second instance as appropriately characterizing and including a first value of the first subset and a second value of the second subset Deployment may include the steps of generating a machine-readable instructions for displaying. A structured presentation represents the association between an instance and the values that characterize the attributes of the instance by an array of instance identifiers and values.

  This and other aspects can include one or more of the following features. The first subset of values can be determined as appropriate by dividing the values in the first collection into groups, but each group contains a subset of the first collection of values . The first subset of values may be determined as appropriate by selecting the first subset based at least in part on the number of values in the first subset. The first subset of values can be determined as appropriate by comparing the values in the first subset with user-specified constraints on those values. The first subset of values can be determined as appropriate by determining that the values in the first subset were derived from a high quality document. The first subset of values can be determined as appropriate by determining that the values in the first subset are derived from documents associated with the second instance. The first subset of values is determined as appropriate by determining that the values in the first subset are derived from documents associated with other attributes that characterize both the first instance and the second instance be able to. The description of the first instance may include an identifier of the first instance that appears in the pre-existing structured presentation. The description of the second instance can include an identifier of the second instance that appears in the pre-existing structured presentation. The description of the first attribute can include a description of a new attribute that is added to the pre-existing structured presentation. An unstructured electronic document collection can include electronic documents available on the Internet. A structured presentation may be a table or a collection of cards. A structured presentation can be presented visually on a display screen, including physically transforming one or more elements of the display screen.

  Other embodiments of this aspect include corresponding systems, devices, and methods.

  Other aspects of the inventive subject matter described herein may be embodied by a system comprising a device and one or more computers programmed to interact with the device and perform operations. These operations involve receiving descriptive data describing pre-existing structured presentations, and pre-existing structured presentation characteristics with the contents of electronic documents in an unstructured collection of electronic documents. Identify electronic documents that compare and identify new values associated with pre-existing structured presentations, and add new values to pre-existing structured presentations to form new structured presentations And outputting instructions for visually presenting the new structured presentation on the device. Visual presentation of pre-existing structured presentation visually presents information in a system layout that matches the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values.

  Other embodiments of this aspect include corresponding computer program products, apparatus, and methods.

  Other aspects of the inventive subject matter described herein may be embodied by a system comprising a device and one or more computers programmed to interact with the device and perform operations. These operations receive descriptive data describing a first instance, a second instance, and a first attribute, and the first instance of two or more documents in the unstructured electronic document collection Extracting a first collection of first attribute values and extracting a second collection of first attribute values of a second instance from two or more documents of an unstructured electronic document collection Determining a first subset of the first collection of values as appropriately characterizing the first attribute of the first instance, and a second subset of the second collection of values, A structured presentation including determining a first attribute of the second instance as appropriately characterizing and including a first value of the first subset and a second value of the second subset Generating machine readable instructions for displaying an action, wherein the structured presentation represents an association between an instance and a value characterizing the instance attribute by an instance identifier and an array of values; Transmitting readable instructions to the device.

  Other embodiments of this aspect include corresponding computer program products, apparatus, and methods.

  Other aspects of the subject matter described herein include a first instance and a first instance of a structured presentation that characterizes actions for displaying the structured presentation on a display device and user interaction with the displayed structured presentation. May be embodied by a machine-implemented method including an action of receiving data including an attribute specification of 1 and an action of displaying a search interface previously hidden on the display device in response to receiving the data . A structured presentation visually presents information in an organized structured arrangement that fits the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values. The search interface includes information or interactive elements that identify the location of the first value that characterizes the first attribute of the first instance in the electronic document collection.

  This and other aspects include one or more of the following features. Receiving data characterizing user interaction with the displayed structured presentation includes receiving manual user designation of the first instance and the first attribute associated with a cell in the structured presentation, or the structured presentation Receiving data characterizing user interaction with the cells within can be included. The cell can be associated with the first instance and the first attribute using an array of cells with respect to the identifier of the first instance and the first attribute in the structured presentation. Receiving data characterizing user interaction with a cell may also include receiving data characterizing user interaction with an empty cell. Displaying a previously hidden search interface triggers a search of an electronic document collection and displays interactive elements that can be selected by the user to identify the first value. Displaying an interactive value input element that can be selected by the user to specify a value that characterizes the first attribute; displaying a snippet characterizing the context of the first value in the first document of the electronic document collection; And one or more of displaying the results of a pre-search of the electronic document collection to identify the first value. The first value may appear in the structured presentation as a value that characterizes the first attribute of the first instance. Displaying the previously hidden search interface can also include displaying the identifier of the first electronic document in the electronic document collection, where the first value is derived from the first electronic document. It is. The method may also include determining that the first electronic document does not function and cannot provide the first value and displaying a visual indication that the first document does not function. it can. An option for the user to select the first value consistently from the first document regardless of changes in the association of the first document with the first instance and the first attribute, or the first instance and An option may be presented to select the first value from the first document that is most relevant to the first attribute. The method also includes searching the unstructured collection of electronic documents to identify the first value in response to user interaction with the search interface and adding the first value to the structured presentation. Can do. Receiving the designation of the first instance and the first attribute may include receiving a designation of the collection of attributes or the collection of instances. The method may also include updating the display of the structured presentation over time.

  Other embodiments of this aspect include corresponding systems, devices, and computer program products.

  Other aspects of the inventive subject matter described herein may be embodied by a system comprising one or more computers programmed to interact with client devices and perform operations. These operations receive data characterizing user interaction that specifies the first cell of the structured presentation displayed on the display device, and a pre-search is performed to write the first value to the first cell. Determining that the pre-search has been performed, and displaying information characterizing the pre-search on the display device in response to determining that the pre-search has been performed. A structured presentation visually presents information in an organized structured arrangement that fits the structured design. A structured presentation represents the association between an instance and the values that characterize the attributes of the instance by an array of values in the cell.

  This and other aspects include one or more of the following features. The data characterizing the user interaction specifying the first cell may include a manual user specification of the first instance and the first attribute associated with the first cell. The information characterizing the pre-search can include information for identifying the electronic document from which the first value is derived. Information that characterizes the prior search includes a collection of electronic documents that may have been the source of the first value, information that identifies the first electronic document in the electronic document collection from which the first value has been extracted, One or more of the snippets that characterize the context of the first value in the first document of the electronic document collection may be included. Information characterizing the pre-search can be displayed, for example, in a search interface display element that was previously hidden. These operations include determining that the first electronic document is inoperable and cannot provide the first value, and displaying a visual indication that the first document is inoperable Can also be included. These operations may also include updating the display of values in the first cell of the structured presentation in response to user interaction. The collection of electronic documents can include electronic documents available on the Internet. The electronic document can include a web page. A structured presentation may be a collection of cards.

  Other embodiments of this aspect include corresponding computer program products, apparatus, and computer program products.

  Other aspects of the inventive subject matter described herein are embodied by a system comprising one or more computers programmed to interactively operate and perform operations on a client device comprising a display device. sell. These operations include displaying a structured presentation on a display device, receiving data characterizing user interaction with the displayed structured presentation, and previously displaying data on the display device in response to receiving the data. Includes displaying a hidden search interface. A structured presentation visually presents information in an organized structured arrangement that fits the structured design. A structured presentation represents the association between an instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values. The data includes a first instance of the structured presentation and a first attribute specification. The search interface includes information or interactive elements that identify the location of the first value that characterizes the first attribute of the first instance in the electronic document collection.

  Other embodiments of this aspect include corresponding computer program products, apparatus, and computer program products.

  The details of one or more embodiments are set forth in the accompanying drawings and the description below. Other features and advantages will be apparent from the description and drawings, and from the claims.

1 is a schematic diagram illustrating a system in which information from an electronic document collection is presented to a user in a structured presentation. FIG. 6 is a schematic diagram illustrating one implementation of another system in which information from an electronic document collection is presented to a user in a structured presentation. 1 is a schematic diagram of an exemplary structured presentation. 1 is a schematic diagram of an exemplary structured presentation. 1 is a schematic diagram of an exemplary structured presentation. 2 is a flowchart of an exemplary process for presenting information from an electronic document collection to a user in a structured presentation. 2 is a flow diagram of an example process for identifying two or more related documents in an electronic document collection. 2 is a flow diagram of an example process for identifying two or more related documents in an electronic document collection. FIG. 6 is a flow diagram of a process for proposing and / or adding a new instance to a structured presentation. Fig. 6 is a schematic diagram of a user interface component for receiving user input specifying a modification of a structured presentation. FIG. 6 is a schematic diagram of a user interface component for receiving user input specifying a technique for adding a new instance to a structured presentation. FIG. 6 is a schematic diagram of a user interface component for receiving user input specifying constraints that are used in user specified constraint options to add a new instance to a structured presentation. 2 is a flow diagram of an exemplary process for adding new attributes to a structured presentation. Fig. 6 is a schematic diagram of a user interface component for adding new attributes to a structured presentation. 2 is a flow diagram of an exemplary process for adding new attribute values to a structured presentation. 2 is a flow diagram of an exemplary process for adding new attribute values to a structured presentation. Fig. 4 is a schematic diagram of a user interface component for selecting candidate values to be added to a structured presentation. FIG. 6 is a schematic representation of a structured presentation including highlights of lack of attribute values presented therein. Fig. 6 is a schematic diagram of a user interface component for selecting candidate attributes to be added to a structured presentation. Fig. 6 is a schematic diagram of a user interface component for selecting candidate instances to be added to a structured presentation. FIG. 6 is a schematic diagram of a process by which new instances can be added to extend a pre-existing structured presentation. 3 is a flow diagram of an example process for adding an instance to a structured presentation based on the content of documents in an electronic document collection. 4 is a flow diagram of an example process for constructing an instance proposal from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. It is a figure showing the structure of the instance proposal from the electronic document in the electronic document collection based on the characteristic of the structured presentation which exists beforehand. 4 is a flow diagram of an example process for constructing an instance proposal from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. It is a figure showing a part of template of the hypertext markup language used as a pattern for describing a moving image. FIG. 6 is a schematic diagram of a process by which a collection of new instance proposals can be constructed based on information in a pre-existing structured presentation. 4 is a schematic diagram of a table associating attributes and instances in an electronic document collection. FIG. 6 is a flow diagram of a process for constructing an instance proposal from a collection of instances and attributes based on pre-existing structured presentation characteristics. FIG. 4 is a flow diagram of a process for constructing a collection of new instance proposals based on information in a pre-existing structured presentation. FIG. 4 is a flow diagram of a process for constructing a collection of new instance proposals based on information in a pre-existing structured presentation. Fig. 6 is a schematic diagram of a table that associates attributes, instances, and their values in a data collection. FIG. 4 is a flow diagram of a process for constructing a collection of new instance proposals based on information in a pre-existing structured presentation. It is a figure showing the structure of the instance proposal from the electronic document in the electronic document collection based on the characteristic of the structured presentation which exists beforehand. FIG. 4 is a schematic illustration of a collection of processes that can be used to construct a collection of new instance proposals based on information in a pre-existing structured presentation. FIG. 4 is a flow diagram of a process for constructing a collection of new instance proposals based on information in a pre-existing structured presentation. FIG. 6 is a schematic diagram of a process by which new attributes can be added to extend a pre-existing structured presentation. 3 is a flow diagram of an example process for adding attributes to a structured presentation based on the content of documents in an electronic document collection. 6 is a flowchart of an exemplary process for constructing attribute proposals from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. It is a figure showing the structure of the attribute proposal from the electronic document in the electronic document collection based on the characteristic of the structured presentation which exists beforehand. 6 is a flowchart of an exemplary process for constructing attribute proposals from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. It is a figure showing a part of template of the hypertext markup language (HTML) used as a pattern for describing a moving image. FIG. 6 is a schematic diagram of a process by which a collection of new attribute proposals can be constructed based on information in a pre-existing structured presentation. 4 is a schematic diagram of a table associating attributes and instances in an electronic document collection. FIG. 6 is a flow diagram of a process for constructing an attribute proposal from a collection of instances and attributes based on pre-existing structured presentation characteristics. FIG. 5 is a flow diagram of a process for constructing a collection of new attribute proposals based on pre-existing information in a structured presentation. FIG. 4 is a flow diagram of a process for identifying related instances for use in constructing an attribute proposal based on information in a pre-existing structured presentation. FIG. 5 is a flow diagram of a process for constructing a collection of new attribute proposals based on pre-existing information in a structured presentation. It is a figure showing the structure of the attribute proposal from the electronic document in the electronic document collection based on the characteristic of the structured presentation which exists beforehand. FIG. 4 is a schematic illustration of a collection of processes that can be used to construct a collection of new attribute proposals based on information in a pre-existing structured presentation. FIG. 5 is a flow diagram of a process for constructing a collection of new attribute proposals based on pre-existing information in a structured presentation. 1 is a schematic diagram illustrating a system in which attribute values derived from two or more electronic documents in an electronic document collection are presented to a user in a structured presentation. 1 is a schematic diagram illustrating one implementation of a system in which attribute values derived from two or more electronic documents in an electronic document collection are presented to a user in a structured presentation. Fig. 6 is a schematic diagram of a table that can associate attributes, values, and instances in an electronic document collection. 6 is a flow diagram of an example process for presenting attribute values derived from two or more electronic documents in an electronic document collection to a user in a structured presentation. 2 is a flow diagram of a process for selecting one or more values to present in a structured presentation. 2 is a flow diagram of a process for selecting one or more values to present in a structured presentation. 2 is a flow diagram of an example process for selecting one or more values to present in a structured presentation. 4 is a schematic diagram illustrating a situation in which attribute values derived from electronic documents in an electronic document collection are presented to a user in a structured presentation. FIG. 6 is a schematic diagram illustrating a process in which both attributes and attribute values are derived from an electronic document in an electronic document collection and presented to a user in a structured presentation. 3 is a flow diagram of a process for adding values to a structured presentation based on the content of documents in an electronic document collection. Fig. 6 is a schematic representation of a structured presentation with a hidden search interface. Fig. 6 is a schematic representation of a structured presentation with a hidden search interface. Fig. 6 is a schematic representation of a structured presentation with a hidden search interface. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. 2 is a flow diagram of a process for adding values to a structured presentation by extracting values from the content of documents in an electronic document collection. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. FIG. 6 illustrates a display element on which a previously hidden search interface is presented. 3 is a flow diagram of a process for adding values to a structured presentation based on the content of documents in an electronic document collection.

  Like reference symbols in the various drawings indicate like elements.

  FIG. 1 is a schematic diagram illustrating a system 100 in which information from an unstructured electronic document collection 102 is presented to a user in a structured presentation 106. In addition to the electronic document collection 102, the system 100 includes a display screen 104 and a data communication infrastructure 108. In operation, the system 100 extracts information from an unstructured collection of electronic documents 102 and presents the extracted information on the display screen 104 in a structured presentation 106.

  The electronic document collection 102 is an unstructured collection in that the organization of information in individual documents in the electronic document collection 102 does not have to conform to a predetermined structure that can be utilized in extracting information. For example, consider three electronic documents in the electronic document collection 102, that is, electronic documents 110, 112, 114. Documents 110, 112, 114 were added to collection 102 by three different users who organized the contents of each electronic document separately. Users do not need to collaborate to ensure that the information in documents 110, 112, 114 is in a particular format. Furthermore, if one user wants to change the format of document 110, that user can do so regardless of the format of the document added by the other user. The user does not need to notify other users of the change. In fact, documents may be added to the collection 102 by entities that are not only collaborative, but also competitors, such as three car manufacturers or three digital camera sales companies. Even though there is a specific linkage between entities that add documents to the collection 102, there is no official mechanism to ensure that the information in the documents is similarly organized within those documents. In addition, there is no formal mechanism to ensure that the organization of information in each of the documents in collection 102 is unchanged.

  In contrast, structured presentation 106 is structured and presents information derived from documents in a collection 102 of organized system arrangements. Thus, the grouping, segmentation, and arrangement of information in the structured presentation 106 follows the structured design even though the information therein is drawn from different contexts in the diverse set of documents in the collection 102. Further, changes made to one aspect of the design of the structured presentation 106 can propagate throughout the structured presentation 106.

  Examples of structured presentations include spreadsheet tables, collections of cards or other records, and other structured presentation formats. Such structured presentations are the positioning and identification of values, the spatial arrangement of information on the display, the organization of the structured presentation and various aspects of information (e.g., column headings, row headings, unit identifiers, and the like), values It may follow the rules for specifying the graphical representation of and other characteristics.

  The structuring of information in structured presentations generally makes it easier for viewers to understand the information. For example, the viewer can recognize the nature of the information stored in the structured presentation by reading the headline. The viewer can easily identify and compare the values described in the structured presentation based on the arrangement and positioning of those values shown on the display. For example, the user can easily verify that some values in the structured presentation are all related to different car attributes (ie, characteristics) and easily compare those values.

  System 100 is not limited to simply writing values derived from documents in collection 102 into structured presentation 106. Instead, in many implementations, the system 100 is responsible for the entities described in the structured presentation 106 (ie, “instances”), the values that characterize the attributes of those instances, and the appropriate structure of the structured presentation 106. Can be determined. Such a determination can be made based on information derived from different documents in the collection 102 that are not limited to having a particular format, a permanent format, or both. For example, attributes appearing in the structured presentation 106 may be based on attributes used in documents in the collection 102 to characterize some instances, as will be described in more detail below. As another example, the units of value that appear in structured presentation 106 (eg, meters, feet, inches, miles) may be based on the units of value that appear in documents in collection 102. As another example, instances that appear in structured presentation 106 may be determined based on a collection of instances that appear in documents in collection 102.

  Further, in many implementations, such information can be derived from previously unspecified documents in collection 102. For example, a search query can be used to identify documents in collection 102 and extract information from those documents. There is no need to have pre-existing restrictions on the identification or type of documents from which information can be derived. For example, the identified documents need not be limited to those associated with a particular individual's account or from a particular retail store. Instead, the information can be extracted from documents that have not been previously specified.

  The system 100 can thus present a structured presentation to the user using the diverse information content of the documents in the collection 102 in a variety of different ways. If the electronic document collection 102 includes a large number of documents, the amount of information available can be very large. Furthermore, this can often be performed automatically or with relatively little human interaction as described in more detail below.

  FIG. 2 is a schematic diagram illustrating one implementation of a system 200 in which information from an unstructured electronic document collection 102 is presented to a user in a structured presentation 106. In system 200, data communication infrastructure 108 includes electronic document collection 102, display screen 104, search engine 202, crawler 204, data center 208, and data storage including document compression, indexing and ranking module 210. Interconnect elements and collections of processing elements.

  The search engine 202 is programmed with one or more machine readable instruction sets to search the unstructured electronic document collection 102. Search engine 202 may be implemented on one or more computers that are deployed in one or more geographic locations.

  The crawler 204 is programmed with one or more machine readable instruction sets to perform crawling on the unstructured electronic document collection 102. The crawler 204 can be implemented on one or more computers that are deployed in one or more geographic locations.

  The compression, indexing, and ranking module 210 is programmed with one or more machine-readable instruction sets to compress, index, and rank documents in the collection 102. The compression, indexing, and ranking module 210 can be implemented on one or more computers deployed at one or more geographic locations.

  The data center 208 stores information characterizing the electronic documents in the electronic document collection 102. Information characterizing such an electronic document may be stored in the form of an indexed database that includes indexed keywords and the location of the document in the collection 102 where the keywords can be found. The indexed database may be formed by the crawler 204, for example.

  In some implementations, the information stored in the data center 208 may itself be organized to facilitate presenting the structured presentation 106 to the user. For example, information may be organized by crawler 204 and compression, indexing, and ranking module 210 in anticipation that it is necessary to present structured presentation 106 related to several topics. The structure of the information in the data center 208 may facilitate grouping, segmenting, and arranging information in the structured presentation 106. This organization can be based on a variety of different factors. For example, an ontology can be used to organize information stored in the data center 208. As another example, the historical records of the previous structured presentation 106 can be used to organize the information stored in the data center 208. As another example, the data tables described herein can be used to organize information stored in the data center 208.

  As shown, system 200 includes a plurality of display screens 104 that can present a structured presentation according to machine-readable instructions. Examples of the display screen 104 include a cathode ray (CRT), a light emitting diode (LED) screen, a liquid crystal display (LCD), a gas plasma display, and the like. Display screen 104 may be an integral part of a self-contained data processing system, such as a personal data assistant (PDA) 215, desktop computer 217, or mobile phone. In general, instructions for presenting a structured presentation are modified to the display screen 104 specifications after being received by such a self-contained data processing system. But this is not always the case. For example, the display screen 104 may be part of a more distributed system in which processing of instructions to present a structured presentation is completed before the instructions reach the display screen 104. For example, display screen 104 can be incorporated into a “dumb” device, such as a television receiver or computer monitor, which receives instructions for presenting structured presentation 106 from a local or remote source.

  In operation, the system 200 can convert unstructured information in the collection 102 into a structured presentation 106 that is presented to the viewer. Such conversion may be performed in the context of a web search where a search engine receives an information request based on information extracted from an electronic document in collection 102 and responds to the information request.

  For example, a personal data assistant (PDA) 215 or desktop computer 217 can exchange information with the user, thereby receiving a search query using, for example, a web browser application. Query description 212 may be transmitted to search engine 202 via wireless data link 219 and / or wired data link 221. In response, search engine 202 can use query description 212 to identify information in data center 208 that can be used in presenting structured presentation 106 on display screen 104. The identified information can be derived from two or more unspecified electronic documents in the unstructured electronic document collection 102. In some cases, the query description 212 may include search terms that are used by the search engine 202 to retrieve information and present the structured presentation 106 to a user. For example, by using a search term in query description 212, other instances within data center 208 are associated with the instance, attributes that characterize such instances, values that characterize individual instances, and / or other structured presentation 106 A collection of aspects can be identified.

  The search engine 202 can also generate a response 214 to the query description 212. The response 214 can be used to present the structured presentation 106 to the user. In general, the response 214 includes machine-readable instructions that can be interpreted by a data processing device in the systems 215, 217 to present the structured presentation 106. For example, the response 214 can be encoded in HTML to specify the characteristics and content of the structured presentation 106. In other implementations, the response 214 can include a text snippet or other information from the data center 208 that is used to present the structured presentation 106. For example, the response 214 can include a collection of values, names of new attributes, and estimates of the probability that the value displayed in the structured presentation 106 is correct, as will be described in more detail below.

  In many cases, system 200 uses information stored in data center 208 to identify the location of one or more documents associated with the query described in query description 212. For example, the search engine 202 can compare keywords in the query description 212 with keyword indexes stored in the data center 208. This comparison result can be used to identify documents in the collection 102 associated with the query description 212. The location of such identified document can be included in response 214, eg, as a hyperlink to a document that responds to the described query.

  In some implementations, the system 200 may store attributes and / or their respective values to facilitate grouping, segmentation, and arrangement of information in the structured presentation 106. For example, when a user interacts with and modifies a client system such as systems 215, 217, a collection of instances, their attributes, and their values as a structured presentation 106 in data center 208 Can be stored. For example, when storing instances, attributes, and values in one structured presentation 106 presented to a first viewer in the data center 208 and providing subsequent structured presentations 106 to other viewers Can be used.

  FIG. 3 is a schematic diagram of an exemplary structured presentation 106, ie, including table 300. Table 300 is an organized system arrangement of one or more identifiers of instances, as well as the values of certain attributes of those instances. Instances are individually identifiable entities and generally share at least some common attributes. An attribute is a property, feature, or characteristic of an entity. For example, Tom, Dick, and Harry are personal instances. Each such individual has attributes such as name, height, weight, and the like. As another example, each city instance has a geographical location, a mayor, and a population. As yet another example, a product instance may have a model name, manufacturer, and year.

  Instance attributes can be characterized by values. Thus, the value of a particular attribute of a particular instance characterizes that particular instance. For example, an individual's name may have a value of “Tom”, a city population may have a value of “4 million”, and a product model name may have a value of “Wrangler”. In some implementations, structured presentations, such as table 300, may also include attribute identifiers as well as unit identifiers used in representing values.

  The grouping, segmentation, and arrangement of information in the table 300 can be selected so that the user can easily understand the information. In this regard, the table 300 includes a collection of rows 302. Each row 302 includes a collection of attribute values 307 associated with the instance identifier 306. Thus, the arrangement and positioning of attribute values 307 and instance identifiers 306 in row 302 graphically represent the association between them. For example, the user can recognize an association between the attribute value 307 and the instance identifier 306 in the same row 302.

  Table 300 also includes a collection of columns 304. Each column 304 includes a collection of attribute values 307 associated with the attribute identifier 308. Thus, the arrangement and positioning of attribute values 307 and attribute identifiers 308 in column 304 graphically represent the association between them. For example, the user can recognize an association between the attribute value 307 and the attribute identifier 308 in the same column 304 based on the cooperation.

Each row 302 is a structured record 310 in that each row 302 associates a single instance identifier 306 with an associated collection of attribute values 307. Further, the arrays and positioning used to represent their association in one structured record 310 are replicated in the other structured record 310 (ie, in the other row 302). In fact, in many cases, all structured records 310 in structured presentation 106 are limited to having the same arrangement and positioning of information. For example, the value “307” of the attribute “ATTR_2” is restricted to appear in the same column 304 in all rows 302. As another example, all attribute identifiers 308 have the same spatial relationship as values 307 appearing in the same column 304. In addition, changes made to the arrangement and positioning of information in one structured record 310 generally propagate to other structured records 310 in the structured presentation 106. For example, a new attribute (e.g., "Attr_2 _^3/4") if the new attribute value 307 characterizing be added to one of the structured record 310, the value is structured presentation attributes for all instances "Attr_2 _^3/4" A new column 304 is added to structured presentation 106 so that it can be added to 106.

  In some implementations, the value 307 in the table 300 can be presented in several units of measurement. Examples of units of measurement include feet, yards, inches, miles, seconds, gallons, liters, degrees Celsius, and similar units. In some cases, the unit of measurement used in presenting the value 307 is indicated by the unit identifier 309. The unit identifier 309 may be displayed beside the value 307 and / or beside the associated attribute identifier 308, for example. The association between the unit identifier 309 and the value 307 indicating the unit of measurement is indicated to the viewer by such an arrangement. In many cases, all of the values 307 associated with a single attribute (eg, all of the values 307 in a single column 304) are limited to presentation in the same unit of measure.

  Information extracted from the electronic document collection 102 by the systems 100, 200 can affect the presentation of the table 300 to the user in various ways. For example, the value 307 to be written to the table 300 can be determined using information extracted from the electronic document collection 102. As another example, by using information extracted from the electronic document collection 102, new attributes and / or new instances to be added to the table 300 can be proposed.

  In some implementations, the instance identifier 306 can be selected based on one or more search strings. For example, when search engine 202 receives a search string “hybrid vehicles” from a user, a system such as system 200 generates a table 300 based on information extracted from electronic document collection 102 using the search string. , Can be written to the table 300. For example, the system 200 accesses the data center 208, identifies the instance identifier 306 in the electronic document associated with the search string, and sets a set of common attributes for the identified instance (and identifiers 308 and Values 307) for those attributes can be determined. In fact, the system 200 can determine an instance identifier 306, an attribute identifier 308, and an associated value 307 based on the received search string.

  In some implementations, one or more attribute identifiers 308, instance identifiers 306, and / or values 307 may be received from a user viewing the display of table 300. As described in more detail below, a system such as system 200 is based on information extracted from electronic document collection 102 using one or more received attribute identifiers 308, instance identifiers 306, and / or values 307. The table 300 can be generated and written to the table 300. Indeed, the system 200 can configure the new instance identifier 306, the attribute identifier 308, and even the associated value 307 based on the received attribute identifier 308, instance identifier 306, and / or value 307.

  FIG. 4 is a schematic diagram of another implementation of a structured presentation, ie, one that includes a table 400. In addition to including attribute identifiers 308, instance identifiers 306, values 307, unit identifiers 309 organized in rows 302 and columns 304, table 400 also includes a number of interactive elements for users to exchange information. . In particular, table 400 includes a collection of instance selection widgets 405, a collection of action triggers 410, a collection of column action trigger widgets 415, and a notes column 420. including.

  The instance selection widget 405 is a user interface component that allows a user to select a structured record 310 in the table 400. For example, the instance selection widget 405 may be a collection of clickable check boxes that are associated with a particular structured record 310 using an array and positioning associated with that structured record 310. The instance selection widget 405 includes a mouse (e.g., hovering over a component and clicking a specific mouse button), a stylus (e.g., pressing a user interface component displayed on the touch screen with the stylus), a keyboard Or “clickable” in that the widget 405 can be manipulated interactively using other input devices that invoke functions provided by the component.

  Action trigger 410 is a user interface component that allows a user to trigger the execution of an action on one or more structured records 310 in table 400 selected using instance selection widget 405. For example, the action trigger 410 can be a clickable text phrase, each of which can be used by a user to trigger an action described in the phrase. For example, the “keep and remove others” action trigger 410 triggers the deletion of a structured record 310 that has not been selected using the instance selection widget 405 from the display of the table 400. As another example, the “remove selected” action trigger 410 triggers the removal of the structured record 310 that has been selected using the instance selection widget 405 from the display of the table 400. As yet another example, the “show on map” action trigger 410 triggers the display of the location of the structured record 310 that is selected using the instance selection widget 405 on the map. For example, if the selected instance is a car, the location of the car dealer selling the selected car can be displayed on the map. As another example, if the selected instance is a spring vacation destination, those destinations can be displayed on a map.

  Column action trigger widget 415 is a user interface component that allows a user to apply an action to all of the cells in a single column 304. When the user interacts with the clickable “+” symbol, other user interface components are displayed, giving the user a set of possible actions to perform. This set of actions includes, for example, deleting the entire column 304 from the structured presentation 400, or searching to find values for all cells in the column 304 that are currently blank.

  Annotation column 420 is a user interface component that allows a user to associate information with instance identifier 306. In particular, the annotation column 420 includes one or more annotations 425 that are each associated with the structured record 310 using the alignment and positioning associated with that structured record 310. Unlike the column 304, the information content of the annotation 425 is unlimited in that the annotation 425 is not claimed to be the value of a particular attribute. Instead, the information contained in the annotation 425 can characterize an unrelated aspect of the instance identified in the structured record 310.

  In some implementations, the table 400 may include additional information other than the values of certain attributes. For example, the table 400 can include a collection of images 430 that are associated with the instance identified in the structured record 310 using the alignment and positioning associated with that structured record 310. As another example, table 400 can include a collection of text snippets 435 extracted from electronic documents in collection 102. The source of the snippet may be a high order result of a search performed using the instance identifier 306 as a search string. Text snippet 435 is associated with the instance identified in structured record 310 using the alignment and positioning associated with that structured record 310.

  As another example, the table 400 may include one or more hypertext links 440 to individual electronic documents in the collection 102. For example, a linked document may be the result of a high order search performed using the instance identifier 306 as a search string. As another example, the linked document may be the source of the extracted value 307 for writing to the table 400. In some cases, interactive manipulation of the hypertext link 440 may trigger navigation to the source electronic document based on information embedded in the hypertext 440 (eg, a website address).

  FIG. 5 is a schematic diagram of another implementation of a structured presentation, ie, one that includes a collection 500 of cards. Card collection 500 is an organized system arrangement of one or more identifiers of instances, as well as the values of certain attributes of those instances. Instance attributes can be specified by value. Further, the card collection 500 generally includes an identifier of an attribute, and an identifier of a unit used when representing a value as appropriate.

  The grouping, segmentation, and arrangement of information in the card collection 500 can be selected so that the user can easily understand the information. In this regard, the card collection 500 includes a collection of cards 502. Each card 502 includes a collection of attribute values 307 associated with the instance identifier 306. Accordingly, the arrangement and positioning of the attribute value 307 and the instance identifier 306 in the card 502 graphically represents the association between them. For example, the user can recognize the association between the attribute value 307 and the instance identifier 306 in the same card 502.

  In the illustrated implementation, the card 502 in the card collection 500 also includes a collection of attribute identifiers 308. Attribute identifier 308 is organized in column 504 and attribute value 307 is organized in column 506. Columns 504, 506 are placed adjacent to each other and aligned so that individual attribute identifiers 308 are placed next to attribute values 307 that characterize the identified attributes. Using this positioning and alignment, the viewer can recognize the association between attribute identifiers 308 and attribute values 307 characterizing those attributes.

  Each card 502 is a structured record 310 in that each card 502 associates a single instance identifier 306 with an associated collection of attribute values 307. Further, the arrangement and positioning used to represent their association in one card 502 is replicated in the other card 502. In fact, in many cases, all of the cards 502 are limited to having the same arrangement and positioning of information. For example, the value 307 that characterizes the attribute “ATTR_1” is restricted so that the spatial relationship with the instance identifier 306 is the same for all cards 502. As another example, the order and positioning of attribute identifiers 308 in all of the cards 502 is the same.

  In addition, changes made to the arrangement and positioning of information in one card 502 generally propagate to other cards 502 in the card collection 500. For example, if a new attribute value 307 characterizing a new attribute (for example, “ATTR_1 3/4”) is inserted between attribute values “value_1_1” and “value_2_1” in one card 502, The positioning of the corresponding attribute value 307 is similarly changed.

  In some implementations, the cards 502 in the card collection 500 may have other features. For example, the card 502 can include interactive elements for interacting with the user, such as instance selection widgets, action triggers, attribute selection widgets, annotation input functions, and the like. As another example, a card 502 in the card collection 500 can include additional information other than the value of a particular attribute, for example, a text snippet associated with an image and / or an identified instance. As another example, cards 502 in card collection 500 can include one or more hypertext links to individual electronic documents in collection 102. Such a feature can be associated with a particular instance by displaying it on a card 502 that includes an instance identifier 306 that identifies that instance.

  In operation, a viewer can interactively manipulate the system presenting the card collection 500 to change the display of one or more cards 502. For example, the viewer can trigger a side-by-side display of two or more of the cards 502 to facilitate comparison of specific instances identified on those cards. As another example, the viewer can trigger a rearrangement of cards 502, termination of display of a particular card 502, or a similar operation. As another example, a viewer can trigger selection, modification, addition, and / or deletion of attributes and / or instances displayed in card 502. As yet another example, the viewer can trigger an operation of sorting the cards into a plurality of mountains, for example, according to the value of the attribute value 307 in the card.

  In some implementations, the card 502 is displayed “double-sided”. For example, the first side can include a graphical representation of the instance identified by the instance identifier 306, and the second side can include the instance identifier 306 and the value 307. This is useful, for example, when a user searches for a particular card in the card collection 500, where the user can identify a particular card simply by looking at the graphic representation on the first side of the card 502. Can do.

  FIG. 6 is a flowchart of an exemplary process 600 for presenting information from an electronic document collection to a user in a structured presentation. Process 600 may be performed by one or more computers that perform operations by executing one or more machine-readable instruction sets. For example, process 600 can be performed by search engine 202 within system 200. In some implementations, the process 600 is performed in response to receiving a trigger, eg, a user request, which can create or modify a structured presentation.

  The system performing process 600 may identify two or more responsive electronic documents in the electronic document collection (step 605). The response document can be identified in many different ways. In some cases, the document is identified based on “new” information (eg, a new search query) received from the viewer. For example, the system can compare a newly received search query with the content of an electronic document in the electronic document collection using string comparison. As another example, the system may access a data center, such as data center 208, to compare the search terms in the search query to the keyword index and identify the location of the response electronic document.

  In some cases, documents are identified based on “old” information already found in the structured presentation. Information found in the structured presentation includes identification of instances, attributes, values, and units that represent values. The system executing process 600 can use this old information to identify response electronic documents in the electronic document collection. For example, a document containing an instance already found in a structured presentation can be identified as being a response document. As another example, a document that characterizes an instance using attributes already found in a structured presentation can be identified as being a response document. Other examples of such identification are described below.

  The system performing process 600 may also collect information from the identified electronic document (step 610). The collected information may relate to one or more instances, attributes, and / or values. The system performing process 600 can collect this information directly from the documents in the electronic document collection or from an already collected collection of information characterizing the electronic documents in the electronic document collection. For example, in the context of system 200 (FIG. 2), the system executing process 600 identifies a document in collection 102, accesses the identified document, and directly receives information from the original document in collection 102. Can be extracted. As another example in the context of system 200 (FIG. 2), the system performing process 600 accesses a collection of information in data center 208, such as an index of keywords and the location of documents that contain those keywords, an ontology. Information may be collected from a database that stores historical records of previous structured presentations that were presented using information extracted from documents in collection 102 and / or.

  The system performing process 600 may use the collected information and provide instructions for presenting the structured presentation based on the collected information (step 615). For example, a system performing process 600 may generate machine-readable instructions for presenting a structured presentation, eg, table 300, 400 or collection of cards 500.

  FIG. 7 is a flow diagram of an example process 700 for identifying response documents in an electronic document collection. Process 700 may be performed independently of or in conjunction with other data processing activities. For example, process 700 may be performed at step 605 in process 600 (FIG. 6).

  The system executing process 700 receives the search query (step 705). For example, the system can receive one or more search strings (eg, “hybrid vehicles”) from a user. As another example, the system can receive search strings from other processes or systems. In some implementations, the search string is received through an application programming interface (API), a common gateway interface (CGI) script, or other programming interface. In other implementations, the search string is received through a web portal, web page, or website, or the like.

  In response, the system executing process 700 identifies two or more documents that include instances, attributes, and / or values that respond to the search query (step 710). These documents can be identified by categorizing the role that the search terms in the search query should play in the structured presentation. For example, the search terms in the search query are structured based on, for example, specific search terms in the search query, explicit instructions by the user on how to classify the search terms in the search query, and / or the search context It can be categorized as a categorization of instances that should appear in the presentation. For example, the search terms in the search query “cities in California” are “San Diego”, “Los” because the plural word “cities” is characterized by an attribute, that is, qualified by “in California”. Can be categorized as an categorization of instances such as "Angeles" and "Bakersfield". As another example, the search terms in the search query "Ivy League schools" are instances ("Cornell", "Columbia", and "" because the plural word "schools" is characterized by the attribute "Ivy League". Brown ”etc.).

  In some cases, additional information must be used to classify search terms within the search query. For example, the search query “Ivy League” can be viewed as an categorization of an instance of “school” or as an exemplary instance of a category “athletic conferences” that includes instances such as “Atlantic Coast Conference” and “PAC-10”. Is also reasonable. In such a case, these search terms can be classified based on, for example, an explicit instruction from the user regarding how to classify or the background status of the search terms in the search session. For example, if the user has previously entered the phrases “Atlantic Coast Conference” and “PAC-10” as search queries, the search query “Ivy League” would appear in a structured presentation along with other instances. It can be regarded as an instance.

  These documents may be identified directly in the electronic document collection 102 or indirectly based on information in the electronic data center 208. Such identification information can include, for example, the URL of the last document found after crawling.

  FIG. 8 is another flow diagram of an example process 800 for identifying two or more response documents in an electronic document collection. Process 800 may be performed independently of or in conjunction with other data processing activities. For example, process 800 may be performed at step 605 in process 600 (FIG. 6). As another example, process 800 may be performed in conjunction with process 700 at step 605 in process 600 (FIG. 6). For example, the processes 700, 800 can be part of an iterative interactive process, in which a search query is received and used to identify a first collection of response documents and the identified documents A first structured presentation that includes content derived from is presented to the user, receives user modifications, and uses the modified structured presentation description to identify a second collection of related documents. In some implementations, the process 800 can be executed multiple times. In some implementations, the process 800 can be performed without user input, for example by the crawler 204 in the system 200 (FIG. 2).

  The system performing process 800 receives a description of the existing content of the structured presentation (step 805). In particular, the system can receive a description of units that present instances, attributes, values, and / or values in an existing structured presentation. The description may include, for example, an instance and attribute identifier and / or a range of attribute values. The description can also include categorization of instances and / or attributes. Such categorization can be determined, for example, using an ontology or based on a categorization assigned to a structured presentation by a viewer. For example, if a user titles a structured presentation “Ivy League Schools”, this title can be considered as a categorization of the instances in the structured presentation.

  In response, the system performing process 800 may identify one or more documents that include instances, attributes, and / or values associated with existing content (step 810). For example, the system compares an instance and / or attribute identifier with an indexed keyword to determine one or more of the instances and / or attributes that a particular document already appears in the existing content of the structured presentation or It can be determined whether to include multiple instances and / or attributes. As another example, the system identifies a new instance, its attributes, and the values of such attributes from such a document, and identifies those values as values that already appear in the existing content of the structured presentation. By comparison, it can be determined whether the new instance is potentially related to the existing content of the structured presentation.

  These documents can be identified directly in the electronic document collection 102 or using identification information in the electronic data center 208. Such identification information can include, for example, the memory location of the last document found by crawling.

  FIG. 9 is a flow diagram of a process 900 for proposing and / or adding a new instance to a structured presentation. Process 900 may be performed by one or more computers that perform operations by executing one or more machine-readable instruction sets. These digital data processing devices can exchange information with the user on input and output devices such as a keyboard, mouse, touch screen, display screen, and the like. For example, in the context of system 200 (FIG. 2), user interaction in process 900 can be performed on the client side, such as PDA 215 or desktop computer 217.

  Process 900 may be performed alone or in conjunction with other data processing activities. For example, as described in more detail below, process 900 may be performed in conjunction with various processes to construct an instance proposal in addition to pre-existing structured presentations. Examples of such configuration processes are described in FIGS. 21-26 and related text. In general, process 900 is performed by a plurality of digital data processing devices. For example, in the context of the system 200 (FIG. 2), activities for composing instance proposals can be performed on the search engine 202 side, while user interaction is such as PDA 215 or desktop computer 217 (FIG. 2) Can be executed on the client side.

  The system executing process 900 can receive a new instance trigger (step 905). A new instance is an instance that is not currently displayed in a structured presentation, eg, structured presentation 106 (FIG. 1). A new instance trigger activates a process for adding a new instance to a structured presentation. For example, a new instance can be triggered by user input received via a mouse, stylus, keyboard, or the like. In other implementations, the new instance can be triggered by other processes or systems. New instance triggers can be received through inter-process communication or application message handlers, to name a few.

  The system performing process 900 may present the user with the option to add a new instance to the structured presentation (step 910). The option is an alternative approach for adding new instances. Examples of options include fully automatic options, automatic options with user-specified constraints, and manual options. These options are described in more detail below.

  The system performing process 900 may present the user with an option to use a user interface device, such as a display screen. In many cases, the display screen presenting options may be the same display screen presenting a structured presentation that adds instances. For example, options can be presented to the user using display screen 104 (FIG. 1).

  The system performing process 900 may receive an optional user selection (step 915). User selections can be received using one or more input devices, such as a keyboard, touchpad, or touch screen. The system may also determine the nature of the option selected by the user (step 920).

  If the system running process 900 determines that the user has selected "Automatic Options", the system will suggest additional instances and / or structure the additional instances without interacting with the user Can be added to the presentation.

  In one implementation of a user-specified automatic option, a new instance can be proposed and / or added based on the characteristics of the structured presentation (step 925). Examples of such properties include the nature of the instances already specified in the structured presentation, the categorization of those instances, and the attributes of those instances. An approach for constructing a new instance based on such characteristics is described in FIGS. 21-26 and the associated text. For example, as described therein, a search query may include attribute identifiers derived from pre-existing structured presentations, attribute values derived from pre-existing structured presentations, and / or Can be built using a combination of These search queries can be used to identify instances to add to the structured presentation using string comparison results or other matching techniques.

  If the system executing process 900 determines that the user has selected the “user-specified constraints” option, the system suggests additional instances based on user-specified constraints on the nature of the additional instances, and / or Or that additional instance can be automatically added to the structured presentation. These constraints can be expressed as one or more parameters that characterize the proposed and / or added instances. For example, these constraints can be expressed as an acceptable value for an instance attribute or as a range of acceptable values for an attribute.

  In one implementation of the user-specified constraint option, the system executing process 900 presents the user with an option to constrain the value of the new instance attribute (step 930). For example, the system can display a list of attributes that characterize instances in the structured presentation, as well as input fields that allow the user to enter constraints on the values of those attributes. In many cases, attributes in such lists also appear in structured presentations to which new instances are added. However, in some implementations, attributes in such lists can be configured based on attributes used to characterize instances elsewhere, such as in documents of an electronic document collection. An exemplary approach for configuring such attributes is described in FIGS. 37-51 and the associated text.

  The system executing process 900 may also receive a user specification of one or more constraints on the value of the new instance attribute (step 935). As explained above, these constraints can limit the value of one or more attributes to a specific value or range of values. For example, one attribute characterizing a car is “number of cylinders”. A user-specified constraint on the value of this attribute restricts the number of cylinders in a new car instance to a specific value (for example, “6”) or a range of values (for example, “6 to 8” or “greater than 6”) Can do.

  The system performing process 900 may also propose and / or add new instances based on user-specified constraints and structured presentation characteristics (step 940). Examples of characteristics of a structured presentation include the nature of instances already specified in the structured presentation, the categorization of those instances, and the attributes of those instances. An approach for constructing a new instance based on such characteristics is described in FIGS. 37-51 and the associated text. As another example, the search query may be attribute identifiers derived from pre-existing structured presentations, attribute values derived from pre-existing structured presentations, and / or combinations thereof, or even Can be built using constraints specified by the user. These search queries can be used to identify instances using string comparison results or other matching techniques. The identified instances can then be proposed and / or added to the structured presentation.

  If the system executing process 900 determines that the user has selected the “manual option”, the system can add additional instances to the structured presentation according to the user's instructions.

  In one implementation of the manual option, the system executing process 900 can receive a new instance from the user (step 945). For example, a user can enter an instance name using a keyboard or other user input device. The system executing process 900 can add a new instance to the structured presentation (step 950). In general, the name of the new instance can be added directly to the structured presentation as the instance identifier 306 in the new structured record 310. In some implementations, the new structured record 310 can be a new row 302 (FIGS. 3 and 4) or a new card 502 (FIG. 5).

  In some embodiments, the system performing process 900 may also perform additional operations based on the received new instance. For example, the system can use the new instance to improve the set of proposed instances or the set of proposed attributes.

  FIG. 10 is a schematic diagram of a user interface component 1000 for receiving user input specifying a modification of a structured presentation. For example, at step 905 of process 900 (FIG. 9), user interface component 1000 can be used to receive a new instance trigger.

  The user interface component 1000 includes an attribute correction area 1005 and an instance correction area 1010. The attribute modification area 1005 includes a header 1015, a collection 1020 of attribute identifiers 1025 each associated with an attribute identifier selection widget 1030, and a new attribute addition trigger 1035.

  The header 1015 includes text or other information that identifies that user interaction with the attribute modification area 1005 actually allows the user to modify the attribute. The attribute identifier 1025 is text or other information that identifies an attribute included in the structured presentation. For example, attribute identifier 1025 may be the same text that appears as attribute identifier 308 in structured presentations 300, 400, 500 (FIGS. 3, 4, 5). The attribute identifier selection widget 1030 is an interactive display element that allows a user to select and deselect attributes for display in a structured presentation. For example, in collection 1020, each attribute identifier selection widget 1030 is associated with a single attribute identifier 1025 using its arrangement and positioning adjacent to each other. The attribute identifier selection widget 1030 can indicate whether one or more graphical displays, eg, the check mark and tint shown, are used to select or deselect the attribute identifier 1025 for display. For example, if the user interactively manipulates the checked attribute identifier selection widget 1030 associated with attribute identifier 1025 “Attribute_1”, the color and checked status in the attribute identifier selection widget 1030 is changed and the structured presentation Is triggered to delete the attribute identifier associated with “Attribute_1” (along with a value corresponding to “Attribute_1”).

  Add new attribute trigger 1035 is an interactive display element that the user can use to trigger the addition of a new attribute to the structured presentation. The method of configuring the new attribute to be added is described in FIGS. 37-51 and the associated text. The addition of new attributes is also described in more detail below, for example, in FIGS.

  The instance modification area 1010 includes a new instance addition trigger 1040 and an instance filter trigger 1045. Add new instance trigger 1040 is an interactive display element that a user can use to trigger the addition of a new instance to a structured presentation. For example, the add new instance trigger 1040 can be used at step 905 in the process 900 (FIG. 9).

  Instance filter trigger 1045 is an interactive display element that a user can use to trigger filtering of instances in a structured presentation. Instance filtering generates a collection of instances that meet one or more criteria. For example, filtering can generate a collection of instances that have a certain number of values or a number of values within a specified range. Therefore, the number of instances included in the structured presentation can be reduced by filtering.

  Filtering triggered by instance filter trigger 1045 presents a user interface component that allows the user to specify one or more filtering criteria, and does not display instances that cannot meet the criteria Modifying the structured presentation.

  In some implementations, the user interface component 1000 can dynamically respond to modifications made by the user using the user interface component 1000 or in some other way. For example, when a user executes a trigger and adds a new attribute to the structured presentation, the identifier of the new attribute can be added to the collection 1020 and presented in the user interface component 1000. For example, if the user adds “Attribute_9” to the structured presentation, the attribute identifier “Attribute_9” may be added to the user interface component 1000 with an associated action trigger 1030.

  FIG. 11 is a schematic diagram of a user interface component 1100 for receiving user input specifying a technique for adding a new instance to a structured presentation. For example, using the user interface component 1100 presents an option to add a new instance to the structured presentation at step 910 and receives an optional user selection at step 915 of the process 900 (FIG. 9). Can do.

  The user interface component 1100 includes a collection of technical descriptions for adding new instances to a header 1105, a prompt 1110, and structured presentations 1115, 1120, 1125 associated with selection widgets 1130, 1135, 1140, respectively.

  The header 1105 includes text or other information that identifies that user interaction with the user interface component 1100 actually allows the user to specify a technique for adding a new instance. Prompt 1110 prompts the user to specify techniques for interacting with user interface component 1100 to add a new instance.

  A description 1115 describes that a new instance is added by a user-specified constraint option as a result of the user specifying this technique. The user interaction of the selection widget 1130 allows the user to specify user-specified constraint options described by the description 1115.

  The description 1120 describes that a new instance is added by a user-specified constraint option as a result of the user specifying this technique. The description 1120 includes a constraint addition widget 1145 and a constraint clear widget 1150. User interaction with the add constraint widget 1145 triggers the addition of a new constraint used in the user specified constraint option. All current constraints are cleared by user interaction with the constraint clear widget 1150. The user interaction of the selection widget 1135 allows the user to specify user-specified constraint options described by the description 1120.

  Description 1125 describes that a new instance is added by a manual option as a result of the user specifying this technique. Description 1125 includes a new instance identifier input field 1155. User interaction with the new instance identifier input field 1155 allows the user to identify the new instance, eg, by name. The user interaction of the selection widget 1140 allows the user to specify the manual options described by the description 1125.

  FIG. 12 is a schematic diagram of a user interface component 1200 for receiving user input specifying constraints that are used with user-specified constraint options to add a new instance to a structured presentation. The user interface component 1200 can be used independently of other user interface components (eg, on a dedicated window or portal) or in conjunction with other user interface components. For example, the user interface component 1200 can be inserted into the user interface 1100 immediately below the technical description 1120 (FIG. 11). For example, the user interface component 1200 presents an option to specify the value of an attribute for a new instance that is added to the structured presentation at step 930, and such value of the attribute at step 935 of process 900 (Figure 9). Can be used to receive user specifications.

  The user interface component 1200 includes a collection of one or more attribute selection widgets 1205, each associated with a value specification area 1215, 1220. The attribute selection widgets 1205 and 1210 are interactive display elements that allow the user to select an attribute whose value is constrained. In the illustrated implementation, each attribute selection widget 1205, 1210 is a drop-down box widget that lists attribute identifiers. In some implementations, the listed attribute identifier may be the same as the attribute identifier 308 in the structured presentation to which the new instance is added.

  The value specification areas 1215 and 1220 are interactive display elements that allow the user to specify one or more constraints on the value of the attribute identified in each attribute selection widget 1205 and 1210. In the illustrated implementation, the value specification area 1215 includes a pair of text input fields 1225 that allow a user to specify an acceptable range of values for the attributes identified in the attribute selection widget 1205. The value specification area 1220 comprises a collection of interactive checkboxes 1230 that allow the user to specify acceptable values for the attributes identified in the attribute selection widget 1210.

  In operation, when a user selects a particular attribute identifier using the attribute selection widget 1205, 1210, a change in the associated value specification area 1215, 1220 may be triggered. For example, the nature of the interactive element and the values and / or ranges that can be specified in the associated value specification areas 1215, 1220 can be changed. In some implementations, these changes may be based on the distribution of the values of such attributes in the structured presentation where new instances are added. For example, if only four values of the attribute “maker” appear in the structured presentation, these same four values can be presented for specification in the associated value specification area. In other implementations, changes made to the associated value specification areas 1215, 1220 may be based on the values of attributes that characterize similar instances in the electronic document collection 102. For example, the attribute “maker” of the car instance can be characterized in documents in the electronic document collection 102 using more diverse values. These values can be identified and presented for specification within the associated value specification area.

  FIG. 13 is a flowchart of an exemplary process 1300 for adding a new attribute to a structured presentation. Process 1300 may be performed by one or more computers that perform operations by executing one or more machine-readable instruction sets. These digital data processing devices can exchange information with the user via input and output devices such as a keyboard, mouse, touch screen, display screen, and the like. For example, in the context of system 200 (FIG. 2), user interaction in process 1300 can be performed on the client side, such as PDA 215 or desktop computer 217.

  Process 1300 may be performed alone or in conjunction with other data processing activities. For example, as described in more detail below, process 1300 may be performed in conjunction with various processes to construct attribute proposals in addition to pre-existing structured presentations. Examples of such configuration processes are described in FIGS. 37-51 and related text, and in FIGS. 21-26 and related text. In general, process 1300 is performed by a plurality of digital data processing devices. For example, in the context of the system 200 (FIG. 2), activities for composing attribute proposals can be performed on the search engine 202 side, while user interaction is such as PDA 215 or desktop computer 217 (FIG. 2) Can be executed on the client side.

  The system executing process 1300 may receive a new attribute trigger (step 1305). A new attribute is an attribute that is not currently displayed in a structured presentation, eg, structured presentation 106 (FIG. 1). The new attribute trigger activates a process for adding new attributes to the structured presentation. For example, a new attribute can be triggered by user input received via a mouse, stylus, keyboard, or the like. In other implementations, the new attribute may be triggered by other processes or systems. New attribute triggers can be received through interprocess communication or application message handlers, to name a few. For example, in some implementations, the system can receive a new attribute trigger from the user interface component 1000 through a user selection of an add new attribute trigger 1035 (FIG. 10).

  The system executing process 1300 may present an option for specifying new attributes (step 1310). For example, the system provides a list of new attributes used to characterize instances in structured presentations, as well as interactive display elements that allow the user to select one or more of those attributes Can be displayed. In some implementations, the attributes in such a list can be configured based on attributes used to characterize instances elsewhere, such as in documents of an electronic document collection. An exemplary approach for configuring such attributes is described in FIGS. 37-51 and the associated text.

  The system executing process 1300 may receive a new attribute specification from the user (step 1315). An attribute specification can, for example, characterize the attributes or characteristics of the new attribute, including the name or other identifier of the new attribute, keywords associated with the new attribute, reliable sources of information about the new attribute, and the like . The attribute designation can be received from the user via one or more input devices, eg, a keyboard, touchpad, or touch screen.

  The system executing process 1300 can add the specified new attribute to the structured presentation (step 1320). For example, a system performing process 1300 can add a new attribute identifier 308 and column 304 to tables 300, 400 (FIGS. 3, 4). As another example, the system can add a new attribute identifier 308 in column 504 along with the corresponding attribute value 307 in column 506 of card collection 500 (FIG. 5). In some implementations, the system performing process 1300 may not only add new attributes to the structured presentation, but also add it to the user interface component to receive user input that specifies modifications to the structured presentation. it can. For example, the system can add a new attribute to the attribute modification area 1005 of the user interface component 1000 (FIG. 10).

  The system executing process 1300 may write the attribute value based at least in part on the user specification (step 1325). The system can write attribute values using a variety of techniques, as described in more detail below.

  FIG. 14 is a schematic diagram of a user interface component 1400 for adding new attributes to a structured presentation. The user interface component 1400 can interact with the user regarding the specification of one or more attributes or characteristics of the new attribute. These attributes or characteristics can be used, for example, when adding new attributes and attribute values to a structured presentation. For example, using the user interface component 1400 may present an option to add a new attribute class to the structured presentation at step 1310 and receive a user specification of the new attribute at step 1315 of the process 1300 (FIG. 13). it can.

  The user interface component 1400 includes a header 1405 and a collection of attribute identifiers 1410, 1415, 1420, 1425 that identify the characteristics that characterize the new attribute. Each characteristic identifier 1410, 1415, 1420, 1425 is associated with a characteristic specification widget 1430, 1435, 1440, 1445 and identifies a characteristic that can be specified by user interaction with that widget. The header 1405 includes text or other information that identifies that user interaction with the user interface component 1400 actually allows the user to add a new instance to the structured presentation.

  The attribute identifier 1410 identifies that the user can specify a class of attributes that are added to the structured presentation by interacting with the attribute specification widget 1430. The attribute class indicates how the attribute and its value are identified. For example, an attribute class can specify the technology used to identify an attribute and its value in an electronic document collection. Examples of attribute classes include “auto-find values”, “search results”, “review”, and “note” classes. Details regarding these attribute classes are further described below. The attribute specification widget 1430 is an interactive display element that allows the user to specify a class of attributes to be added to the structured presentation. In the illustrated implementation, the attribute specification widget 1430 is a drop-down box widget.

  The attribute identifier 1415 identifies that the user can specify the name of the new attribute or other identifier by interactively operating the attribute specification widget 1435. The attribute specification widget 1435 is an interactive display element that allows the user to specify the name or other identifier of a new attribute to be added to the structured presentation. In the illustrated implementation, the attribute specification widget 1435 includes a text input field. In general, the attribute identifier identified in characteristic identifier 1415 can be added directly in the structured presentation as attribute identifier 308.

  The characteristic identifier 1420 identifies that the user can specify a keyword that characterizes the new attribute by interactively operating the characteristic specification widget 1440. The feature specification widget 1440 is an interactive display element that allows the user to specify one or more keywords that characterize the attributes added to the structured presentation. In the illustrated implementation, the attribute specification widget 1440 includes a text entry field in which one or more keywords can be entered. The keyword may include, for example, an attribute identifier that characterizes the background of the attribute identifier or a synonym for the search term. For example, if the attribute identifier is “bank”, the keywords identified in the attribute designation widget 1440 include “NASCAR” and “speedway”, and the attribute points to “bank” of the racetrack as opposed to the financial institution. Can show that.

  In operation, the keywords specified in the attribute specification widget 1440 can be used to identify instances, attributes, and / or attribute values when searching an electronic document collection. For example, keywords can be used when constructing new attributes and / or new instances as described in FIGS. 21-26 and related text and FIGS. 37-51 and related text.

  The attribute identifier 1425 identifies that the user can specify “favorite sites” that characterize the new attribute by interactively operating the attribute specification widget 1445. “Favorite sites” are documents in the electronic document collection. User designation of a document as “favorite sites” indicates that the user considers the content of the document to be relevant to the new attribute and at the same time likely to be true. So, for example, when configuring new instances and attributes to add to pre-existing structured presentations (as described further below), assign a high confidence value to the content of “favorite sites” be able to. User designation of a document as “favorite sites” can also be used as an indicator that the content of the document is reliable with respect to attribute values that are written to the structured presentation.

  The property specification widget 1445 is an interactive display element that allows a user to specify one or more documents in the electronic document collection as “favorite sites”. In the illustrated implementation, the attribute specification widget 1445 includes a text input field that can enter, for example, one or more domain names or other locations of the electronic document.

  In some implementations, the feature “de-designate” widget allows the user to identify one or more documents in the electronic document collection as being a “disfavored” site. A user designation as a “disfavored site” for a document indicates that the user does not trust the document as a source of attribute values. Such a feature de-designation widget may include a text entry field where, for example, one or more domain names or other locations of the electronic document can be entered.

  FIG. 15 is a flow diagram of an example process 1500 for adding new attribute values to a structured presentation. Process 1500 may be performed by one or more computers that perform operations by executing one or more machine-readable instruction sets. Process 1500 may be performed alone or in conjunction with other data processing activities. For example, as described in more detail below, process 1500 may be performed in conjunction with various processes for adding new attributes to a structured presentation, eg, process 1300 (FIG. 13).

  The system executing process 1500 can receive a user designation of a class of new attributes (step 1505). As described above, the class of attribute indicates how the attribute and its value are identified. Receiving the new attribute class may be part of receiving the new attribute designation in step 1315 of process 1300 (FIG. 13). In some implementations, a user specification of a class of new attributes can be received via the attribute specification widget 1430 at the user interface component 1400 (FIG. 14).

  The system executing process 1500 can determine which class is designated for the new attribute (step 1510). Based on the specified class, the system executing process 1500 can determine which of the various sub-processes to add a new attribute value to the structured presentation is executed. For example, the system associates a subprocess associated with the “note” class, a subprocess associated with the “reviews” class, a subprocess associated with the “search results” class, or an “already found” class. It is possible to decide to add attribute values according to the subprocess being performed.

  If the system running process 1500 decides to add a new attribute value using the subprocess associated with the “note” class, the system may write the annotation received from the user to the attribute value. Yes (step 1515). For example, in the background situation of FIG. 4, the value in the annotation column 420 in the table 400 can be received from the user and used to write the value of the new attribute.

  If the system running process 1500 decides to add a new attribute value using the subprocess associated with the “reviews” class, the system will search and identify the electronic document that contains the review (Step 1520). A review is a rigorous evaluation of one or more instances that are characterized by new attributes. In some cases, the review can be created by someone with expertise to assess the instance, such as a critic. Reviews may be identified based on, for example, labels or other text that identifies those reviews as reviews. For example, review using a specific domain name (e.g. http://www.google.com/prdhp, http://www.epinions.com/, http://www.amazon.com/) The containing electronic document can be identified. Electronic documents that contain reviews can be found in an electronic document collection, eg, collection 102.

  The system performing process 1500 may write the attribute value using content from the identified review (step 1525). For example, the system can extract values from the review using one or more text or table-based extraction patterns and present those extracted values in a structured presentation. These extraction patterns can preferentially select segments of the review document that are “sentiment focused”. Emotion concentration segments are identified as strong emotions that speak positively or negatively on a particular subject. For example, restaurant reviews can include emotionally focused segments such as “the food is exceptionally good” and “the service was very poor indeed”. . Presentation of these extracted values in the structured presentation can be part of the writing of the structured presentation in step 1325 of process 1300 (FIG. 13).

  If the system executing process 1500 decides to add a new attribute value using the subprocess associated with the “search results” class, the system will retrieve the search results from the electronic document collection, eg, collection 102. Can be generated (step 1530). This search can generate a result set that is not limited to reviews, but rather can include a variety of electronic documents. Electronic documents can be found in an electronic document collection, eg, collection 102.

  The search results can be generated by a search based on the identifier of the new attribute, as well as the identifier of the instance characterized by that attribute. Some implementations use additional keywords associated with the new attribute to improve the search results, for example, keywords received from the user via the attributed widget 1440 of the user interface component 1400 (Figure 14). be able to.

  The system performing process 1500 can write the content from the search result set to attribute values in the structured presentation (step 1535). For example, the system may extract one or more values from a search result set using one or more text or table-based extraction patterns and present those extracted values in a structured presentation. it can. Writing the contents of the search result set to these attribute values can be part of writing the structured presentation in step 1325 of process 1300 (FIG. 13).

  If the system running process 1500 decides to add a new attribute value using a subprocess associated with the “already found” class, the system has already found an electronic document collection, for example, Values that are extracted from the electronic document collection 102 can be identified (step 1540). The “already found” value can be stored, for example, in a collection of information characterizing the electronic document, eg, data center 208 in system 200 (FIG. 2). In some implementations, such a collection of information can include historical records of previous structured presentations. The system performing process 1500 may write the already extracted value to the structured presentation attribute value (step 1545). Writing the contents of the search result set to these attribute values can be part of writing the structured presentation in step 1325 of process 1300 (FIG. 13).

  FIG. 16 is a flow diagram of an example process 1600 for adding a new attribute value to a structured presentation. In particular, the process 1600 relates to selecting attribute values to use when writing structured presentation attribute values. Process 1600 may be performed by one or more computers that perform operations by executing one or more machine-readable instruction sets. Process 1600 may be performed alone or in conjunction with other data processing activities. For example, process 1600 may be performed at step 1325 of process 1300 (FIG. 13), at step 1525 of process 1500 (FIG. 15), at step 1535 of process 1500 (FIG. 15), and / or at step of process 1500 (FIG. 15). May be implemented at 1545.

  The system performing process 1600 may identify candidate attribute values (step 1605). Candidate attribute values can be extracted, for example, directly from content (such as reviews or other documents in an electronic document collection) or identified from a collection of attribute values that have already been extracted. For example, in the context of FIG. 2, the system can access the data center 208 and extract one or more stored attribute values.

  The system performing process 1600 may determine the certainty in the identified candidate value (step 1610). The certainty in the candidate value should characterize the probability that the candidate value correctly characterizes the attributes of the instance. The certainty of value accuracy is based on, for example, the number of times the value is used to characterize an instance attribute, the quality of the document from which the value is used to characterize the instance attribute, and the like Can be determined.

  The system performing process 1600 may determine whether certain of the candidate values are low, medium, or high (step 1615). If the attribute value certainty is low, this indicates that the candidate value is unlikely to characterize the instance attribute correctly. If the attribute value confidence is high, this indicates that the candidate value is likely to characterize the instance attribute correctly.

  If the system executing process 1600 determines that some of the candidate values are highly certain, the system can write the extracted value to the attribute value in the structured presentation (step 1545). This can be done automatically, ie without user input.

  If the system performing process 1600 determines that some of the candidate values are medium certain, the system can provide the candidate values to the user (step 1625). For example, the system can generate a user interface component that presents candidate values in association with identifiers of instances and attributes potentially characterized by those candidate values.

  The system performing process 1600 may receive a user selection of some of the presented values (step 1630). User selections can be received as one or more user inputs. For example, a user interface component that presents candidate values can comprise one or more selection widgets that allow a user to select candidate values for writing to a structured presentation. This selection can be received from the user using a mouse, keyboard, or other user input device.

  The system executing process 1600 may write the selected value to the attribute value (step 1635). For example, a system performing process 1600 can present selected values in a structured presentation.

  In some implementations, the selected attribute value can be used to further improve the attributes, values, and / or instances presented in the structured presentation. For example, if a user specifies that the value of an instance attribute is thousands of dollars, the magnitude of that value can be used to exclude values of significantly different magnitude from the structured presentation. As another example, if a user specifies that an instance attribute has a value of thousands of dollars, the magnitude of that value can be used to exclude instances that have that attribute value that are significantly different in magnitude. it can.

  If the system running process 1600 determines that some of the candidate values are uncertain, the system running process 1600 may highlight such a lack in the structured presentation. Yes (step 1640). These deficiencies can be highlighted, for example, by leaving uncertain input items or highlighting low confidence values using color or other marks. The system may also be able to receive candidate values to correct those deficiencies from interactive elements, such as users interacting with text fields in uncertain input items or annotation cells adjacent to missing input items. .

  FIG. 17 is a schematic diagram of a user interface component 1700 for selecting candidate values to be added to a structured presentation. User interface component 1700 can interact with the user to select values that characterize the new attribute in the structured presentation. For example, the user interface component 1700 can be presented to the user at step 1625 and receive a user selection at step 1630 of the process 1600 (FIG. 16).

  User interface 1700 includes a header 1705 and a table 1710. The header 1705 includes text or other information that identifies that user interaction with the user interface component 1700 allows the user to select the value of the attribute of the instance for display in the structured presentation. Table 1710 includes a collection of candidate value information organized in columns 1715, 1720, 1725, as well as a collection of row selection widgets 1730.

  In particular, column 1715 also includes a collection of candidate value identifiers along with column header 1735. Candidate value identifiers may be extracted directly from documents in the electronic document collection 102 or indirectly through the data center 208. In some implementations, the value can also include a unit identifier 309 that specifies the unit of measurement for the particular value 307. Column header 1735 identifies that a candidate value identifier is found in column 1715.

  Column 1720 also includes a collection of confidence values along with column header 1740. The confidence value indicates the likelihood that the candidate value identified in column 1715 is correct. The confidence value can be represented by a numerical value or a word. For example, the confidence value can be expressed as a percentage as a probability, for example, whether the value is a correct value or on a numerical scale. Column header 1740 identifies that a confidence value is found in column 1720.

  Column 1725 also includes a collection of source identifiers along with column header 1745. The source identifier identifies one or more sources of candidate values identified in column 1715. The source can be identified using, for example, the title of the electronic document, the domain name, the author's name, or the like. In some implementations, the source identifier can include a text snippet that includes the candidate values identified in column 1715. Column header 1745 identifies that the source identifier is found in column 1720.

  The selection widget collection 1730 comprises one or more user interactive elements to receive input from the user. User input can identify that the candidate value identified in column 1715 is added to the structured presentation.

  In some implementations, the user interface component 1700 can present candidate values in an order based on confidence values. For example, the candidate value with the highest confidence value may be presented at the top of column 1715 and the candidate value with the lowest confidence value may be presented at the bottom of column 1715.

  In some implementations, the user interface component 1700 may also include a snippet of text around the attributes and values in the particular source identified in column 1725. By using such a snippet, the user can see the value in the background situation.

  FIG. 18 is a schematic representation of a structured presentation 1800 that includes a lack of attribute value highlight 1802 presented therein. In the example shown, the certainty in the candidate values for characterizing the attributes "ATTR_1" and "ATTRIBUTE_N" of the instance "INSTANCE_1" is the candidate value for characterizing the attribute "ATTR_2" of the instance "INSTANCE_2" Like the certainty in the low. For the attribute “ATTR_1” of instance “INSTANCE_1”, this lack of trust is highlighted by an empty cell 1804. In the case of the attribute “ATTRIBUTE_N” of the instance “INSTANCE_1” and the attribute “ATTR_2” of the instance “INSTANCE_2”, this lack of trust is highlighted by a color mark 1806. Such highlighting provides an intuitive form of feedback regarding the nature of a particular attribute value. That is, the user can browse the table 300 and immediately determine which value is suspected of being correct. The system may receive user input that corrects one or more of the highlighted deficiencies. For example, the system can receive manually entered attribute values, additional constraints, or other user input described in this designation that can be used by the system to reliably identify additional attribute values. .

  In some implementations, user interaction with a cell that has a deficiency highlighted can trigger a search that is aimed at correcting the deficiency. For example, user interaction with an empty cell 1804 may trigger a search. The search can use, for example, a customizable query based on the category of the instance in the display, the identifier of the instance characterized by the new value, and / or the identifier of the attribute characterized by the new value. After returning the set of search results, the system can further receive an interaction specifying a value that corrects the deficiency. In some implementations, the set of returned search results can include attribute-specific highlighting in text snippets that distinguish potential values.

  FIG. 19 is a schematic diagram of a user interface component 1900 for selecting candidate attributes to be added to a structured presentation. User interface component 1900 can interact with the user to select attributes that characterize the instance in the structured presentation. For example, the user interface component 1900 can be presented to the user so that attributes can be selected for addition to the structured display at step 1320 of the process 1300 (FIG. 13).

  User interface 1900 includes a header 1905 and a table 1910. The header 1905 includes text or other information that identifies that user interaction with the user interface component 1900 allows the user to select an instance attribute to display in the structured presentation. Table 1910 includes a collection of candidate attribute information organized in columns 1915, 1920, 1925, as well as a collection of row selection widgets 1930.

  In particular, column 1915 also includes a collection of candidate attribute identifiers along with column header 1935. Candidate attribute identifiers may be extracted directly from the documents in the electronic document collection 102 or indirectly through the data center 208. In some implementations, the attribute may also include a unit identifier 309 that specifies the unit of measure for the value of the candidate attribute to be converted. Column header 1935 identifies that a candidate attribute identifier is found in column 1915.

  Column 1920 also includes a collection of confidence values along with column header 1940. The confidence value indicates the likelihood that the candidate attribute identified in column 1915 is correct. The confidence value can be represented by a numerical value or a word. For example, the confidence value can be expressed as a percentage as a probability, for example, whether the attribute is the correct attribute or is on a numeric scale. Column header 1940 identifies that a confidence value is found in column 1920.

  Column 1925 also includes a collection of source identifiers along with column header 1945. The source identifier identifies one or more sources of candidate attributes identified in column 1915. The source can be identified using, for example, the title of the electronic document, the domain name, the author's name, or the like. In some implementations, the source identifier can include a text snippet that includes the candidate attributes identified in column 1915. Column header 1945 identifies that the source identifier is found in column 1920.

  The selection widget collection 1930 includes one or more user interactive elements to receive input from the user. The user input can identify that the candidate attribute identified in column 1915 is added to the structured presentation.

  In some implementations, the user interface component 1900 can present candidate attributes in an order based on confidence values. For example, the candidate attribute with the highest confidence value may be presented at the top of column 1915 and the candidate attribute with the lowest confidence value may be presented at the bottom of column 1915.

  In some implementations, the user interface component 1900 may also include text snippets around instances and attributes in the particular source identified in column 1925. By using such a snippet, the user can see the attributes in the background situation.

  FIG. 20 is a schematic diagram of a user interface component 2000 for selecting candidate instances to be added to a structured presentation. User interface component 2000 can interact with the user to select instances to add to the structured presentation. For example, the user interface component 2000 can be presented to the user so that an instance can be selected for addition to the structured display at steps 925, 940 of the process 900 (FIG. 9).

  The user interface 2000 includes a header 2005 and a table 2010. The header 2005 includes text or other information identifying that user interaction with the user interface component 2000 allows the user to select an instance for display in the structured presentation. Table 2010 includes a collection of candidate instance information organized in columns 2015, 2020, and 2025, as well as a collection of row selection widgets 2030.

  In particular, column 2015 also includes a collection of candidate instance identifiers along with column header 2035. Candidate instance identifiers may be extracted directly from documents in the electronic document collection 102 or indirectly through the data center 208. Column header 2035 identifies that a candidate instance identifier is found in column 2015.

  Column 2020 also includes a collection of confidence values along with column header 2040. The confidence value indicates the likelihood that the candidate instance identified in column 2015 will be added. The confidence value can be represented by a numerical value or a word. For example, the confidence value can represent, as a percentage, the probability that an instance meets a user-specified constraint, for example. Column header 2040 identifies that a confidence value is found in column 2020.

  Column 2025 also includes a collection of source identifiers along with column header 2045. The source identifier identifies one or more sources of candidate instances identified in column 2015. The source can be identified using, for example, the title of the electronic document, the domain name, the author's name, or the like. In some implementations, the source identifier may include a text snippet that includes identifiers of candidate instances in column 2015. Column header 2045 identifies that the source identifier is found in column 2020.

  The selection widget collection 2030 includes one or more user interactive elements to receive input from the user. The user input can identify that the candidate instance identified in column 2015 is added to the structured presentation.

  In some implementations, the user interface component 2000 can present candidate instances in an order based on a confidence value. For example, the candidate instance with the highest confidence value may be presented at the top of column 2015 and the candidate instance with the lowest confidence value may be presented at the bottom of column 2015.

  In some implementations, the user interface component 2000 may also include a snippet of text around the instance identifier in the particular source identified in column 2025. By using such a snippet, the user can see an instance in the background situation.

  Changes made to structured presentations using the systems and processes described herein are an iterative process in which those changes are used to identify additional instances, attributes, and / or values. Can be part. For example, the process 800 (FIG. 8) can be repeated several times. As the range of existing content increases, additional instances, attributes, and / or values that are identified can have increased certainty.

  FIG. 21 is a schematic diagram of a process 2100 in which a new instance can be added to extend a pre-existing structured presentation. Process 2100 may be performed by one or more computer systems, eg, system 200 (FIG. 2), that perform operations with one or more machine-readable instruction sets.

  Process 2100 includes an extract operation 2105 and a merge operation 2110 that add a new instance to a pre-existing structured presentation based on information derived from documents in electronic document collection 102. In particular, the process 2100 proposes one or more new instances based on information presented in a pre-existing structured presentation 106. For example, if the structured presentation includes multiple instances corresponding to several videos, the system 200 may suggest additional instances of the videos according to information derived from the electronic document collection. That is, the system 200 can identify and suggest additional instances according to the similarity of the attribute identifier, the unit of measurement of the attribute value, the value of the attribute value, or a combination thereof. For example, the system 200 can suggest a video with similar showtime, theater, or airtime.

  As shown in FIG. 21, the extraction operation 2105 extracts a collection of new instance proposals from the electronic document collection 102 using pre-existing characteristics of the structured presentation 106. Examples of properties include instances in a pre-existing structured presentation, attributes in a pre-existing structured presentation, and values of attributes in a pre-existing structured presentation. The pre-existing characteristics of the structured presentation 106 can be represented as a collection of machine-readable information that can be represented by one or more computer systems that perform operations with one or more machine-readable instruction sets. Can receive. For example, search engine 202 (FIG. 2) can receive pre-existing characteristics of structured presentation 106.

  In an extraction operation 2105, one or more new instance proposals can be constructed based on the content of the documents in the electronic document collection 102 and the characteristics of the pre-existing structured presentation 106. As described further below, a variety of different techniques for composing a new instance proposal may be used.

  A merge operation 2110 may merge some or all of the new instance proposals with the pre-existing structured presentation 106 to form an extended structured presentation 106. For the viewer, the expanded structured presentation can be displayed on a display device, such as the display screen 104, for example.

  All new instance proposals configured in the extract operation 2105 are merged with the pre-existing structured presentation 106 and need not be displayed to the viewer. For example, in some implementations, a collection of new instance proposals can be presented to the viewer with interactive elements that allow the viewer to select one or more instances to be added. However, in other implementations, new instance proposals can be added automatically without user interaction and without screening new instance proposals prior to display. Further details regarding mergers can be found, for example, in FIGS. 9-20 and associated text.

  FIG. 22 is a flow diagram of an example process 2200 for adding an instance to a structured presentation based on the content of documents in an electronic document collection. Process 2200 may be performed by a system of one or more computers that perform operations with one or more machine-readable instruction sets. For example, process 2200 may be performed by search engine 202 within system 200.

  In some implementations, process 2200 may be performed in response to receiving input from a user or other system or process that triggers an update of structured presentation 106, for example. For example, process 2200 can be performed in response to a user request to add one or more new instances to structured presentation 106. As another example, process 2200 may be performed by a search engine, eg, search engine 202 (FIG. 2) in response to receiving a search query.

  The system performing process 2200 may receive one or more characteristics of a pre-existing structured representation (step 2205). For example, the system can receive one or more attribute identifiers of a pre-existing structured presentation. As another example, the system can receive one or more instance identifiers that appear in a pre-existing structured presentation.

  The system performing process 2200 may construct one or more instance proposals from documents in the electronic document collection based on one or more characteristics of a pre-existing structured presentation (step 2210). ). Instance proposals can be constructed based on these characteristics in a number of different ways. For example, in one implementation, the system may construct an instance proposal from documents in the electronic document collection 102 by constructing a search query using attribute identifiers derived from pre-existing structured presentations. it can. These search queries can be used to identify instances that can share similar attributes using string comparison results or other matching techniques. Examples of other approaches are further described below.

  The system executing process 2200 may provide one or more instance suggestions to the user (step 2215). For example, a list of instance proposals can be displayed for the user on the same display screen that displays a pre-existing structured presentation.

  The system performing process 2200 may receive a user selection of one or more instance proposals (step 2220). For example, a user interface component interacts with the user and receives one or more user inputs (e.g. mouse clicks, key presses, or other user inputs) that select one or more instance suggestions. it can.

  The system executing process 2200 may add the selected instance proposal as a new structured record to the structured presentation (step 2225). For example, if the structured presentation is a table such as table 300 (FIG. 3), the system can add a new row 302. As another example, if the structured presentation is a collection of cards, such as a collection of cards 500 (FIG. 5), the system can add a new card 502.

  FIG. 23 is a flow diagram of an example process 2300 for constructing an instance proposal from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. Process 2300 may be performed alone or in conjunction with other activities. For example, process 2300 may be performed at step 2210 in process 2200 (FIG. 22).

  Process 2300 may be performed by a system of one or more computers that perform operations with one or more machine-readable instruction sets. For example, process 2300 may be performed by search engine 202 in system 200 (FIG. 2).

  The system performing process 2300 may identify one or more documents that include a structured component related to an instance specified in a pre-existing structured presentation (step 2305). A structured component is a number of parts or regions of an electronic document that are structured. Examples of structured components include tables, lists, records, collections of attribute-value pairs, and the like. Thus, structured components are similar to structured presentations, with predefined structures and Attribute values and instance identifiers can be organized to match.

  The entire electronic document including the structured component need not be structured. For example, an electronic document may have a table sandwiched between two paragraphs of unstructured text. Furthermore, structured components in different documents need not have the same format or follow a predetermined or permanent structure. In fact, the organization of information within one structured component can generally be changed independently of the organization of information within the structured component that appears in other documents. For example, if a structured list of schools in one person's resume is changed to remove the graduation year, the other structured lists of schools in other resumes will be changed as well There is no need to guarantee that

  A system performing process 2300 may identify documents that include structured components in a variety of ways. For example, tables and other structured components can be identified using metadata labels, eg, HTML tags, found in the document itself. As another example, a structured component can be identified by identifying repetitive elements (eg, a series of commas or tab delimiters) in the document.

  Structured components are related to the specified instance in a pre-existing structured presentation if they contain information related to the specified instance. For example, a structured component that characterizes one or more of the specified instances with one or more attribute values is related to the specified instance in a pre-existing structured presentation Can be thought of as As another example, a structured component that characterizes one or more of the same attributes of an instance that is different from the instance specified in the pre-existing structured presentation is related to the specified instance Can be thought of as In many implementations, the instance and / or attribute identifiers need not be the same. Rather, conceptually related instances and attributes can be used to identify documents that contain structured components.

  Thus, in some implementations, the system performing process 2300 may be in the same or related instance identifier and / or preexisting structured presentation as found in the preexisting structured presentation. Identify one or more documents that contain structured components related to instances specified in pre-existing structured presentations by identifying documents that contain the same or related attribute identifiers as found Can do.

  The system performing process 2300 may select one or more instance proposals from the structured component (step 2310). This selection process can reduce the number of instances proposed to the user. The selection of instance proposals can be performed in various ways. For example, the system may make an instance suggestion based on a category of values of instances in a structured component, attributes of instances in the structured component, and / or attributes of instances in the structured component, as further described below. You can choose.

  FIG. 24 is a representation 2400 that represents the configuration of an instance proposal from an electronic document in an electronic document collection based on pre-existing structured presentation characteristics. In particular, representation 2400 illustrates the configuration of an instance proposal that uses one implementation of process 2300 (FIG. 23).

  As shown, the pre-existing structured presentation specifies a collection of instances 2405 (ie, instances “Philadelphia” and “Chicago”). Further, different documents in an electronic document collection (eg, collection 102) include different structured components 2410, 2415, 2420. The structured components 2410, 2415, 2420 may be identified as being associated with the instance 2405 designated based on, for example, the same instance identifiers “Philadelphia” and “Chicago” appearing therein.

  As shown, the structured components 2410, 2415, 2420 include a variety of different potential instance proposals based on different background situations. In particular, in the context of structured component 2410, instances "Philadelphia" and "Chicago" are part of a tabular component that represents various city characteristics. In the context of the structured component 2415, the instances “Philadelphia” and “Chicago” are part of the structured component that represents a portion of the standings of the National League East District at some point in the 1970s. In the context of the structured component 2420, the instances “Philadelphia” and “Chicago” are part of a tabular component that represents various animation characteristics.

  Instead of suggesting all the various instances found in structured components 2410, 2415, 2420 to the user, select an instance selection from components 2410, 2415, 2420 based on the attributes used to characterize those instances can do. In particular, as shown, the pre-existing structured presentation 106 uses the values of the attributes “year”, “rating”, and “box office receipts” to create instances “Philadelphia” and “Chicago”. Characterize. The structured component 2410 characterizes the instances “Philadelphia” and “Chicago” using the values of the attributes “population” and “area”. The structured component 2415 characterizes the instances “Philadelphia” and “Chicago” using the values of the attributes “wins”, “losses”, and “GB” (ie, game differences). The structured component 2420 characterizes the instances “Philadelphia” and “Chicago” using the values of the attributes “year”, “runtime”, and “rating”.

  The system can select from instances within the structured components 2410, 2415, 2420 based on these characterized attributes. For example, the system identifies the correspondence between attribute identifiers “year” and “rating” in pre-existing structured presentation 106 and attribute identifiers “year” and “rating” in structured component 2420. , Instances “Peter Pan” and “Star Wars” can be selected as suggestions to add to the pre-existing structured presentation 106.

  As described in FIGS. 37-51 and the associated text, the system may also suggest or add additional attribute identifiers. For example, the structured component 2420 includes an attribute identifier “runtime”. Thus, in such a system, the attribute identifier “runtime” can be proposed with or without a corresponding attribute value.

  In some implementations, even if instances derived from structured components 2410, 2415 are not proposed in a particular configuration, such instances are stored for use in future information requests. can do. For example, even if the cities represented by the structured component 2410 have not been selected as instance proposals, those cities will have their respective attribute identifiers (eg, “population” and “area”) and data collections (eg, data Can be stored along with attribute values in the center 208). If the user subsequently requests information about one or more cities, such a system can access this stored information and provide additional information to the user.

  FIG. 25 is a flow diagram of an example process 2500 for constructing an instance proposal from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. Process 2500 may be performed alone or in conjunction with other activities. For example, process 2500 may be performed at step 2210 in process 2200 (FIG. 22).

  Process 2500 may be performed by a system of one or more computers that perform operations with one or more machine-readable instruction sets. For example, process 2500 may be performed by search engine 202 in system 200 (FIG. 2).

  The system performing process 2500 may identify one or more documents associated with one or more specified instances (step 2500). For example, a system executing process 2500 uses a string comparison to document one or more of the specified instance and its attributes and / or values and an electronic document collection such as electronic document collection 102. Can be matched. In other examples, the system executing process 2500 can access stored information (such as information in data center 208) to identify the electronic document associated with the specified instance.

  The system performing process 2500 may extract a template for one or more of the identified documents (step 2510). The document template is used as a pattern for the arrangement of the content of individual documents in a sub-collection of documents in the electronic document collection. The documents in a subcollection are generally from a single source, such as a single company. For example, a book store can use a single document template as a pattern for an array of content describing different books. In another example, a furniture retailer can use a single document template as a pattern for an array of different sofa brochure content. For example, a template for an e-brochure for a sofa allows the user to select the sofa brand name, sofa photo, sofa color, textual sofa description, and sofa An array of tables characterizing dimensions, availability, and prices can be specified on the brochure. In this way, the document template can organize information about the instances in much the same way as a structured presentation and to match a predefined structure.

  In general, a document template can be used as a pattern for the entire content of an electronic document, and can even specify an array of structured components within the document, as described above. However, since the document template only specifies the content array of the sub-collection of documents in the unstructured electronic document collection, the electronic document collection itself remains unstructured. For example, even if AMAZON.COM uses one template as a pattern for an array of descriptions for all books sold by AMAZON.COM, BARNESANDNOBLE.COM and other book stores will use the same template as the book. It is not necessarily used as a pattern for the description arrangement of books sold by the store. Furthermore, document templates can be changed without necessarily propagating the changes to the entire collection of electronic documents, and even to a sub-collection.

  FIG. 26 is a diagram showing a part of a hypertext markup language (HTML) template 2600 used as a pattern for describing a moving image (that is, moving image “Philadelphia”). The hypertext markup language (HTML) code of the template portion 2600 is machine readable and human readable. For example, the web page can be generated using the HTML code of the template portion 2600 in the browser.

  In the example shown, the template portion 2600 is divided into two subsections 2605, 2610. Subsection 2605 is used as a pattern for an array of text that identifies the video “Philadelphia”. Subsection 2610 is used as a pattern for various attribute identifiers and their value arrays. In general, the patterns in subsections 2605, 2610 are repeated many times in a particular sub-collection of documents within an electronic document to describe different videos.

  An HTML parser can be used to extract the output format from the template portion 2600, and thus the output format can be used to identify documents having the same template. For example, HTML tags <title>, <div>, other HTML tags, and positions relative to each other can be identified by an HTML parser. Such an HTML parser can determine that the HTML tag <title> appears before the HTML tag <div>. Thus, the HTML parser can extract the output format from the template portion 2600 from the content arranged according to the template.

  Returning to FIG. 25, after extracting the template, the system performing process 2500 may identify one or more documents having the same template (step 2515). For example, the system can compare a template of a document in the electronic document collection with the extracted template.

  The system performing process 2500 may also construct one or more instance proposals from documents that are identified as having the same template (step 2520). In particular, the system can use template iteration within a sub-collection of documents to infer that the documents in the sub-collection contain the same type of content for the same category of instances. In other words, the system can infer that the background status of two documents is the same because the same template is used as a pattern for different documents.

  After documents with similar background status are identified, the template itself can be used to construct an instance proposal. For example, the HTML tag in the template portion 2600 (FIG. 26) identifies that the movie title described in the document is “Philadelphia (1993)”. By searching for similarly tagged text in documents that share the same template, the system can identify titles for other movies.

  Further, in some implementations, additional content in the document template can be used in constructing the instance proposal. For example, identifying a particular value (eg, George Lucas) as a “movie director” can be used to select a particular instance proposal from a sub-collection of documents. In other words, the subsection 2610 of the template portion 2600 can be parsed or analyzed in some other manner to determine whether an attribute has a similar value, identifier, or other characteristic. In such a situation, the instance identifier may be extracted from subsection 2605.

  FIG. 27 is a schematic diagram of a process 2700 that can construct a collection of new instance proposals 2115 based on pre-existing information in the structured presentation 106. Process 2700 may be performed by a system of one or more computers that perform operations with one or more machine-readable instruction sets (FIG. 2).

  The process 2700 performs an extraction operation 2705 on the instance / attribute collection 2710 based on information in the pre-existing structured presentation 106. An instance / attribute collection 2710 is a collection of information that associates an instance with its attributes and, in some implementations, with the values of those attributes. The information in collection 2710 can be used as a response to receiving a trigger (e.g., a search query) or as part of the process of indexing electronic document collection 102 in anticipation of receiving a trigger, e.g. It can be extracted from the documents in the document collection 102. In some implementations, the information in collection 2710 can include the content of previous structured presentations presented to the current user or other users. In general, the instance proposal is provided to a user who selects an instance proposal to be added to the structured presentation, eg, as described in steps 2215, 2220, 2225 (FIG. 22).

  An association between an instance and its attributes can be established within collection 2710 by structuring information storage within collection 2710. FIG. 28 is a schematic diagram of a table 2800 that associates attributes and instances in collection 2710. Table 2800 contains a collection of records 2802, 2804, 2806, 2808, 2810, 2812, 2814, each record with an instance identifier as a description of the document location and attributes that characterize the identified instances within those documents. Associate. Information in records 2802, 2804, 2806, 2808, 2810, 2812, 2814 may be organized in a collection of columns 2815, 2820, 2825, 2830, 2835, 2840. In particular, column 2815 can include an instance identifier. Column 2820 may include a description of the location of the electronic document that includes the instance identified in column 2815. Columns 2825, 2830, 2835, 2840 may identify attributes that characterize the instance identified in column 2815 in the document whose location is described in column 2820.

  As shown, different electronic documents can include different categories and quantities of information that characterize the same instance. For example, a document whose location is identified in column 2820 of record 2804 contains two attributes for instance "INSTANCE_2", whereas a document whose location is identified in column 2820 of record 2810 is three attributes of instance "INSTANCE_2" including. Furthermore, the attributes in record 2804 (ie, attribute “ATTR_5” and attribute “ATTR_6”) are partially different from the attributes in record 2810 (ie, attribute “ATTR_5”, attribute “ATTR_8”, and attribute “ATTR_9”). .

  A data collection 2710 (such as table 280) that associates attributes and instances can be formed in a number of different ways. For example, a document that includes an internal structured component may be identified. Examples of such internal structured components are tables and lists that appear in HTML documents. The relationship between attributes and instances in these internal structured components can be copied to form a data collection 2710.

  As another example, collection 2710 can be formed from the content of previous structured presentations presented to the current user or other users.

  As yet another example, after a first document is identified as containing attributes and instances, the template for that document can be used to extract attributes and instances from other documents that contain the same template. For example, if a stereo retailer describes different stereos offered for sale using the same document template, it uses the arrangement of information in the first electronic document for the first stereo to Information can be extracted from other electronic documents.

  In some implementations, techniques such as natural language parsing can be used to identify instances and attributes. For example, the electronic document can be parsed to identify phrases such as “[Instance] has a / an [attribute]” in the electronic document.

  In some implementations, the data collection 2710 can classify instances and their attributes. For example, instances and attributes can be categorized as North American cities, National League East District teams, or popular movies. Storage of information in the data collection 2710 can be based on such classification. For example, different categories can be stored in different files, records, and the like.

  Referring again to FIG. 27, it can be seen that the process 2700 proposes one or more new instances based on information presented in the pre-existing structured presentation 106. For example, if the structured presentation includes multiple instances corresponding to several videos, the system 200 may suggest additional instances of the videos according to information derived from the data collection 2710. That is, the system 200 can identify and suggest additional instances according to the similarity of attribute identifiers. For example, the system 200 can suggest a video with similar showtime, theater, or airtime.

  FIG. 29 is a flow diagram of a process 2900 for constructing an instance proposal from a collection of instances and attributes based on pre-existing structured presentation characteristics. Process 2900 may be performed by a system of one or more computers that perform operations with one or more machine-readable instruction sets. For example, process 2900 may be performed by search engine 202 in system 200 (FIG. 2).

  Process 2900 may be performed alone or in conjunction with other activities. For example, process 2900 may be performed in process 2700 (FIG. 27). As another example, process 2900 may be performed at step 2210 in process 2200 (FIG. 22) alone or in conjunction with one or both of processes 2300, 2500 (FIGS. 23, 25).

  The system executing process 2900 can access the collection of instances and their attributes (step 2905). For example, a system executing process 2900 can access an instance and attribute collection 2710 (FIG. 27) stored in data center 208 (FIG. 2).

  The system performing process 2900 may identify one or more related instances based on the properties of the instance attributes specified in the pre-existing structured presentation (step 2910). For example, the system can compare the instance attributes of the instances stored in the instance and attribute collection 2710 (FIG. 27) with the instances specified in the structured presentation 106. The system uses this comparison result to share any of the stored instances, if any, with the attribute identifier, or the attribute specified in the structured presentation 106 that already has the attribute concerned. You can decide what to do. For example, pre-existing structured presentation 106 presumes to use the attributes “ATTR_3” and “ATTR_5” to characterize the collection of instances. After reviewing the instance and attribute collection 2710, such as table 2800 (Figure 28), the system characterizes the instances "INSTANCE_1" and "INSTANCE_2" and the same attributes "ATTR_3" and "ATTR_5" in records 2802, 2806 Can be proposed based on what is used.

  FIG. 30 is a flow diagram of a process 3000 for constructing a collection of new instance proposals 2115 based on pre-existing information in the structured presentation 106. Process 3000 may be performed by one or more computer systems that perform operations according to one or more machine-readable instruction sets, eg, system 200 (FIG. 2). Process 3000 may be performed alone or in conjunction with other activities. For example, process 3000 may be performed at process 2700 (FIG. 27). As another example, process 3000 is performed at step 2210 in process 2200 (FIG. 22) alone or in conjunction with one or more of processes 2300, 2500, 2900 (FIGS. 23, 25, 29). Can be done. In some implementations, process 3000 may be performed in response to receiving input (eg, from a user of system 200 or from another system or process). For example, process 3000 may be performed by search engine 202 in response to receiving a search string.

  The system performing process 3000 may identify one or more trusted sources for one or more specified instances (step 3005). For example, the system can access a collection of trusted sources of documents in the electronic document collection 102 assembled, for example, by a programmer.

  As another example, the system may receive user-specific input that identifies one or more trusted sources of documents in the electronic document collection 102 as “trusted” from the user's perspective. For example, the display screen 104 displaying the pre-existing structured presentation 106 can include a GUI component that allows the viewer to specify a trusted source of the document. The identity of the trusted source can be received in conjunction with receiving a search query. For example, viewers can use JD POWER AND ASSOCIATES, AMAZON.COM, and MAJOR LEAGUE BASEBALL at http://www.jdpower.com/, http://www.amazon.com/, and http: // www. Can be identified as a trusted source of documents at mlb.com/. In some implementations, user-specific input can identify the subject matter for which the source is reliable. For example, MAJOR LEAGUE BASEBALL can be identified as a reliable source of baseball statistics, but cannot be considered as a reliable source of information about drug tests.

  As yet another example, a system performing process 3000 can analyze a collection of user-specific inputs that identify trusted sources from multiple users to assemble a “generic” collection of trusted sources. . For example, many users can identify AMERICAN AUTOMOBILE ASSOCIATION (AAA) as trusted. AAA can then be added to the collection of trusted sources based on statistical analysis of these identification results.

  The system performing process 3000 may determine additional attributes from a trusted source for the instance specified in the pre-existing structured presentation (step 3010). For example, the system can access documents provided by a trusted source and identify one or more documents that characterize a specified instance using one or more attributes. The system can extract attribute identifiers from these documents using parsers or other string comparison techniques.

  As another example, the system can access a data collection that associates attributes and instances, such as table 2800 (FIG. 28). The system records 2802, 2804, 2806, 2808, 2810, 2812, 2814, etc., and records 2802, 2804, 2806, 2808, 2810, instances and locations identified in pre-existing structured presentations. It can be filtered based on both whether the documents identified by 2812, 2814 were from a trusted source. For example, if AMAZON.COM is a trusted source, a collection that associates attributes and instances can be scanned to identify documents that have the http://www.amazon.com/ domain.

  The system executing process 3000 may compare these additional attributes to instances such as table 2800 (FIG. 28) and attributes in the attribute collection (step 3015). For example, the system can use string comparison, or other comparison techniques, to compare additional attributes with attributes stored in instances and attribute collections.

  The system executing process 3000 may identify instances and instances in the attribute collection based on these comparison results (step 3020). For example, the system can determine the number of attributes used to characterize an instance in a document from a trusted source and the attributes associated with the instance and other instances in the attribute collection.

  FIG. 31 is a flow diagram of a process 3100 for constructing a collection of new instance proposals 2115 based on pre-existing information in the structured presentation 106. Process 3100 may be performed by one or more computer systems, eg, system 200 (FIG. 2), that perform operations with one or more machine-readable instruction sets. Process 3100 may be performed alone or in conjunction with other activities. For example, process 3100 may be performed in process 2700 (FIG. 27). As another example, process 3100 may be performed alone or with one or more of processes 2300, 2500, 2900, 3000 (FIGS. 23, 25, 29, 30) at step 2210 in process 2200 (FIG. 22). Can be executed in conjunction. In some implementations, process 3100 may be performed in response to receiving input (eg, from a user of system 200 or from another system or process). For example, process 3100 may be performed by search engine 202 in response to receiving a search string.

  The system executing process 3100 may identify one or more related instances based on the instance attribute values specified in the pre-existing structured presentation. For example, the system can identify the associated instance by comparing the attribute value of the specified instance with the attribute value of the other instance and determine whether the other instance is a suitable proposal. Such a comparison may require, for example, that the attribute values are the same or that the attribute values are within a certain range. Such ranges can be determined based on attribute value ranges that have already been characterized, for example, by instances specified in the user via interactive elements or pre-existing in the structured presentation. it can.

  In some implementations, the system performing process 3100 can convert attribute values to normal units of measure before comparing the attribute values. For example, if the specified unit of measure is feet, but the unit of measure for one or more extracted attribute values is meters, the system converts the metric value to feet using normal techniques. can do. A schematic representation of one such approach is detailed below.

  FIG. 32 is a schematic diagram of a table 3200 that associates attributes, instances, and their values in a data collection. Because table 3200 associates attributes and instances, table 3200 can also be used as instance / attribute collection 2710 (FIG. 27). Table 3200 may be generated based on information derived from a collection of electronic documents, eg, electronic document collection 102. The table 3200 can be generated, for example, during the execution of the crawling process and stored in the data center 208 for later use, for example.

  The table 3200 includes a collection of records 3202, 3204, 3206, 3208, 3210, 3212, 3214, each record having an instance identifier, a description of the document's location, attributes that characterize the identified instances in those documents, And associated with values that characterize those attributes in those documents. Information in records 3202, 3204, 3206, 3208, 3210, 3212, 3214 may be organized in a collection of columns 3215, 3220, 3225, 3230, 3235, 3240. In particular, column 3215 can include an instance identifier. Column 3220 may include a description of the location of the electronic document that includes the instance identified in column 3215. Columns 3225, 3235 may identify attributes that characterize the instance identified in column 3215 in the document whose location is described in column 3220. Columns 3230, 3240 can include values that characterize the attributes shown in columns 3225, 3235.

  In the example shown, each record 3202, 3204, 3206, 3208, 3210, 3212, 3214 is associated with a different instance (eg, INSTANCE_10 through INSTANCE_N). Each of these instances is characterized in at least one document by attribute identifiers ATTR_3, ATTR_6. As such, if the instance proposal was configured based solely on attributes that could be used to characterize INSTANCE_10 through INSTANCE_N, then all INSTANCE_10 through INSTANCE_N could be proposed to the user.

  In many situations this is unacceptable. For example, many of the same attributes of all vocational schools or universities around the world (eg, student count, student / teacher ratio, location, etc.) may be characterized in any electronic document available on the Internet. However, a list of proposals that includes all vocational schools or universities does not help students looking for a school to attend.

  Thus, as described above for process 3100 (Figure 31), the associated instance compares the attribute value of the instance specified in the pre-existing structured presentation with the attribute value of the other instance. Can be identified. For example, the specified instance in the structured presentation characterizes the instance attribute “ATTR_3” with the unit “unit_a” value “VAL_3” and the instance attribute “ATTR_6” with the unit “unit_c” value “VAL_6” , Systems such as search engine 202 identify that the instances identified in records 3202, 3206 (i.e., `` INSTANCE_10 '' and `` INSTANCE_12 '') can be proposed to the user based on a common value (even if the units are different) can do. Thus, the system can convert the values in cells 3245, 3250, and 3255, 3260 into a common unit of measure and compare the values to determine if the values are similar. In this way, even if the units representing these values are different, similar instances can be selected.

  In addition, the instance identified in record 3208 (that is, INSTANCE_13) shares the common value of attribute “ATTR_3” with the specified instance in the structured presentation, but the instance indicated in record 3208 is It does not have to be proposed. In particular, the value characterizing the attribute “ATTR_6” of this instance is the value “VAL_8”, which is different from the value characterizing this attribute of the specified instance in the structured presentation. Based on this difference, the instance identified in record 3208 can be excluded from the list of suggested instances.

  Different criteria can be used to put and exclude instances in the list of proposed instances. For example, the number of attribute values that must be similar may be different. As another example, in some implementations, a user can specify the number and / or nature of attribute values that are considered when constructing a list of suggested instances. In yet another example, instance ranking is performed based on the correspondence between these attribute values and the attribute values of one or more specified instances in a pre-existing structured presentation. it can. As yet another example, a range of values can be determined based on values that characterize the attributes of one or more instances specified in the structured presentation, and can be proposed using this range. Related instances to be included in the list of instances to be identified.

  As described in FIGS. 9-20 and the associated text, in some implementations, the system selects a value from a collection of different values based on criteria that reflect the probability that the value is appropriate. Can do. Examples of such are user-specified ranges, the number of documents that characterize attributes with specific values, and / or the quality of documents that characterize attributes with specific values.

  FIG. 33 is a flow diagram of a process 3300 for constructing a collection of new instance proposals 2115 based on pre-existing information in the structured presentation 106. Process 3300 may be performed by a system of one or more computers that performs operations with one or more machine-readable instruction sets, eg, system 200 (FIG. 2). Process 3300 may be performed alone or in conjunction with other activities. For example, process 3300 may be performed at process 2700 (FIG. 27). As another example, process 3300 is a step 2210 in process 2200 (FIG. 22), alone or one of processes 2300, 2500, 2900, 3000, 3100 (FIGS. 23, 25, 29, 30, 31). Can be performed in conjunction with one or more. In some implementations, process 3300 may be performed in response to receiving input (eg, from a user of system 200 or from another system or process). For example, process 3300 may be performed by search engine 202 in response to receiving a search string.

  The system executing process 3300 may access a categorized collection of instances and attributes (step 3305). For example, the system can access the instance and attribute collection 2710 to access one or more categorized collections of instances and attributes generated during previous searches.

  The system executing process 3300 may identify the category that includes the specified instance (step 3310). In some implementations, the system may identify categories that include instances based on similar attributes, similar attribute values, combinations of these characteristics, and / or other technologies.

  The system performing process 3300 may select one or more instance proposals from the identified category (step 3315). For example, in some implementations, an instance proposal can be selected from an identified category based on the similarity between the attribute value of the specified instance and the attribute value of the instances in that category.

  FIG. 34 is a representation 3400 representing the configuration of an instance proposal from an electronic document in an electronic document collection based on pre-existing structured presentation characteristics. In particular, representation 3400 shows the configuration of an instance proposal that uses one implementation of process 3300 (FIG. 33).

  As shown, the pre-existing structured presentation specifies a collection of instances 2405 (ie, instances “Philadelphia” and “Chicago”). Further, instances derived from different documents in an electronic document collection (eg, collection 102) have already been categorized into different collections 3410, 3415, 3420. The categorized instance collections 3410, 3415, 3420 can be identified as being associated with the instance 2405 designated based on, for example, the same instance identifiers “Philadelphia” and “Chicago” appearing therein.

  As shown, the categorized instance collections 3410, 3415, 3420 are categorized in a variety of different ways. In particular, the categorized instance collection 3410 is categorized as a collection of “North American Cities”. Categorized instance collection 3415 is categorized as a collection of “National League East Teams”. The categorized instance collection 3420 is categorized as a collection of “Popular Movies”. In the context of the system 200, the categorized instance collections 3410, 3415, 3420 can be stored in the data center 208 (FIG. 2). That is, system 200 can generate one or more categories of instances based on search strings that have already been received. Thus, after search engine 202 performs a search based on the search string, search engine 202 can categorize the results and store them in data center 208. During subsequent searches, these categorized results can be accessed and analyzed to generate instance suggestions.

  A categorized instance collection that includes instances specified in pre-existing structured presentations, for example, the similarity between the attributes that characterize the specified instances and the attributes that characterize instances in different categories Can be identified based on For example, a common use of the attributes “year” and “rating” uses a pre-existing structured presentation and a categorized instance collection 3420, and a categorized instance collection 3420 uses an instance 2405 Can be identified as including.

  In some implementations, a subset of instances in the categorized instance collection may be selected as an instance proposal based on values that characterize the instances included in the category. For example, the instance “Star Wars” is on the list of instance proposals based on the value that characterizes the “rating” attribute of “Star Wars”, which is similar to the value that characterizes the “rating” attribute of “Philadelphia” and “Chicago” Can be included. As another example, the instance “Peter Pan” is on the list of instance proposals based on a value that characterizes the “rating” attribute of “Peter Pan” that is different from the value that characterizes the “rating” attribute of “Philadelphia” and “Chicago”. Can be excluded.

  FIG. 35 is a schematic diagram of a collection of processes 3500 that can be used to construct a collection of new instance proposals 2115 based on pre-existing information in the structured presentation.

  The processes in collection 3500 can be viewed as a filter that is applied continuously to a large collection of potential instances 3505 and produces a smaller collection 3510 of instance proposals. The process collection 3500 includes a category filter 3515, a related attribute filter 3520, and a related value filter 3525. Category filter 3515 can include, for example, some aspects of process 3300 (FIG. 33). Related attribute filters 3520 can include, for example, aspects of process 2300 (FIG. 23), process 2500 (FIG. 25), process 2900 (FIG. 29), and / or process 3000 (FIG. 30). Related value filter 3525 can include, for example, an embodiment of process 2300 (FIG. 23), an embodiment of process 2500 (FIG. 25), an embodiment of process 3100 (FIG. 31), and / or an embodiment of process 3300 (FIG. 33). . Each filter can exclude potential instances 3505 from the instance proposal collection 3510 that can be presented to the user or added directly to the structured presentation.

  Filters 3515, 3520, 3525 can be applied in any order. However, in general, filters 3515, 3520, 3525 are applied in order of granularity. In particular, filters 3515, 3520, 3525 that reduce the number of potential instances by a maximum amount are applied first, and filters 3515, 3520, 3525 that reduce the number of potential instances by a minimum amount are applied last.

  In some implementations, any of the filters 3515, 3520, 3525 can be omitted from the collection 3500 and / or additional filters can be added to the collection 3500. For example, a user specified filter can be added to collection 3500 that can filter potential instances 3505 according to input supplied by the user.

  FIG. 36 is a flow diagram of a process 3600 for constructing a collection of new instance proposals 2115 based on pre-existing information in the structured presentation 106. Process 3600 may be performed by a system of one or more computers that performs operations with one or more machine-readable instruction sets, eg, system 200 (FIG. 2). Process 3600 may be performed alone or in conjunction with other activities. For example, process 3600 may be performed iteratively in conjunction with one or more of the processes in process collection 3500. In some implementations, the process 3600 may be performed in response to receiving a search string.

  The system performing process 3600 may perform an initial match between the instance specified in the pre-existing structured presentation 106 and the instance derived from the document collection (step 3605). The initial matching can be based on one or more of the filtering processes in the process collection 3500.

  The system performing process 3600 may determine whether the number of matches is too large, too small, or appropriate (step 3610). If there are too few matches, the system can expand the name of the specified instance (step 3615). For example, a system executing process 3600 may use alternative spellings, abbreviations, synonyms, alternative names, nicknames, and / or other keywords for instances specified in one or more of the processes in process collection 3500. Can be used.

  The system performing process 3600 may also extend one or more ranges of attribute values used in related value filtering 3525 (FIG. 35) (step 3618). This range can be expanded based on input received from the user or automatically, without user input. For example, in some implementations, the system can expand the range based on the distribution of attribute values for a selected group of instances, eg, to include a certain percentage of instances or a predetermined number of instances.

  The system performing process 3600 may also reduce the number of attributes and / or instances used in related attribute filtering 3520 (FIG. 35) (step 3620). The number of attributes and / or instances can be reduced, for example, based on the number of potential instances that are excluded by a particular attribute and / or instance. For example, if a requirement to use a particular attribute to characterize a potential instance excludes all of the potential instances, this attribute can be omitted from the associated attribute filtering. The attributes and / or instances to be removed can be determined, for example, automatically, without user input, or based on input received from the user.

  Again, the system can try to match between the instance specified in the pre-existing structured presentation 106 and the instance drawn from the document collection, but this time it has changed The parameter is used (step 3622). This matching can also be done using one or more of the filtering processes in the process collection 3500.

  If it is determined that there are too many matches (step 3610), the system performing process 3600 may narrow one or more ranges of attribute values used in the associated value filtering 3525 (Figure 35). (Step 3625). This range can be narrowed based on input received from a user or automatically, without user input. For example, in some implementations, the system can narrow the range based on the distribution of attribute values for a selected group of instances, for example, to exclude a certain percentage of instances or a predetermined number of instances. .

  The system performing process 3600 may also increase the number of attributes and / or instances used in the associated attribute filter 3520 (FIG. 35) (step 3628). The number of attributes and / or instances can be increased based on, for example, the number of potential instances excluded by a particular attribute and / or instance. The attributes and / or instances to be added can be determined, for example, automatically, without user input, or based on input received from the user.

  The system performing process 3600 may screen matches based on the changed parameters (step 3630). In particular, a narrowed range and / or an increased number of attributes and / or instances can be used in the associated value filtering 3525 (FIG. 35).

  If it is determined that the number of matches is acceptable (step 3610), the system performing process 3600 may suggest matched instances to the user (step 3635). For example, a system performing process 3600 may present one or more instance suggestions on a display screen, eg, a GUI on display screen 104.

  FIG. 37 is a diagram of a process 3700 in which new attributes can be added to extend a preexisting structured presentation. Process 3700 may be performed by a system comprising one or more computers that perform operations by executing one or more machine-readable instruction sets, eg, system 200 (FIG. 2).

  Process 3700 includes an extract operation 3705 and a merge operation 3710 that add new attributes to a pre-existing structured presentation based on information derived from documents in electronic document collection 102. In particular, the process 3700 proposes one or more new attributes based on information presented in the pre-existing structured presentation 106. For example, if the structured presentation includes multiple instances corresponding to several videos, the system 200 may suggest additional video attributes according to information derived from the electronic document collection. That is, the system 200 can identify and suggest additional attributes according to instance identifier similarity, instance categories, attribute values, or combinations thereof.

  As shown in FIG. 37, an extraction operation 3705 extracts a collection of new attribute proposals from the electronic document collection 102 using pre-existing characteristics of the structured presentation 106. Examples of characteristics include instances in pre-existing structured presentations, categories of pre-existing structured presentations, and attribute values in pre-existing structured presentations. It is done. The pre-existing structured presentation 106 characteristic can be represented as a collection of machine-readable information that is executed by executing one or more machine-readable instruction sets to perform one or more operations. The computer can receive. For example, search engine 202 (FIG. 2) can receive pre-existing characteristics of structured presentation 106.

  In an extraction operation 3705, one or more new attribute proposals can be constructed based on the content of the documents in the electronic document collection 102 and the characteristics of the pre-existing structured presentation 106. As described further below, a variety of different techniques for composing new attribute proposals can be used.

  In merge operation 3710, some or all of the new attribute proposals can be merged into pre-existing structured presentation 106 to form extended structured presentation 106. For the viewer, the expanded structured presentation can be displayed on a display device, such as the display screen 104, for example.

  All new attribute proposals configured in the extraction operation 3705 are merged with the pre-existing structured presentation 106 and need not be displayed to the viewer. For example, in some implementations, a collection of new attribute proposals can be presented to the viewer with interactive elements that allow the viewer to select one or more attributes to be added. However, in other implementations, new attribute proposals can be added automatically without user interaction and without screening new attribute proposals prior to display. Further details regarding mergers can be found, for example, in FIGS. 9-20 and associated text.

  FIG. 38 is a flow diagram of an example process 3800 for adding attributes to a structured presentation based on the content of documents in an electronic document collection. Process 3800 may be performed by a system of one or more computers that perform operations by executing one or more machine-readable instruction sets. For example, process 3800 can be performed by search engine 202 in system 200.

  In some implementations, the process 3800 may be performed in response to receiving input from a user or other system or process that triggers an update of the structured presentation 106, for example. For example, process 3800 can be performed in response to a user request to add one or more new attributes to structured presentation 106. As another example, process 3800 may be performed by a search engine, eg, search engine 202 (FIG. 2) in response to receiving a search query.

  The system performing process 3800 may receive one or more characteristics of a pre-existing structured presentation (step 3805). For example, the system can receive one or more instance identifiers that appear in a pre-existing structured presentation. As another example, the system may receive a description of a category that includes an instance identified in a pre-existing structured presentation.

  The system performing process 3800 may construct one or more attribute proposals from documents in the electronic document collection based on one or more characteristics of the pre-existing structured presentation (step 3810). ). Attribute proposals can be constructed based on these characteristics in a number of different ways. For example, in one implementation, the system may construct an attribute proposal from documents in the electronic document collection 102 by constructing a search query using instance identifiers derived from pre-existing structured presentations. it can. These search queries can be used to identify attributes that can characterize the same or similar instances using string comparison results or other matching techniques. Examples of other approaches are further described below.

  The system performing process 3800 may provide one or more attribute suggestions to the user (step 3815). For example, a list of attribute proposals can be displayed for the user on the same display screen that displays a pre-existing structured presentation.

  The system performing process 3800 may receive a user selection of one or more attribute proposals (step 3820). For example, a user interface component interacts with the user and receives one or more user inputs (e.g., mouse clicks, key presses, or other user inputs) that select one or more attribute suggestions. it can.

  The system performing process 3800 may add the selected attribute proposal to the structured presentation (step 3825). In particular, the selected attribute proposal can be used to extend an existing structured record in the structured presentation. For example, if the structured presentation is a table such as table 300 (FIG. 3), the system can add a new column 304. As another example, if the structured presentation is a collection of cards, such as a collection of cards 500 (FIG. 5), the system can add a new attribute identifier 308 to the card 502.

  FIG. 39 is a flow diagram of an example process 3900 for constructing attribute proposals from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. Process 3900 may be performed alone or in conjunction with other activities. For example, process 3900 may be performed at step 3810 in process 3800 (FIG. 38).

  Process 3900 may be performed by a system of one or more computers that perform operations by executing one or more machine-readable instruction sets. For example, process 3900 can be performed by search engine 202 in system 200 (FIG. 2).

  The system performing process 3900 may identify one or more documents that include structured components related to the instance specified in the pre-existing structured presentation (step 3905). A structured component is a number of parts or regions of an electronic document that are structured. Examples of structured components include tables, lists, records, collections of attribute-value pairs, and the like. Thus, structured components are similar to structured presentations, with predefined structures and Attribute values and instance identifiers can be organized to match.

  The entire electronic document including the structured component need not be structured. For example, an electronic document may have a table sandwiched between two paragraphs of unstructured text. Furthermore, structured components in different documents need not have the same format or follow a predetermined or permanent structure. In fact, the organization of information within one structured component can generally be changed independently of the organization of information within the structured component that appears in other documents. For example, if a structured list of schools in one person's resume is changed to remove the graduation year, the other structured lists of schools in other resumes will be changed as well There is no need to guarantee that

  A system performing process 3900 may identify documents that include structured components in a variety of ways. For example, tables and other structured components can be identified using metadata labels, such as HTML tags, found in the document itself. As another example, a structured component can be identified by identifying repetitive elements (eg, a series of commas or tab delimiters) in the document.

  Structured components are related to the specified instance in a pre-existing structured presentation if they contain information related to the specified instance. For example, a structured component that characterizes one or more of the specified instances with one or more attribute values is related to the specified instance in a pre-existing structured presentation Can be thought of as As another example, a structured component that characterizes one or more of the same attributes of an instance that is different from the instance specified in the pre-existing structured presentation is related to the specified instance Can be thought of as In many implementations, the instance and / or attribute identifiers need not be the same. Rather, conceptually related instances and attributes can be used to identify documents that contain structured components.

  Thus, in some implementations, the system performing process 3900 is within the same or related instance identifier and / or preexisting structured presentation as found in the preexisting structured presentation. Identify one or more documents that contain structured components related to instances specified in pre-existing structured presentations by identifying documents that contain the same or related attribute identifiers as found Can do.

  The system performing process 3900 may select one or more attribute proposals from the structured component (step 3910). In this selection process, attributes proposed to the user can be selected to reduce the number. The selection of attribute proposals can be performed in various ways. For example, the system may select an attribute suggestion based on an instance value in the structured component and / or an attribute value of the instance in the structured component, as further described below.

  FIG. 40 is a representation 4000 representing the composition of attribute proposals from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. In particular, representation 4000 shows the configuration of an attribute proposal that uses one implementation of process 3900 (FIG. 39).

  As shown, the pre-existing structured presentation specifies a collection of instances 4005 (ie, instances “Philadelphia” and “Chicago”). Further, different documents in an electronic document collection (eg, collection 102) include different structured components 4010, 4015, 4020. The structured components 4010, 4015, 4020 can be identified as being associated with the instance 4005 designated based on, for example, the same instance identifiers “Philadelphia” and “Chicago” appearing therein.

  As shown, the structured components 4010, 4015, 4020 include a variety of different potential attribute proposals based on different background situations. In particular, in the context of the structured component 4010, the instances “Philadelphia” and “Chicago” are part of a tabular component that represents various city characteristics. In the context of the structured component 4015, the instances “Philadelphia” and “Chicago” are part of the structured component that represents a portion of the standings of the National League East District at some point in the 1970s. In the context of the structured component 4020, the instances “Philadelphia” and “Chicago” are part of a tabular component that represents various animated features.

  Instead of suggesting all the various attributes found in structured components 4010, 4015, 4020 to the user, select attribute selection from components 4010, 4015, 4020 based on the attributes used to characterize those instances can do. In particular, as shown, the pre-existing structured presentation 106 uses the values of the attributes “year”, “rating”, and “box office receipts” to create instances “Philadelphia” and “Chicago”. Characterize. The structured component 4010 characterizes the instances “Philadelphia” and “Chicago” using the values of the attributes “population” and “area”. The structured component 4015 characterizes the instances “Philadelphia” and “Chicago” using the values of the attributes “wins”, “losses”, and “GB” (ie, game differences). The structured component 4020 characterizes the instances “Philadelphia” and “Chicago” using the values of the attributes “year”, “runtime”, and “rating”.

  The system can select from attributes within the structured components 4010, 4015, 4020 based on these characterized attributes. For example, the system identifies the correspondence between attribute identifiers “year” and “rating” in pre-existing structured presentation 106 and attribute identifiers “year” and “rating” in structured component 4020. , The attributes “director” and “runtime” can be selected as a proposal to add to the pre-existing structured presentation 106.

  As described in FIGS. 21-36 and the associated text, in some implementations, the system may also suggest or add additional instance identifiers. For example, the structured component 4020 includes instance identifiers “Peter Pan” and “Star Wars”. Thus, such a system can be proposed to include these instance identifiers in the structured presentation.

  In some implementations, even if attributes derived from structured components 4010, 4015 are not proposed in a particular configuration, such attributes are stored for use in future information requests. can do. For example, even if the cities represented by the structured component 4010 have not been selected as attribute proposals, those cities will have their respective attribute identifiers (eg, “population” and “area”) and data collections (eg, data Can be stored along with attribute values in the center 208). If the user subsequently requests information about one or more cities, such a system can access this stored information and provide additional information to the user.

  FIG. 41 is a flow diagram of an example process 4100 for constructing attribute proposals from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. Process 4100 may be performed alone or in conjunction with other activities. For example, process 4100 may be performed at step 3810 in process 3800 (FIG. 38).

  Process 4100 may be performed by a system of one or more computers that perform operations by executing one or more machine-readable instruction sets. For example, process 4100 may be performed by search engine 202 in system 200 (FIG. 2).

  The system performing process 4100 may identify one or more documents associated with one or more specified instances (step 4105). For example, the system performing process 4100 uses a string comparison to document one or more of the specified instance and its attributes and / or values and an electronic document collection such as electronic document collection 102. Can be matched. In other examples, the system performing process 4100 can access stored information (eg, information in data center 208) to identify the electronic document associated with the specified instance.

  The system performing process 4100 may extract a template for one or more of the identified documents (step 4110). The document template is used as a pattern for the arrangement of the content of individual documents in a sub-collection of documents in the electronic document collection. Documents in a sub-collection are generally from a single source, eg, a single company. For example, a book store can use a single document template as a pattern for an array of content describing different books. In another example, a furniture retailer can use a single document template as a pattern for an array of different sofa brochure content. For example, a template for an e-brochure for a sofa allows the user to select the sofa brand name, sofa photo, sofa color, textual sofa description, and sofa An array of tables characterizing dimensions, availability, and prices can be specified on the brochure. In this way, the document template can organize information about the instances in much the same way as a structured presentation and to match a predefined structure.

  In general, a document template can be used as a pattern for the entire content of an electronic document, and can even specify an array of structured components within the document, as described above. However, since the document template only specifies the content array of the sub-collection of documents in the unstructured electronic document collection, the electronic document collection itself remains unstructured. For example, even if AMAZON.COM uses one template as a pattern for an array of descriptions for all books sold by AMAZON.COM, BARNESANDNOBLE.COM and other book stores will use the same template as the book. It is not necessarily used as a pattern for the description arrangement of books sold by the store.

  Furthermore, document templates can be changed without necessarily propagating the changes to the entire collection of electronic documents, and even to a sub-collection.

  FIG. 42 is a diagram showing a part of a hypertext markup language (HTML) template 4200 used as a pattern for describing a moving image (that is, moving image “Philadelphia”). The hypertext markup language (HTML) code of the template portion 4200 is machine readable and human readable. For example, the web page can be generated using the HTML code of the template portion 4200 in the browser.

  In the example shown, the template portion 4200 is divided into two subsections 4205, 4210. Subsection 4205 is used as a pattern for an array of text that identifies the video “Philadelphia”. Subsection 4210 is used as a pattern for the array of various attribute identifiers and their values. In general, the patterns in subsections 4205, 4210 are repeated many times in a particular sub-collection of documents within an electronic document to describe different videos.

  An HTML parser can be used to extract the output format from the template portion 4200, and thus the output format can be used to identify documents having the same template. For example, HTML tags <title>, <div>, other HTML tags, and positions relative to each other can be identified by an HTML parser. Such an HTML parser can determine that the HTML tag <title> appears before the HTML tag <div>. Thus, the HTML parser can extract the output format from the template portion 4200 from the content arranged according to the template.

  Referring again to FIG. 41, after extracting the template, the system performing process 4100 may use the template to identify one or more new attributes (step 4115). For example, the system can identify an array of attributes derived from pre-existing structured representations in the template. This array can be used to infer other attributes. The system executing system 4100 may also construct one or more attribute proposals from the attributes identified using the template (step 4120). Thus, these templates themselves can be used to construct attribute proposals. For example, the HTML tag in the template portion 4200 (FIG. 42) identifies that the movie is titled.

  “Philadelphia (1993)” is characterized by the attributes “Director”, “Writer”, and “Release Date”. Any of these attributes can be used to construct one or more attribute proposals.

  Further, in some implementations, additional content in the document template can be used in constructing attribute suggestions. For example, attribute values can be used in constructing attribute proposals. For example, if the value of attribute “year” is, for example, 1976, attribute “start time” may be excluded from the collection of attribute proposals to characterize the movie.

  FIG. 43 is a diagram of a process 4300 in which a collection of new attribute proposals 3715 can be constructed based on pre-existing information in the structured presentation 106. Process 4300 may be performed by one or more computer systems, eg, system 200 (FIG. 2), that perform operations by executing one or more machine-readable instruction sets.

  Process 4300 performs an extract operation 4305 on instance / attribute collection 4310 based on pre-existing information in structured presentation 106. An instance / attribute collection 4310 is a collection of information that associates an instance with its attributes and, in some implementations, with the values of those attributes. The information in the collection 4310 can be used as a response to receiving a trigger (e.g., a search query) or as part of the process of indexing the electronic document collection 102 in anticipation of receiving a trigger, e.g. It can be extracted from the documents in the document collection 102. In some implementations, the information in the collection 4310 can include the content of previous structured presentations presented to the current user or other users. In general, the attribute proposal is provided to a user who selects an attribute proposal to be added to the structured presentation, as described in steps 3815, 3820, 3825 (FIG. 38).

  An association between an instance and its attributes can be established in collection 4310 by structuring information storage in collection 4310. FIG. 44 is a schematic diagram of a table 4400 associating attributes and instances in collection 4310. Table 4400 contains a collection of records 4402, 4404, 4406, 4408, 4410, 4412, 4414, each record with an instance identifier as a description of the document's location and attributes that characterize the identified instances within those documents. Associate. Information in records 4402, 4404, 4406, 4408, 4410, 4412, 4414 may be organized in a collection of columns 4415, 4420, 4425, 4430, 4435, 4440. In particular, column 4415 can include an instance identifier. Column 4420 may include a description of the location of the electronic document that includes the instance identified in column 4415. Columns 4425, 4430, 4435, 4440 may identify attributes that characterize the instance identified in column 4415 in the document whose location is described in column 4420.

  As shown, different electronic documents can include different categories and quantities of information that characterize the same instance. For example, a document whose location is identified in column 4420 of record 4404 contains two attributes for instance "INSTANCE_2", while a document whose location is identified in column 4420 of record 4410 is three attributes of instance "INSTANCE_2" including. Further, the attributes in record 4404 (ie, attribute “ATTR_5” and attribute “ATTR_6”) are partially different from the attributes in record 4410 (ie, attribute “ATTR_5”, attribute “ATTR_8”, and attribute “ATTR_9”). .

  A data collection 4310 (such as table 4400) that associates attributes and instances may be formed in a number of different ways. For example, a document that includes an internal structured component may be identified. Examples of such internal structured components are tables and lists that appear in HTML documents. The relationship between attributes and instances in these internal structured components can be copied to form data collection 4310.

  As another example, collection 4310 can be formed from the content of previous structured presentations presented to the current user or other users.

  As yet another example, after a first document is identified as containing attributes and instances, the template for that document can be used to extract attributes and instances from other documents that contain the same template. For example, if a stereo retailer describes different stereos offered for sale using the same document template, it uses the arrangement of information in the first electronic document for the first stereo to Information can be extracted from other electronic documents.

  In some implementations, techniques such as natural language parsing can be used to identify instances and attributes. For example, the electronic document can be parsed to identify phrases such as “[Instance] has a / an [attribute]” in the electronic document.

  In some implementations, the data collection 4310 can classify instances and their attributes. For example, instances and attributes can be categorized as North American cities, National League East District teams, or popular movies. Storage of information in the data collection 4310 can be based on such classification. For example, different categories can be stored in different files, records, and the like.

  Referring again to FIG. 43, the process 4300 proposes one or more new attributes based on information presented in the pre-existing structured presentation 106. For example, if the structured presentation includes multiple instances corresponding to several videos, the system 200 may suggest additional attributes for the videos according to information derived from the data collection 4310. That is, the system 200 can identify and suggest additional attributes based on the attributes used to characterize the same instance. For example, the system 200 can suggest other attributes that are typically used to characterize a video, such as showtime, theater, or airtime.

  FIG. 45 is a flow diagram of a process 4500 for constructing an attribute proposal from a collection of instances and attributes based on pre-existing structured presentation characteristics. Process 4500 may be performed by a system of one or more computers that perform operations by executing one or more machine-readable instruction sets. For example, process 4500 may be performed by search engine 202 in system 200 (FIG. 2).

  Process 4500 may be performed alone or in conjunction with other activities. For example, process 4500 may be performed in process 4300 (FIG. 43). As another example, process 4500 may be performed at step 3810 in process 3800 (FIG. 38) alone or in conjunction with one or both of processes 3900, 4100 (FIGS. 39, 41).

  The system executing process 4500 can access the collection of instances and their attributes (step 4505). For example, a system executing process 4500 can access an instance and attribute collection 4310 (FIG. 43) stored in data center 208 (FIG. 2).

  The system performing process 4500 may identify one or more related attributes based on the instance attribute characteristics specified in the pre-existing structured presentation (step 4510). For example, the system can compare the instance attributes of the instances stored in the instance and attribute collection 4310 (FIG. 43) with the instances specified in the structured presentation 106. The system uses this comparison result to share any of the stored instances, if any, with the attribute identifier, or the attribute specified in the structured presentation 106 that already has the attribute concerned. You can decide what to do. For example, pre-existing structured presentation 106 presumes to use the attributes “ATTR_3” and “ATTR_5” to characterize the collection of instances. After reviewing instance and attribute collection 4310, such as table 4400 (Figure 44), the system uses attributes "ATTR_7" and "ATTR_7" to characterize instances "INSTANCE_1" and "INSTANCE_2" in records 4402, 4406. Suggestions can be made based on use in conjunction with ATTR_3 and ATTR_5.

  FIG. 46 is a flow diagram of a process 4600 for constructing a collection of new attribute proposals 3715 based on pre-existing information in the structured presentation 106. Process 4600 may be performed by one or more computer systems, eg, system 200 (FIG. 2), that perform operations by executing one or more machine-readable instruction sets. Process 4600 may be performed alone or in conjunction with other activities. For example, process 4600 may be performed in process 4300 (FIG. 43). As another example, process 4600 may be performed alone or in conjunction with one or more of processes 3900, 4100, 4500 (FIGS. 39, 41, 45) at step 3810 in process 3800 (FIG. 38). Can be done. In some implementations, process 4600 may be performed in response to receiving input (eg, from a user of system 200 or from another system or process). For example, process 4600 may be performed by search engine 202 in response to receiving a search string.

  The system performing process 4600 may identify one or more trusted sources for one or more specified instances (step 4605). For example, the system can access a collection of trusted sources of documents in the electronic document collection 102 assembled, for example, by a programmer.

  As another example, the system may receive user-specific input that identifies one or more trusted sources of documents in the electronic document collection 102 as “trusted” from the user's perspective. For example, the display screen 104 displaying the pre-existing structured presentation 106 can include a GUI component that allows the viewer to specify a trusted source of the document. The identity of the trusted source can be received in conjunction with the search query. For example, viewers can use JD POWER AND ASSOCIATES, AMAZON.COM, and MAJOR LEAGUE BASEBALL at http://www.jdpower.com/, http://www.amazon.com/, and http: // www. Can be identified as a trusted source of documents at mlb.com/. In some implementations, user-specific input can identify the subject matter for which the source is reliable. For example, MAJOR LEAGUE BASEBALL can be identified as a reliable source of baseball statistics, but cannot be considered as a reliable source of information about drug tests.

  As yet another example, a system performing process 4600 can analyze a collection of user-specific inputs that identify trusted sources from multiple users to assemble a “generic” collection of trusted sources. . For example, many users can identify AMERICAN AUTOMOBILE ASSOCIATION (AAA) as trusted. AAA can then be added to the collection of trusted sources based on statistical analysis of these identification results.

  The system performing process 4600 may determine additional attributes from a trusted source for the instance specified in the pre-existing structured presentation (step 4610). For example, the system can access documents provided by a trusted source and identify one or more documents that characterize a specified instance using one or more attributes. The system can extract attribute identifiers from these documents using parsers or other string comparison techniques.

  As another example, the system can access a data collection that associates attributes and instances, such as table 4400 (FIG. 44). The system records 4440, 4404, 4406, 4408, 4410, 4412, 4414, etc., and records 4402, 4404, 4406, 4408, 4410, instances and locations identified in a pre-existing structured presentation. The documents identified at 4412, 4414 can be filtered based both on whether they came from a trusted source. For example, if AMAZON.COM is a trusted source, a collection that associates attributes and instances can be scanned to identify documents that have the http://www.amazon.com/ domain.

  The system performing process 4600 may compare these additional instances to instances such as table 4400 (FIG. 44) and attributes in the attribute collection (step 4615). For example, the system can use string comparison, or other comparison techniques, to compare additional instances with instances stored in instances and attribute collections.

  The system performing process 4600 may identify attributes in the instance and attribute collection based on these comparison results (step 4620).

  FIG. 47 is a flow diagram of a process 4700 for identifying related instances for use in constructing attribute proposals based on information in pre-existing structured presentation 106. Process 4700 may be performed by one or more computer systems, eg, system 200 (FIG. 2), that perform operations by executing one or more machine-readable instruction sets. Process 4700 may be performed alone or in conjunction with other activities. For example, process 4700 may be performed in process 1500 (FIG. 15). As another example, process 4700 may be performed at step 3810 in process 3800 (FIG. 38) alone or with one or more of processes 3900, 4100, 4500, 4600 (FIGS. 39, 41, 45, 46). Can be executed in conjunction. In some implementations, process 4700 may be performed in response to receiving input (eg, from a user of system 200 or from another system or process). For example, process 4700 may be performed by search engine 202 in response to receiving a search string.

  The system performing process 4700 may identify one or more related instances based on attributes and / or attribute values that characterize the instances specified in the pre-existing structured presentation. For example, the system can identify a related instance by comparing the attribute value of the specified instance with the attribute value of the other instance and determine whether the other instance is related. Such a comparison may require, for example, that the attribute values are the same or that the attribute values are within a certain range. Such ranges can be determined based on attribute value ranges that have already been characterized, for example, by instances specified in the user via interactive elements or pre-existing in the structured presentation. it can.

  In some implementations, the system performing process 4700 may convert attribute values to normal units of measure before comparing the attribute values. For example, if the specified unit of measure is feet, but the unit of measure for one or more extracted attribute values is meters, the system converts the metric value to feet using normal techniques. can do.

  FIG. 48 is a flow diagram of a process 4800 for composing a collection of new attribute proposals 3715 based on pre-existing information in the structured presentation 106. Process 4800 may be performed by one or more computer systems, eg, system 200 (FIG. 2), that perform operations by executing one or more machine-readable instruction sets. Process 4800 may be performed alone or in conjunction with other activities. For example, process 4800 may be performed in process 1500 (FIG. 15). As another example, process 4800 is a step 3810 in process 3800 (FIG. 38), either alone or one of processes 3900, 4100, 4500, 4600, 4700 (FIGS. 39, 41, 45, 46, 47). Can be performed in conjunction with one or more. In some implementations, process 4800 may be performed in response to receiving input (eg, from a user of system 200 or from another system or process). For example, process 4800 may be performed by search engine 202 in response to receiving a search string.

  The system executing process 4800 may access a categorized collection of instances and attributes (step 4805). For example, the system can access the instance and attribute collection 1510 to access one or more categorized collections of instances and attributes generated during previous searches.

  The system executing process 4800 may identify a category that includes the specified instance (step 4810). In some implementations, the system may identify categories that include instances based on similar attributes, similar attribute values, combinations of these characteristics, and / or other technologies.

  The system performing process 4800 may select one or more attribute proposals from the identified category (step 4815). For example, in some implementations, attribute proposals can be selected from an identified category based on the number of times those attributes are used to characterize instances within that category.

  FIG. 49 is a representation 4900 representing the composition of attribute proposals from electronic documents in an electronic document collection based on pre-existing structured presentation characteristics. In particular, representation 4900 shows the configuration of an attribute proposal that uses one implementation of process 4800 (FIG. 48).

  As shown, the pre-existing structured presentation specifies a collection of instances 4005 (ie, instances “Philadelphia” and “Chicago”). Further, instances derived from different documents in an electronic document collection (eg, collection 102) have already been categorized into different collections 4910, 4915, 4920. The categorized instance collection 4910, 4915, 4920 may be identified as being associated with the instance 4005 designated based on, for example, the same instance identifiers “Philadelphia” and “Chicago” appearing therein.

  As shown, the categorized instance collections 4910, 4915, 4920 are categorized in a variety of different ways. In particular, the categorized instance collection 4910 is categorized as a collection of “North American Cities”. Categorized instance collection 4915 is categorized as a collection of “National League East Teams”. Categorized instance collection 4920 is categorized as a collection of “Popular Movies”. In the context of the system 200, the categorized instance collections 4910, 4915, 4920 can be stored in the data center 208 (FIG. 2). That is, system 200 can generate one or more categories of instances based on search strings that have already been received. Thus, after search engine 202 performs a search based on the search string, search engine 202 can categorize the results and store them in data center 208. During subsequent searches, these categorized results can be accessed and analyzed to generate attribute proposals.

  A categorized instance collection that includes instances specified in pre-existing structured presentations, for example, the similarity between the attributes that characterize the specified instances and the attributes that characterize instances in different categories Can be identified based on For example, a common use of the attributes “year” and “rating” is to use a pre-existing structured presentation and a categorized instance collection 4920 as the categorized instance collection 4920 Can be identified as including.

  In some implementations, a subset of attributes in the categorized instance collection may be selected as an attribute proposal based on attributes that characterize the instances included in the category. For example, characterizing a video instance using the attribute “Start time” can be viewed as an indication that only information about the currently playing video is included in the structured display. Thus, attributes such as “playing at” and “coupons available” may be included in the list of attribute proposals. As another example, the attribute “year” may be excluded from the list of attribute proposals based on characterizing the animated instance in the pre-existing structured display using the attribute “Start time”.

  FIG. 50 is a schematic diagram of a collection 5000 of processes that can be used to construct a collection of new attribute proposals 3715 based on pre-existing information in structured presentation 106.

  The processes in collection 5000 can be viewed as a filter that is applied continuously to a large collection of potential attributes 5005 and produces a smaller collection 5010 of attribute proposals. Each filter can exclude potential attributes 5005 from the attribute proposal collection 5010 that can be presented to the user or added directly to the structured presentation.

  Filters 5015, 5020, 5025 can be applied in any order. However, in general, the filters 5015, 5020, 5025 are applied in order of particle size. In particular, filters 5015, 5020, 5025 that reduce the number of potential attributes by a maximum amount are applied first, and filters 5015, 5020, 5025 that reduce the number of potential attributes by a minimum amount are applied last.

  In some implementations, any of the filters 5015, 5020, 5025 can be omitted from the collection 5000 and / or additional filters can be added to the collection 5000. For example, a user-specified filter can be added to collection 5000 that can filter potential attributes 5005 according to input supplied by the user.

  FIG. 51 is a flow diagram of a process 5100 for composing a collection of new attribute proposals 3715 based on pre-existing information in the structured presentation 106. Process 5100 may be performed by one or more computer systems, eg, system 200 (FIG. 2), that perform operations by executing one or more machine-readable instruction sets. Process 5100 may be performed alone or in conjunction with other activities. For example, process 5100 may be performed iteratively in conjunction with one or more of the processes in process collection 5000. In some implementations, the process 5100 may be performed in response to receiving a search string.

  The system performing process 5100 may perform an initial match between the instance specified in the pre-existing structured presentation 106 and the attributes derived from the document collection (step 5105). The initial matching can be based on one or more of the filtering processes in the process collection 5000.

  The system performing process 5100 may determine whether the number of matches is too large, too small, or appropriate (step 5110). If there are too few matches, the system can widen the name of the specified instance (step 5115). For example, a system executing process 5100 may use alternative spellings, abbreviations, synonyms, alternative names, nicknames, and / or other keywords for instances specified in one or more of the processes in process collection 5000. Can be used.

  The system performing process 5100 may also reduce the number of attributes and / or instances used in the related instance and / or attribute filtering 5020 (FIG. 50) (step 5120). The number of attributes and / or instances can be reduced, for example, based on the number of potential instances that are excluded by a particular attribute and / or instance. For example, if the requirement that a particular instance is characterized by an attribute excludes all of the potential attributes, this instance can be omitted from the related instances and / or attribute filtering. The attributes and / or instances to be removed can be determined, for example, automatically, without user input, or based on input received from the user.

  Again, the system can attempt to match between the instance specified in the pre-existing structured presentation 106 and the instance derived from the document collection, but this time the modified parameter Is used (step 5122). This matching can also be done using one or more of the filtering processes in the process collection 5000.

  If it is determined that there are too many matches (step 5110), the system performing process 5100 determines the number of attributes and / or instances used in the relevant attribute and / or instance filtering 5020 (Figure 50). It can be increased (step 5128). The number of attributes and / or instances can be increased based on, for example, the number of potential attributes that are excluded by a particular attribute and / or instance. The attributes and / or instances to be added can be determined, for example, automatically, without user input, or based on input received from the user. For example, the instance to add can be determined using process 4800 (FIG. 48).

  The system performing process 5100 may screen matches based on the changed parameters (step 5130). In particular, a narrowed range and / or an increased number of instances can be used in the associated value filtering 5025 (FIG. 50).

  If it is determined that the number of matches is acceptable (step 5110), the system performing process 5100 may suggest matched attributes to the user (step 5135). For example, a system performing process 5100 can present one or more attribute proposals on a display screen, eg, a GUI on display screen 104.

  FIG. 52 is a schematic diagram illustrating a system 5200 in which attribute values 307 derived from two or more electronic documents in an electronic document collection are presented to a user in a structured presentation. In addition to electronic document collection 102, display screen 104, and data communication path 108, system 5200 includes structured data 5205 and merge module 5210. In operation, the system 5200 extracts attribute values from an unstructured collection of electronic documents in the electronic document collection 102, merges the information with information derived from the structured data 5205, and writes it to the structured presentation 106.

  System 5200 can write attribute values to all or only a portion of structured presentation 106. There are many situations in which attribute values are written to only part of a structured presentation. For example, such writing can be done as part of adding a new instance (and thus a new structured record) to the structured presentation 106. As another example, such writing can be done as part of adding a new attribute to the structured presentation 106. As yet another example, such writing can be part of an improvement in some of the existing attribute values within structured presentation 106. For example, certain parts of the original attribute value can be checked to see if they are correct or to verify that the instance being characterized has not changed.

  Structured data 5205 is a structured collection of information. Information in structured data 5205 may be organized according to a predefined data model. For example, structured data 5205 can be organized according to a hierarchical or relational data model and stored in a data storage device. In some cases, all or part of the structured data 5205 can be presented to the user in a structured presentation. For example, in some implementations, structured data 5205 is pre-presented to the user on the same display screen 104 that presents the structured presentation 106 with the new attribute values drawn from the collection 102 written. The structured presentation 106 can be.

  The merge module 5210 is a collection of one or more machine readable instruction sets that are deployed on one or more data processing devices. The merge module 5210 can provide functionality for identifying new attribute values and for placing values in the structured presentation 106 to be written. The operations performed by merge module 5210 are described in further detail below.

  FIG. 53 is a schematic diagram illustrating one implementation of a system 5300 in which attribute values derived from two or more electronic documents in the electronic document collection 102 are presented to a user in a structured presentation. In addition to electronic document collection 102, display screen 104, data communication path 108, search engine 202, crawler 204, and compression, indexing, and ranking module 210, system 5300 includes attribute / value / instance collection in data center 208. 5305 is also provided.

  The attribute / value / instance collection 5305 is a collection of information that associates an instance with its attributes and also with the values of those attributes. Information in collection 5305 is collected as a response to receiving a trigger (e.g., a search query) or as part of the process of indexing electronic document collection 102 in anticipation of receiving a trigger, e.g. It can be extracted from the electronic document in 5305.

  Associations between instances, attributes, and their values may be established within collection 5305 by structuring information storage within collection 5305. For example, FIG. 54 is a schematic diagram of a table 5400 that can associate attributes, values, and instances in collection 5305 (FIG. 53). Table 5400 includes a collection of records 5402, 5404, 5406, 5408, 5410, each record associating an instance identifier with a description of the document location and attributes and values characterizing the identified instance. Information in records 5402, 5404, 5406, 5408, 5410 may be organized in a collection of columns 5415, 5420, 5425, 5430, 5435, 5440, 5445, 5450. In particular, column 5415 can include an instance identifier. Column 5420 can include a description of the location of the electronic document that includes the instance identified in column 5415. Columns 5425, 5435, 5445 may characterize the instance identified in column 5415, and may include descriptions of attributes that are themselves characterized by values in the document whose location is described in column 5420. Columns 5430, 5440, 5450 may contain a description of values that characterize the attributes described in columns 5425, 5435, 5445 of the instance identified in column 5415 in the document whose location is described in column 5420. it can.

  As shown, different electronic documents can include different categories and quantities of information that characterize the same instance. For example, a document whose location is identified in column 5420 of record 5404 includes information characterizing the three attributes of instance “INSTANCE_1”, but a document whose location is identified in column 5420 of record 5406 is of instance “INSTANCE_1”. Contains information that characterizes two attributes. In addition, the attributes characterized in record 5404 (i.e. attribute `` ATTR_5 '', attribute `` ATTR_6 '', attribute `` ATTR_7 '') are attributed to record 5406 (i.e. attribute `` ATTR_3 '', attribute `` ATTR_4 '') Different.

  Furthermore, even values used to characterize the same attributes of the same entity may be different in different electronic documents. For example, the document whose location is identified in column 5420 of record 5402 contains the value “VALUE_3A” that characterizes the attribute “ATTR_3” of instance “INSTANCE_1”, but the document whose location is identified in column 5420 of record 5406 is the same. Contains the value “VALUE_3B” that characterizes the same attribute “ATTR_3” of the instance “INSTANCE_1”.

  There are many different potential sources of such discrepancies between values that characterize the same attributes of the same entity in different documents. For example, a document may contain fake information that incorrectly characterizes an entity's attributes. Not only inadvertent mistakes, attribute values can change over time. Examples of this include the value of the “height” attribute of the instance of a high school basketball player, the value of the “list price” attribute of the house instance, and the value of the “mayor” attribute of the city instance. Some documents can be updated with the correct values, but some documents are original but currently hold incorrect values.

  Furthermore, even a completely accurate document may characterize the same attributes of the same entity in different ways. For example, different documents can represent the same value using different units. As another example, different documents can represent the same value with different accuracies (e.g., `` about a two hour drive to Phoenix '' and `` a 130 minute drive to Phoenix at the posted speed limits "(130 minutes drive to Phoenix at the speed limit shown on the sign)).

  Such inconsistencies are inherently inherent in the context of the unstructured electronic document collection, eg, the document collection 102. In this regard, as described above, different users presenting information in different ways can add different documents to the collection 102. Formal users to add documents do not collaborate so that the information is presented consistently, and a formal mechanism to ensure that the presentation of information in different documents remains unchanged Nor.

  A data collection 5305 (eg, table 5400) that associates attributes, values, and instances may be formed in a number of different ways. For example, a document that includes an internal structured component may be identified. Examples of such internal structured components include tables and lists that appear in HTML documents, and the like. Relationships between attributes, values, and instances in these internal structured components can be copied to form data collection 5305.

  As another example, after a first document is identified as containing attributes, values, and / or instances, the template for that document is used to attribute, values, and instances from other documents that contain the same template. Can be extracted. For example, if a stereo retailer describes different stereos offered for sale using the same document template, it uses the arrangement of information in the first electronic document for the first stereo to Information can be extracted from other electronic documents.

  In yet another example, a single document template can be used to extract attributes, values, and / or instances from the document. For example, a template may specify multiple attributes and an array of values characterizing those attributes with respect to an instance identifier. If some of these attributes and / or values are known, an array of those known attributes and / or values can be identified and used to identify other attributes and / or values it can.

  For example, a template for a single web page might use an array of attribute / value pairs “Director: Orson Welles”, “Writer: Orson Welles, Herman J. Mankiewicz”, and “Release Date” for the identifier of the video instance “Citizen Kane”. : May 1, 1941 "can be specified. If the attributes and values "Director: Orson Welles" and "Release Date: May 1, 1941" were already known, use the attribute / value array for these video instance identifiers "Citizen Kane" "Writer: Orson Welles, Herman J. Mankiewicz" can be extrapolated.

  In some implementations, techniques such as natural language parsing can be used to identify instances, attributes, and their values. For example, the electronic document can be parsed to identify phrases such as “[Instance] has a / an [attribute] of [value]” in the electronic document.

  FIG. 55 is a flow diagram of an example process 5500 for presenting attribute values derived from two or more electronic documents in an electronic document collection to a user in a structured presentation. Process 5500 may be performed by one or more computers that perform operations by executing one or more machine-readable instruction sets. Process 5500 may be performed independently of or in conjunction with other data processing activities. For example, process 5500 may be performed as part of process 600 (FIG. 6).

  The system executing process 5500 may receive the instance identifier and attribute identifier (step 5505). The system performing process 5500 can retrieve instance and attribute identifiers directly from the user (e.g., in the form of a search query) or indirectly (e.g., as part of structured data collection 905 (Figure 9)). Can receive.

  The system performing process 5500 may identify an electronic document associated with the received instance that includes the value of the attribute (step 5510). For example, the system can access the attribute / value / instance collection 5305 in the data center 208 (FIG. 53) to identify the associated electronic document. As another example, a search engine (eg, search engine 202) can identify related documents by performing a keyword search using the instance and attribute identifier. In some cases, such keyword searches can be supplemented with linguistic parsing, or other techniques that facilitate value identification.

  The system performing process 5500 may define a subset of values for the identified attributes of the identified instances for presentation in the structured presentation. These subsets of values may include one or more values that are considered or may be appropriate for writing to the structured presentation. In particular, a subset of these values can be thought of as characterizing the identified attributes of the identified instance exactly and in a manner consistent with the desire of the viewer of the structured presentation. As described further below, the viewer's desire for a structured presentation can be based on, for example, the results of selection of values received from the viewer, or for information such as, for example, structured data 905 (FIG. 9). Confirmation can be based on the characterization of the same or other attributes of the same or other instances in a pre-existing structured collection.

  The system performing process 5500 may comprise instructions for displaying the structured presentation with the subset of values written (step 5520). Thus, the structured presentation can be presented based on information collected from the collection of electronic documents (ie, a subset of values collected from the electronic document collection) (step 615, FIG. 6).

  In some implementations, the process 5500 may be performed several times, for example, the number of instance identifiers and / or attribute identifiers.

  FIG. 56 is a flow diagram of a process 5600 for determining one or more values to present in a structured presentation. Process 5600 may be performed alone or in conjunction with other activities. For example, process 5600 may be performed at step 5515 in process 5500.

  The system performing process 5600 may divide the values of attributes from two or more documents in the electronic document collection into two or more groups (step 5605). Grouped values can be derived directly from the electronic document or can be derived from a description of the content of the electronic document, such as attribute, value, and instance associations such as table 5400 (FIG. 54).

  The system can group values using one or more different standards for determining when the values should be grouped. For example, one standard may require that grouped values are the same. For example, two values “4” can be grouped together. Other standards may require that numerical values be within a specific range that is the same. For example, the values “3.14” and “3.14159” can be grouped together. Other standards may require that the word values be identical or have the same meaning. For example, the words “czar”, “czar”, and “tsar” can all be grouped together. Other standards may require that word values represent the same concept in the concept ontology. For example, the words “pink” and “mauve” can be grouped together. Other standards allow values written in different formats to be grouped together. For example, the dates “July 25, 1982” and “7/25/1982” can be grouped together. In other standards, values written in different units can be grouped together. For example, the measurement units “1 m” and “100 cm” can be grouped together. Other standards allow groups of values that are written in different but nearly equal units. For example, measurement units “1 m” and “39 inches” can be grouped together.

  Using these groupings, the system running process 5600 performs one or more of the following subprocesses in any order, selects a group, and therefore a collection of attribute values A subset of values can be selected.

  In the first sub-process, the system performing process 5600 may select the group with the highest “value” to present in the structured presentation (step 5610). In some implementations, the “value” of a group reflects the count of values within that group. According to statistical terms, the system performing process 5500 can select a value that has a high frequency in the electronic document collection. In fact, this approach allows documents in an electronic document collection to “vote” on the value of an instance attribute.

  In other implementations, the “value” of the group is not only a value count, but also an individual count weight or score with parameters that reflect the characteristics of the document from which the value is derived. For example, the count is weighted based on, for example, the page rank of the document from which the value is drawn, the weighting factor for that document given by the user, the number of other values drawn from the document, or the “age” of the document be able to. For example, a more recently created document may be considered more accurate when characterizing attributes of a particular instance.

  This sub-process approach is effective, for example, in eliminating inadvertent attribute characterization errors when values on one electronic document are typographical errors. However, alone, this approach may select inappropriate values under certain circumstances. For example, even if a large number of documents characterize a volume attribute in liters, a viewer may be interested in having that attribute characterized in gallons in a structured presentation.

  In the second sub-process, the system performing process 5600 may receive a user specification of constraints on, for example, an acceptable value range, or an acceptable value unit (step 5615). For example, the system can include a GUI component that allows a user to select a range of values or units of measurement constraints on a display screen, eg, display 104 (FIG. 1). The constraint condition may be that there is no upper limit (for example, “value> 1”) or that the upper and lower limits are determined (for example, a value in the range of 1 to 10).

  After receiving this constraint, the system performing process 5600 may select a group that matches the received constraint for presentation in the structured presentation (step 5620). For example, if the user selects “meters” as the appropriate unit of measurement, the system implementation may select one or more groups of values expressed in meters.

  This sub-process approach is effective in ensuring that the values posted in structured presentations are presented in an organized system layout. For example, the unit of measure of the value used to characterize Michael Jordan's height, for example, can be constrained to be the same as the unit of measure of the value used to characterize Magic Johnson's height. Such an organized system arrangement allows the user to easily compare the values of the same attribute in different instances without having to worry about the unit in which the value is presented.

  In the third sub-process, the system executing process 5600 can determine the “quality” of the document from which the attribute value in each group was derived (step 5625). The “quality” of the document may reflect the probability that the information in the document is accurate and does not mischaracterize the value of the attribute. For example, commercial suppliers of goods generally provide accurate information about those goods. Thus, the “quality” of information provided by a commercial supplier can be considered higher than the “quality” of information provided by an individual. As another example, bias can be taken into account when determining the quality of the document from which information is retrieved. For example, information derived from a source that is said to be an independent information source (eg, Congressional Budget Office) may be considered of higher quality than information derived from political parties.

  As another example, the quality of a document may be based on the user's specification of the source of the document or the quality of the document itself. For example, the user can indicate that the car reliability information drawn from the consumer union (the publisher of the consumer report) is high quality, but the car reliability information drawn from Road & Truck magazine is not. . The system performing process 5600 may also select a group that includes values derived from the highest quality document (step 5630).

  This sub-process approach is particularly useful in situations where attribute values are the subject of discussion. There are disagreements about various values, for example, the true height of a college point guard, the best cheese steak in Philadelphia, and the number of stars given to a particular hotel. By allowing the viewer to specify a preferred “high quality” provider of values that characterize such attributes, the structured presentation can be modified to meet the viewer's expectations.

  As described above, a combination of these activities and other activities can be performed to select one or more values for presentation in the structured presentation. For example, all groups of values that represent a certain percentage of the total number of values (e.g.,> 10%) are selected in the first screening (step 5610), and then these groups are selected based on unit designation. Further sorting can be performed (step 5620). The remaining groups can be further screened based on the quality of the document from which the attribute values within each group are derived (step 5630). Thus, in some implementations, process 5600 provides one or more values from the remaining groups that are consistent with the unit of measure and that are derived from a source that is not mischaracterized and preferred by the viewer. can do.

  FIG. 57 is a flow diagram of a process 5700 for selecting one or more values to present in a structured presentation. Process 5700 may be performed alone or in conjunction with other activities. For example, process 5700 may be performed at step 5515 in process 5500 (FIG. 55) alone or in conjunction with one or more of the sub-processes of process 5600 (FIG. 56).

  The system performing process 5700 may divide the values of attributes extracted from two or more documents in the electronic document collection into two or more groups (step 5605). The system executing process 5700 may present a description of the group of values to the user (step 5705). For example, the system can list the most common values in each group, or a list of some of the values in each group, on the display, for example, on the display screen 104 (FIG. 1). Can be displayed. In some implementations, the group description of values may include additional information that characterizes those groups. For example, display a count of the number of values in each group, display a percentage that reflects the percentage of extracted values found in each group, and / or display a description of the units of measure within the group Is possible. As another example, an estimate of the quality of the electronic document from which the values in each group were extracted can be displayed. As yet another example, the identification, location, and / or snippet or other excerpt from which the values in each group were extracted can be displayed.

  In some implementations, descriptions of groups of values are sorted with certainty-based ordering. That is, the description of the group of values is ordered according to the certainty regarding the accuracy of the values within each group for the system executing process 5700. The certainty of the accuracy of the values in each group can be determined based on, for example, the number of values in each group, the quality of the document from which those values were extracted, and the like.

  The system performing process 5700 may receive a user selection of a desired group of values (step 5710). For example, the system can receive a user interaction that identifies a selection of a group of desired values.

  In some implementations, the system performing process 5700 may change other aspects of the structured presentation based on user selection (step 5715). For example, if a user selects a group of values in meter units of measurement and characterizes the same attributes of other instances, but there are other values presented in feet units of measure, such values are structured presentations. Converted at 106 and can be posted in meters.

  FIG. 58 is a flow diagram of an example process 5800 for selecting one or more values to present in a structured presentation. Process 5800 may be performed alone or in conjunction with other activities. For example, process 5800 may be used alone or in conjunction with one or more of the sub-processes of process 5600 (FIG. 56) and / or process 5700 (FIG. 57) at step 5515 in process 5500 (FIG. 55). Can be executed.

  The system performing process 5800 may identify electronic documents in a structured data collection, eg, an electronic document collection associated with instances and other attributes in structured data collection 905 (step 5805). As described above, the structured data collection 905 can be a version of the structured presentation 106.

  Documents related to instances and other attributes in the structured data collection can be identified in various ways. For example, a system executing process 5800 may access a data collection that associates instances, their attributes, and the values that characterize these attributes, eg, attribute / value / instance collection 5305 (FIG. 53). A document that contains information related to instances and other attributes in the structured data collection can be identified therein, for example, by comparing the identifiers and attributes of the instances in both data collections. As another example, a system performing process 5800 may use an instance identifier and attribute as a search term in one or more search queries. Such search queries can be used alone or in conjunction with other extraction techniques such as language parsing and string comparison to identify relevant documents in an electronic document collection.

  The system performing process 5800 may also select one or more values to present in the structured presentation from the identified document (step 5810). FIG. 59 is a schematic diagram illustrating a situation in which attribute values derived from electronic documents in the electronic document collection 102 are presented to the user in the structured presentation 106. In particular, a system such as system 900 (FIG. 9) derives attribute values from table 5400 that associates attributes, values, and instances derived from electronic documents in electronic document collection 102. The system also merges these attribute values with the initial structured presentation 106 to form the final structured presentation 106. Thus, the initial structured presentation 106 serves as structured data 905 (FIG. 9).

  In the example shown, the initial structured presentation has been modified to associate the value of the new attribute (ie, attribute “AIRPORT”) with the instance identified in the structured presentation. In particular, a new column 5905 has been added to the initial structured presentation. At the top of column 5905 is an attribute identifier 5910 that identifies the new attribute using the word “AIRPORT”. The addition of a new attribute value to the structured presentation may be triggered, for example, based on an interactive operation by the user or automatically, as further described in FIGS. 9-20 and related text.

  A system, such as search engine 202, can access a data collection that associates attributes, values, and instances derived from electronic documents in electronic document collection 102 (such as table 5400). Using such a data collection, the system can select one or more values that characterize one or more new attributes of the instance in the initial structured presentation. For example, in the illustrated situation, the value 5915 (ie, the value “value_ai”) characterizes the attribute “AIRPORT” of the instance “NEW YORK” in the document “DOC_3”. If necessary, the system may select one or more values for the new attribute to display using, for example, one or more of processes 5600, 5700, 5800 (FIGS. 56, 57, 58). it can.

  After the values characterizing the new attributes of one or more of the instances in the initial structured presentation are selected, the final structured presentation 106 can be presented to the viewer. The final structured presentation 106 can include selected values that characterize new attributes of one or more of the instances in the structured presentation. For example, as shown, the value 5915 can be presented in the final structured presentation 106 for viewing by a viewer.

  FIG. 60 is a schematic diagram illustrating a process 6000 in which both attributes and attribute values are derived from an electronic document in an electronic document collection and presented to the user in a structured presentation.

  In process 6000, the initial structured data collection 905 can include a pre-existing structured presentation 6005. Pre-existing structured presentation 6005 can characterize an instance using one or more attribute values, eg, attribute values in column 6010. New attributes that characterize instances in pre-existing structured presentation 6005 should be constructed based on the content of the electronic documents in electronic document collection 102, as described in FIGS. 37-51 and the associated text. Can do. New attributes can be added to the pre-existing structured presentation 6005 at step 6015 and displayed as part of the structured presentation 6020. Such new values of attributes can be constructed based on the contents of the electronic documents in the electronic document collection 102 as described herein. The new value can be added to the pre-existing structured presentation 6005 at step 6025 and displayed as part of the structured presentation 6020. In particular, in the example shown, the new column 6030 includes a new collection of new attribute identifiers 308 (ie, attribute identifier 6035) that identify the new attributes and attribute values 307 (ie, attribute values 6040, 6045) that characterize the new attributes. Can be included. In fact, the pre-existing structured presentation 6005 content is merged with the content pulled from the electronic document collection 102.

  FIG. 61 is a flow diagram of a process 6100 for adding values to a structured presentation based on the content of documents in an electronic document collection. Process 6100 may be performed by a system of one or more computers that perform operations by executing one or more machine-readable instruction sets. For example, process 6100 may be performed by search engine 202 within system 200. Process 6100 may be performed alone or in conjunction with other activities. For example, process 6100 may be performed as part of one or more of processes 700, 600, 800 (FIGS. 6, 7, 8).

  In some implementations, the process 6100 may be performed in response to receiving input from, for example, a user or other system or process that triggers creation of a new structured presentation or update of a structured presentation. For example, process 6100 may be performed in response to a user request to add one or more new attributes to structured presentation 106. As another example, process 6100 may be performed by a search engine, eg, search engine 202 (FIG. 2) in response to receiving a search query.

  The system executing process 6100 may receive an instance and attribute specification in the structured presentation (step 6105). A structured presentation may be a new or pre-existing structured presentation. For example, the system can receive a search query that specifies an instance or category of instances characterized in a structured presentation.

  As another example, a user can interact with a pre-existing structured presentation to specify instances and / or attributes. Instances and / or attributes can be specified internally or manually by user interaction with a pre-existing structured presentation. In the intrinsic specification, the instance and attribute associated with the cell are specified by user interaction with the cell by using the system arrangement of the instance and the attribute identifier in the structured display.

  In contrast, with manual designation, the user manually identifies which cells contain the instance and attribute identifiers associated with the cells. For example, a user can enter a search query into a cell that specifies an array of instance identifiers, attribute identifiers, or both in a structured presentation. For example, a search query that contains the formula “(CELL_1, CELL_2)” is associated with the attribute specified in cell “CELL_2” of the instance identified by cell “CELL_1” and for this attribute of this instance You can specify to perform a search. Such manual specification of instance and attribute identifiers is a spreadsheet table where the location of the instance and attribute identifiers is obvious to the user but is unknown to the data processing device presenting the structured presentation. It is particularly useful in structured presentations such as Other examples of user interaction with pre-existing structured presentations for specifying one or more instances and attributes are detailed below.

  The system executing process 6100 may construct one or more value propositions from the documents in the electronic document collection for the specified attributes of the instance (step 6110). Value proposals can be configured for attributes specified in a number of different ways. For example, in one implementation, the system can construct value suggestions from documents in electronic document collection 102 by performing a search using a search query that is constructed using specified instances and attributes. . For example, a value proposal may be constructed by finding a document that includes a structured component related to a specified instance and attribute, eg, as described in FIGS. 52-60 and the associated text.

  As another example, a search query may require that a specified instance and attribute identifier be found in a language pattern indicating that a value characterizing the instance attribute is likely to appear. Examples of such patterns are "the <attribute> of <entity> is", "<entity> with an <attribute> of", "<entity> has an <attribute> of", "<entity> 's <attribute> is ”, and the like. Such a pattern can be used to extract value proposals from text content in an electronic document.

  The system performing process 6100 may provide one or more value suggestions to the user (step 6115). For example, a list of value propositions can be displayed for the user on the same display screen that displays a pre-existing structured presentation. Displaying a list of value propositions can be performed before selecting values for addition to a pre-existing structured presentation.

  As another example, in some implementations, the value proposal can be hidden in the structured presentation along with the search information and interactive elements. Examples of such implementations are further described below.

  The system performing process 6100 may receive a user selection of value propositions presented in a structured display (step 6120). For example, the interactive element can interact with the user to receive one or more user inputs (eg, mouse clicks, key presses, or other user inputs) that select a value proposition. In some implementations, interactive elements can be hidden in structured presentations, as described further below. The system performing process 6100 may also add the selected value to the structured presentation (step 6125) and display the selected value in the structured presentation.

  FIG. 62 is a schematic representation of a structured presentation with the search interface hidden, ie, structured presentation 6200. The search interface can include search information, one or more search interactive elements, or a combination thereof. An interactive element is a component of a graphical user interface that can interact with a user, for example, can receive input commands. The search interactive element and the search information are related to the search. Search is the process of finding information in an electronic document collection. The search interface, for example, information indicating whether a search is available to write a value to a structured presentation, interactive elements that allow the user to indicate that such a search should be performed, at the time of the search A display that identifies the identified electronic document, an interactive element that allows the user to select an electronic document from multiple electronic documents to write values to the structured presentation, or of these and other functions Combinations can be included.

  The structured presentation 6200 may be any form of structured presentation, including any of the structured presentations described above. For example, structured presentation 6200 may be a data table displayed within a spreadsheet framework, as shown. The structured presentation 6200 data table includes a collection of rows 302 and columns 304. Each row 302 includes each instance identifier 306, and each column 304 includes each attribute identifier 308. The arrangement and positioning of instance identifier 306 and attribute identifier 308 in row 302 and column 304 associates each cell of the spreadsheet framework in which structured presentation 6200 is displayed with the instance and attribute. For example, cell 6205 in structured presentation 6200 is associated with an instance identified as “Tesla Roadster” and an attribute identified as “mpg”. Cell 6210 in structured presentation 6200 is associated with an instance identified as “Chevy Volt” and an attribute identified as “range”. Cell 6215 in structured presentation 6200 is associated with an instance identified as “Myers NmG” and an attribute identified as “top speed”. Cell 6220 in structured presentation 6200 is associated with an instance identified as “Myers NmG” and an attribute identified as “mpg”.

  Instances, attributes, and associations between cells such as cells 6205, 6210, 6215, 6220 can be used to receive instance and attribute specifications from the user. For example, receiving a user interaction selecting cell 6220 can be considered an input specifying an instance identified as “Myers NmG” and an attribute identified as “mpg”. User interaction to select a cell may include, for example, receiving input to position the cursor 6225 on the cell, the user clicking on the cell, or the like. In some implementations, the selection of a cell can be represented by positioning visual indicia such as ambient highlights 6230 in or around the cell.

  In the illustrated implementation, upon selection, the selected cell 6220 does not contain the value 307. There are several reasons for this. For example, the structured presentation 6200 may be a new structured presentation that has not yet been written with values. As another example, structured presentation 6200 may be a pre-existing structured presentation with values removed. As yet another example, the structured presentation 6200 may be a pre-existing structured presentation that, for some reason, derives previous values from a source document that can no longer function as a source of values.

  FIG. 63 is a diagram of another structured presentation 6300 where the search interface is hidden. In the context of the structured presentation 6200, the structured presentation 6300 includes the value 307 in the selected cell 6220. There are several reasons for this. For example, cell 6220 may have been automatically filled with value 307, for example, in response to receiving a search query. As another example, writing to the cell 6220 may be performed by a user manually operating the cell 6220 and inputting a value. As yet another example, cell 6220 may be filled with value 307 in response to a user specifying an instance, attribute, or both associated with cell 6220 (either implicitly or manually). There is sex. In any case, selection of cell 6220 specifies an instance identified as “Myers NmG” and an attribute identified as “mpg” associated therewith.

  FIG. 64 is a schematic diagram of another structured presentation 6400 where the search interface is hidden. Structured presentation 6400 includes visual indicia 6405. Visual indicia 6405 visually indicates that hidden search information or interactive elements are accessible from structured presentation 6400.

  In the illustrated implementation, each visual indicia 6405 is found in a separate cell, such as cell 6205, 6210, 6215, 6220. The positioning and arrangement of visual indicia 6405 in the cell (and concomitantly, the positioning and arrangement of visual indicia 6405 relative to instance identifier 306 and attribute identifier 308 in row 302 and column 304) are as follows: As further described, hidden search information and relevance and function of interactive elements can be visually shown.

  In operation, user interaction with structured presentations 6200, 6300, 6400 can trigger a hidden search interface. As described above, the search interface can include search information, search interactive elements, or both. The search interface can be hidden in the structured presentation in that the search information and interactive elements need not always be visible in the structured presentation. Rather, the hidden search interface can be completely or partially hidden from view while the structured presentation is in several states. For example, the hidden search interface can be hidden while the viewer is likely to be reviewing other information content in the structured presentation. This concealment can expand the portion of the structured presentation that can be used to present other information and reduce the scatter that is visible to improve the readability of the structured presentation.

  FIG. 65 is an illustration of a display element 6500 in which a previously hidden search interface is presented. In some implementations, the display element 6500 can be presented in response to user interaction with the structured presentation itself. Display element 6500 may “pop up” in front of a structured presentation (structured presentation 6200, 6300, 6400, etc.) and present search interactive element 6505 in window 6510 in response to user interaction with cell 6220. it can. For example, search interactive element 6505 and window 6510 may be presented in response to a user clicking cell 6220 using a mouse. Search interactive element 6505 is a hyperlink that includes text indicating that “more options ...” is available to write to cell 6220.

  FIG. 66 is an illustration of a display element 6600 where a previously hidden search interface is presented. In some implementations, the display element 6600 can be presented in response to user interaction with the structured presentation itself. In addition to search interactive element 6505, display element 6600 presents source identifier 6605 in window 6510 in response to user interaction with cell 6220. Source identifier 6605 includes text or other information identifying the electronic document that is the source of value 307 to be written to cell 6220. The source document identified by source identifier 6605 may be a document identified as a result of a pre-search. In some implementations, the source identifier 6605 can also include a hyperlink to the source document.

  FIG. 67 is an illustration of a display element 6700 on which a previously hidden search interface is presented. In some implementations, the display element 6700 can be presented in response to user interaction with the structured presentation itself. In addition to search interactive element 6505 and source identifier 6605, display element 6700 presents snippet 6705 in window 6510 in response to user interaction with cell 6220. The snippet 6705 is text or other information that describes the background status of the value in the electronic document that is the source of the value 307 to be written to the cell 6220.

  FIG. 68 is an illustration of a display element 6800 where a previously hidden search interface is presented. In some implementations, the display element 6800 can be presented in response to user interaction with the structured presentation itself or in response to user interaction with a previously hidden search interactive element 6505. Display element 6800 includes a header 6802, an associated source selection area 6805, and a matching source selection area 6810. The header 6802 may include text or other information that identifies the cell to which the value is added. In the illustrated implementation, cell 6220 is identified by the attribute and value characterized by value 307 in cell 6220 (ie, Myers NmG: mpg).

  The associated source selection area 6805 comprises information and interactive elements that allow the user to specify the relevance of the source electronic document for the specified instance and attribute to be used when selecting values to be written to the structured presentation. be able to. In the illustrated implementation, the user can specify that a single “most relevant” document is the single source of values to write to the structured presentation. The relevance of a document can characterize the degree to which the document matches, for example, attributes and instances defined by the search.

  In the illustrated implementation, the associated source selection area 6805 includes a header 6815, a selection widget 6820, a value identifier 6825, and a source identifier 6830. The header 6815 allows the user to specify that the most relevant electronic document is used as the source of values to be written to the cell identified in the header 6802 by using the associated source selection area 6805. Include text or other information to identify things. Using the selection widget 6820, the user can choose to use the most relevant document as the source of values to write to the structured presentation. Value identifier 6825 includes text or other information that identifies a value derived from the currently most relevant document. The source identifier 6830 includes text or other information that identifies the currently most relevant electronic document. In some implementations, the source identifier 6830 can also include a hyperlink to the currently most relevant document. Since the most relevant documents can change over time, the values identified by value identifier 6825 and the document identified by source identifier 6830 can also change over time.

  The matching source selection area 6810 may comprise information and interactive elements that allow the user to specify that the source electronic document be used consistently when selecting values to write to the structured presentation. In the illustrated implementation, the user can select a document from among three candidate documents to specify that document to be used consistently as a source of values to write to the structured presentation.

  In the illustrated implementation, the consistent source selection area 6810 includes a header 6835, a collection of selection widgets 6840, a collection of value identifiers 6845, and a collection of source identifiers 6850.

  Header 6835 is text or other information that identifies that the user can specify that the source electronic document should be used consistently when selecting values using the associated source selection area 6805 including. Selection widget 6840 allows the user to select a document to be used consistently. In the illustrated implementation, the user can select a document from three different documents. The value identifier 6845 includes text or other information that identifies a current value that can be derived from a particular document for writing to a structured presentation. The source identifier 6850 includes text or other information that identifies the electronic document from which the value identified by the value identifier 6825 is derived. In some implementations, the source identifier 6850 can also include a hyperlink to the electronic document from which the value identified by the value identifier 6825 is derived.

  Both the relevance of an electronic document and the value in the electronic document can change over time. For example, a person who adds an electronic document to an electronic document collection can change the content of the electronic document such that the relevance of the document to instances and attributes changes. As another example, a person who adds an electronic document to an electronic document collection can change the values used to characterize the attributes of the instance. The headers 6815, 6835 may include text or other information that identifies the nature of the change that may occur. For example, in the illustrated implementation, header 6815 is most relevant when the user specifies that the relevance of the source electronic document to the specified instance and attribute is used in selecting values to write to the structured presentation. Contains text that identifies both likely documents and attribute values can change. As another example, in the illustrated implementation, the header 6835 may change the value of the attribute when the user specifies that the source electronic document is used consistently when selecting values to write to the structured presentation. Include text that states.

  FIG. 69 is an illustration of a display element 6900 where a previously hidden search interface is presented. In some implementations, the display element 6900 can be presented in response to user interaction with the structured presentation itself or in response to user interaction with a previously hidden search interactive element 6505. In addition to headers 6802, 6815, 6835, selection widgets 6820, 6840, value identifiers 6825, 6845, and source identifiers 6830, 6850, display element 6900 includes a collection of snippets 6905 and a collection of search interactive elements 6910. Each snippet 6905 is text or other information that describes the background status of each value identified by the value identifiers 6825, 6845 in the electronic document that is the source of the identified value. Search interactive element 6910 is a hyperlink that allows the user to navigate to each electronic document that is the source of the value identified by each value identifier 6845.

  FIG. 70 is a diagram of a display element 7000 on which a previously hidden search interface is presented. In some implementations, the display element 7000 can be presented in response to user interaction with the structured presentation itself, or in response to user interaction with a previously hidden search interactive element 6505. In addition to headers 6802, 6815, 6835, selection widgets 6820, 6840, value identifiers 6825, 6845, and source identifiers 6830, 6850, snippets 6905, and search interactive element 6910, display element 6900 includes search trigger 7005. A search trigger 7005 is an interactive element that triggers a search of the electronic document collection. When displayed with a description of the results of a previous search (for example, value identifier 6825, 6845, source identifier 6830, 6850, snippet 6905, and search interactive element 6910), search trigger 7005 allows the user to You can show that you are not satisfied with the search results. In some implementations, the search triggered by search trigger 7005 may be a “full text search” performed using a general purpose search engine, such as a Google search engine. In some implementations, the search engine may be presented with a query that is automatically generated using the instances and attributes specified by the previous user interaction.

  As illustrated in FIGS. 65-70, the search information that was previously hidden and the nature of the user interaction that triggers the display of the interactive element can determine the category of search information and interactive elements that are displayed. For example, user interaction specifying a single cell within a structured presentation can trigger the presentation of search information and interactive elements related to writing values to the same cell. In other implementations, user interaction with a column, row, or other collection of cells may trigger the presentation of search information and interactive elements associated with writing values to the collection of cells. For example, using user interaction with a column, the user can specify that the value to be written to the column is consistently derived from a single source document or group of source documents. As another example, using user interaction with a row, the user can specify that the value to write to that row is derived from the source document most relevant to the instance and its attributes.

  FIG. 71 is a flow diagram of a process 7100 for adding values to a structured presentation by extracting values from the content of documents in an electronic document collection. Process 7100 may be performed by a system of one or more computers that perform operations by executing one or more machine-readable instruction sets. For example, process 7100 can be performed by search engine 202 in system 200. Process 7100 may be performed alone or in conjunction with other activities. For example, process 7100 may be performed as part of steps 6105, 6115, and 6120 of process 6100 (FIG. 61).

  The system performing process 7100 may receive data characterizing user interaction specifying one or more cells of the structured presentation (step 7105). A structured presentation may be a new or pre-existing structured presentation. Concomitantly, one or more attributes and instances can be specified by interactive operations on one or more cells, as described above.

  The system performing process 7100 may determine whether the one or more values to write to the one or more cells are the result of a pre-search of the electronic document collection (step 7110). Such a determination accesses a data storage device that stores not only the information that characterizes the information that is visible in the structured presentation, but also the information that characterizes the pre-search performed to write to the structured presentation. Can be done. The stored information that characterizes the pre-search is, for example, information that indicates that the search was actually performed, the URL of the source document in the result set of the previous search, and a snippet that characterizes the background status of the values in the source document Can be included.

  If the system executing process 7100 determines that the value was obtained as a result of a previous search, the system can present search information that characterizes the prior search (step 7115). Such information can include, for example, information identifying the source document in the result set from which the value was derived, a snippet characterizing the background status of the value in the source document, and a hyperlink to the source document. For example, the system can present search information that characterizes a single source document in a presentation, such as display elements 6600, 6700 (FIGS. 66, 67). As another example, the system may have multiple source documents in a presentation, such as display elements 6800, 6900, 7000 (FIGS. 68, 69, 70) (sources having values that are different from values that are visibly written to a structured presentation) Search information can be presented (including documents).

  In some implementations, the system performing process 7100 can transition between presenting search information for a single source document and search information for multiple source documents in response to interacting with the user. For example, the system can receive user interaction with a search interactive element, such as search interactive element 6505, and transition between display elements 6600, 6700 and display elements 6800, 6900, 7000 (FIGS. 65-70).

  In some implementations, the system performing process 7100 can perform a new search and provide information characterizing one or more electronic documents in the result set obtained by the new search (step 7120). ). The information to be characterized includes, for example, an electronic document name and URL, an electronic document snippet, an electronic document summary, or the like. The result set characterizes a single source document in a presentation such as display elements 6600, 6700 (Figures 66, 67) or multiple source documents in a presentation such as display elements 6800, 6900, 7000 (Figures 68, 69, 70) be able to. In some implementations, the system can transition between presenting search information for a single source document and search information for multiple source documents in response to interacting with the user. For example, the system can receive user interaction with a search interactive element, such as search interactive element 6505, and transition between display elements 6600, 6700 and display elements 6800, 6900, 7000 (FIGS. 65-70).

  FIG. 72 is an illustration of a display element 7200 that presents a previously hidden search interface. In some implementations, the display element 7200 can be presented in response to user interaction with the structured presentation itself or in response to user interaction with a previously hidden search interactive element 7005. Display element 7200 can receive values resulting from a search, eg, a search performed using a general purpose search engine. Values received by the display element 7200 can be added to the structured presentation to characterize the attributes of the instance. The display element 7200 includes a header 7205, an instance identifier 7210, an attribute identifier 7215, a value input element 7220, a value use trigger element 7225, and a presentation closing element 7230.

  The header 7205 is text or other information that describes that the display element 7200 can receive the value of an instance attribute. The header 7205 may prompt the user to input a value obtained as a result of the search. For example, the header 7205 may be text that asks whether the search was successful.

  The instance identifier 7210 is text or other information that identifies the instance, or category of instances, characterized by a value entered using the display element 7200. In the illustrated implementation, the instance identifier 7210 is text that identifies the instance “China”. The attribute identifier 7215 is text or other information that identifies the attribute of the instance identified by the instance identifier 7210. The attribute identified by the attribute identifier 7215 can be characterized by the value received by the display element 7200. In the illustrated implementation, the attribute identifier 7215 is text that identifies the attribute “Population”. Together, the instance identifier 7210 and the attribute identifier 7215 identify that the Chinese population is characterized by the value received using the display element 7200.

  The value input element 7220 is an interactive element that allows a user to specify a value that characterizes the attribute identified by the attribute identifier 7215 of the instance identified by the instance identifier 7210. The value input element 7220 may be a text input field, for example.

  The value usage trigger element 7225 triggers the user to use the value entered in the value input element 7220 to characterize the attribute identified by the instance attribute identifier 7215 identified by the instance identifier 7210 in the structured presentation. An interactive element that enables. The value usage trigger element 7225 may be, for example, a button that includes text identifying that the value entered in the value input element 7220 is used in the structured presentation as a result of user interaction with the value usage trigger element 7225. .

  Presentation close element 7230 is an interactive element that allows the user to close display element 7200. The display element 7200 is an attribute identified by the attribute identifier 7215 of the instance whose value entered in the value input element 7220 is identified by the instance identifier 7210 in the structured presentation in response to user interaction with the presentation close element 7230. Regardless of whether used to characterize. Presentation close element 7230 can be, for example, a button that includes text identifying that display element 7200 is closed by user interaction with presentation close element 7230.

  FIG. 73 is a diagram of a display element 7300 where a previously hidden search interface is presented. In some implementations, the display element 7300 can be presented in response to user interaction with the structured presentation itself or in response to user interaction with a previously hidden search interactive element 7005. The display element 7300 can receive the value of an instance attribute to add to the structured presentation. In addition to the header 7205, instance identifier 7210, attribute identifier 7215, value input element 7220, value usage trigger element 7225, and presentation close element 7230, the display element 7300 includes a source input element 7305 and a source input element identifier 7310.

  Source input element 7305 is an interactive element that allows the user to specify the source of the values that characterize the attribute identified by the instance attribute identifier 7215 identified by the instance identifier 7210. The source input element 7305 may be, for example, a text input field. The source input element identifier 7310 is text or other information that describes that the source input element 7305 can be used to specify the source of the value.

  In some implementations, the display elements 7200, 7300 may be displayed on the display screen for the user after a failed search. For example, the display elements 7200, 7300 can be displayed in response to receiving information from the user indicating that the user is not satisfied with the results of the previous search. For example, the display of display elements 7200, 7300 may be triggered by user interaction with search trigger 7005 (FIG. 70). As another example, the display elements 7200 and 7300 may be displayed when an automatic search for an instance attribute value is performed but a satisfactory result is not obtained.

  There are many reasons why a search for a value does not give satisfactory results. For example, the specification of attributes and / or instances may be incorrect due to misspellings or other errors, for example. As another example, an attribute or instance may be specified with respect to an unknown or indefinite value that is error-free. For example, the instance “suitable for Jim and Diane” is specified in terms of indeterminate values, namely the identity of Jim and Diane, and the nature of what is “suitable” to them. As another example, the instance “my car” is specified with an indeterminate value, ie, identification of the owner of the car being characterized.

  Searching for values may not provide satisfactory results because the electronic document obtained as a result of the pre-search does not function to provide values for the structured presentation. For example, the source document from which the value is derived may become unavailable. The source document may become unavailable, for example, when the party that added the source document retrieves the document from the electronic document collection. As yet another example, such a source document may still be available, but the value itself is not available in the source document. The value may become unavailable, for example, when the party who added the source document to the electronic document collection changes the content of the source document.

  FIG. 74 is an illustration of a display element 7400 where a previously hidden search interface is presented. Display element 7400 may be presented automatically in response to user interaction or in response to a trigger event. For example, the display element 7400 can be automatically posted in response to the pre-search not working.

  Display element 7400 includes a search interactive element 6505, a source identifier 6605, and an error message 7405 in window 6510. Search interactive element 6505 is a hyperlink that includes anchor text indicating that “more options ...” is available to search for a value to write to cell 6220. Source identifier 6605 is a collection of text that identifies the electronic document that is the source of value 307 to be written to cell 6220.

  Error message 7405 can include text or other information indicating that the results of the pre-search have stopped working. For example, error message 7405 can indicate that value 307 has become unavailable in the source document identified by source identifier 6605. Error message 7405 may describe the nature of the malfunction or may include information that simply indicates that an error has occurred. For example, in the illustrated implementation, error message 7405 indicates that the electronic document itself is still available, but the value is no longer available in the electronic document.

  FIG. 75 is a flow diagram of a process 7500 for adding values to a structured presentation based on the content of documents in an electronic document collection. Process 7500 may be performed by a system of one or more computers that perform operations by executing one or more machine-readable instruction sets. For example, process 7500 can be performed by search engine 202 in system 200. Process 7500 may be performed alone or in conjunction with other activities. For example, process 7100 may be performed as part of process 6100 (FIG. 61).

  The system performing process 7500 may receive an update trigger that triggers an update of one or more values of one or more cells of a pre-existing structured presentation (step 7505). The update trigger can be generated, for example, automatically in response to a certain period of time since the previous update, manually in response to user interaction, or the like. For example, as described above, user interaction with cell 6220 (FIGS. 62, 63, 64) may trigger a cell update. An update trigger may trigger an update of a value or a single cell, a collection of cells, or all cell values in a structured presentation. Concomitantly, as described above, an update trigger can specify one or more attributes of one or more instances.

  The system performing process 7500 may determine whether one or more pre-searches are not working to write values to the structured presentation (step 7510). Such a determination can be made by attempting to access a document from which values to write to the structured presentation are derived.

  If the system performing process 7100 determines that the pre-search is no longer working, the system can update the structured presentation with the content of one or more source documents identified in the pre-search. (Step 7515). The new value used to update the structured presentation need not be the same as the value already used to write to the structured presentation. Rather, the updated structured presentation can include values provided by the source electronic document along with its current content.

  If the system executing process 7100 determines that the pre-search has stopped working, the system can notify the user that the pre-search is inoperable (step 7520). For example, a display element such as display element 7400 can be used to notify the user that it is operational, and this gives the user the opportunity to perform a new search and write a value to the structured presentation. Can do. In some implementations, the system performs a new search, as described with reference to step 7120 of process 7100 (FIG. 71), and one or more in the result set obtained by the new search. Information that characterizes the electronic document can be provided.

  Thus, the display and subject matter herein can be implemented by a number of embodiments of the invention, including, for example, the embodiments numbered below.

  Embodiment 1: A machine-implemented method, which is a step of receiving descriptive data describing a pre-existing structured presentation, wherein the visual presentation of the pre-existing structured presentation is Visually present information in a system layout that conforms to the structured design, and the structured presentation provides an association between the instance and the values that characterize the attributes of the instance in the visual presentation of the structured presentation. A new representation associated with a pre-existing structured presentation by comparing the characteristics of the pre-existing structured presentation with the contents of the electronic document in an unstructured collection of electronic documents, represented by an array of Identify the electronic document that identifies the instance and identify the new instance Forming an extended structured presentation in addition to the structured presentation that is present in advance, an extended structured presentation and a step of outputting an instruction for presentation on a display device, machine implementation.

  Embodiment 2: Adding a new instance identifier includes configuring an instance proposal collection, providing an instance proposal collection to a user, and receiving a user selection of the new instance, wherein the new instance is The method of embodiment 1 included in the collection of instance proposals.

  Embodiment 3: Comparing the characteristics of a preexisting structured presentation with the content of the electronic document includes an electronic document that includes a structured component related to an instance identified in the preexisting structured presentation 3. The method of embodiment 2 comprising identifying documents in the collection.

  Embodiment 4: The step of configuring a collection of instance proposals includes an identifier of an instance identified in a pre-existing structured presentation and identifies a first document in an electronic document collection arranged according to a template Embodiment 2 comprising: identifying a second document arranged according to the template but related to the second instance; and including the second instance in the instance proposal collection. Method.

  Embodiment 5: Comparing the characteristics of a pre-existing structured presentation with the content of the electronic document includes electronic document collection including information about one or more instances in the pre-existing structured presentation 2. The method of embodiment 1, comprising identifying a document within.

  Embodiment 6: Comparing pre-existing structured presentation characteristics with content of an electronic document includes identifying a new instance in a stored collection of instance-attribute associations Form 1 method.

  Embodiment 7: The step of comparing the characteristics of a pre-existing structured presentation with the content of the electronic document is characterized by pre-existing structured presentation characteristics in the pre-existing structured presentation. Embodiment 2. The method of embodiment 1 comprising comparing to the attribute being

  Embodiment 8: The step of comparing the characteristics of a pre-existing structured presentation with the content of the electronic document includes the attributes used to characterize the instances in the pre-existing structured presentation. The method of embodiment 1, comprising the step of comparing with the content.

  Embodiment 9: The step of comparing the characteristics of a pre-existing structured presentation with the content of an electronic document is to electronically calculate the values of attributes used to characterize instances in the pre-existing structured presentation. The method of embodiment 1, comprising the step of comparing with the content of the document.

  Embodiment 10: The step of comparing the characteristics of a pre-existing structured presentation with the content of the electronic document includes comparing the category of the instance including the instance in the pre-existing structured presentation with the content of the electronic document. The method of embodiment 1, comprising the step of:

  Embodiment 11: The method of Embodiment 1, wherein the collection of electronic documents includes electronic documents available on the Internet, and the electronic documents include web pages.

  Embodiment 12: The method of Embodiment 1, wherein the extended structured presentation includes a table.

  Embodiment 13: The method of Embodiment 1, wherein the extended structured presentation includes a collection of cards.

  Embodiment 14: The method of Embodiment 1, further comprising visually presenting the augmented structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Embodiment 15: An apparatus comprising one or more machine-readable data storage media storing instructions operable to cause one or more data processors to perform operations, wherein the operations are unstructured Configure a collection of instance proposals based on the content of two or more documents in a structured electronic document collection, where the electronic document collection is permanent even if the electronic document format in the electronic document collection is limited A stage that is an unstructured collection in terms of providing a user with an instance proposal collection, receiving a user selection of the first instance in the instance proposal collection, and An identifier is presented on the display device to the structured pre In the presentation, where the presentation of the structured presentation visually presents the information in an organized form, and the structured presentation is the visual presentation of the structured presentation, and the instance and instance attributes And the step of representing the association between the values characterizing by an instance identifier and an array of values.

  Embodiment 16: The apparatus of Embodiment 15, wherein configuring the collection of instance proposals includes comparing pre-existing structured presentation characteristics with the content of the electronic documents in the electronic document collection.

  Embodiment 17: Configuring a collection of instance proposals includes identifying a document in an electronic document collection that includes a structured component related to the identified instance in a pre-existing structured presentation. Embodiment 16

  Embodiment 18: The step of configuring an instance proposal collection identifies a first document in an electronic document collection that is associated with an instance identified in a pre-existing structured presentation and arranged according to a template 17. The apparatus of embodiment 16, comprising: identifying a second document arranged according to the template but related to the second instance; and including the second instance in the instance proposal collection .

  Embodiment 19: An embodiment wherein configuring a collection of instance proposals includes identifying documents in an electronic document collection that includes identifiers of one or more instances in a pre-existing structured presentation. 16 devices.

  Embodiment 20: The apparatus of Embodiment 16, wherein configuring the collection of instance proposals includes identifying additional attributes used to characterize the instances in the pre-existing structured presentation.

  Embodiment 21: An embodiment wherein configuring a collection of instance proposals comprises comparing the value of an attribute used to characterize an instance in a pre-existing structured presentation with the value of the instance proposal. 16 devices.

  Embodiment 22: Configuring a collection of instance proposals includes identifying a category of instances that includes instances in a pre-existing structured presentation, and using the instances in that category of instances, instance proposals 17. The apparatus of embodiment 16, comprising the steps of:

  Embodiment 23: The apparatus of Embodiment 15, wherein configuring a collection of instance proposals includes identifying an instance proposal in a collection in which the association of instances and attributes is stored.

  Embodiment 24: Configuring the collection of instance proposals includes comparing the attributes characterized in the pre-existing structured presentation with the content of the documents in the unstructured electronic document collection Form 15 device.

  Embodiment 25: The apparatus of Embodiment 15, wherein the collection of electronic documents includes documents available on the Internet, and the electronic documents include web pages.

  Embodiment 26: The apparatus of Embodiment 15, wherein the structured presentation includes a table.

  Embodiment 27: The apparatus of Embodiment 15, wherein the structured presentation includes a collection of cards.

  Embodiment 28: A client device and one or more computers programmed to operate the client device interactively and perform operations, wherein these operations are pre-existing structured presentations. At the stage of receiving descriptive data to describe, the visual presentation of a pre-existing structured presentation visually presents information in a system layout that conforms to the structured design, and the structured presentation is structured An electronic document unstructured stage that represents the association between an instance and the values that characterize the attributes of the instance in the visual presentation of the presentation, represented by an instance identifier and an array of values, and pre-existing structured presentation characteristics Of electronic documents in a collection Identifying an electronic document that identifies a new instance associated with a pre-existing structured presentation compared to the content, and adding an identifier for the new instance to the pre-existing structured presentation to extend the structure One or more computers comprising: forming a structured presentation; and outputting instructions for presenting the enhanced structured presentation on a display device that is coupled in data communication with the client device; A system comprising:

  Embodiment 29: A client device and one or more computers programmed to operate the client device interactively and perform operations, wherein these operations are two in an unstructured electronic document collection The stage of constructing a collection of instance proposals based on the content of the document or higher, and the electronic document collection is unstructured in that the format of the electronic document in the electronic document collection is neither restrictive nor permanent A collection, a stage using a client device to provide an instance proposal collection to a user, a stage receiving a user selection of the first instance in the instance proposal collection, and data communication with the client device. Adding the identifier of the first instance proposal to the structured presentation presented on the display devices that are coupled in a visual manner, and the visual presentation of the structured presentation visually displays the information in an organized manner. A structured presentation comprising a stage, wherein an association between an instance and a value characterizing an instance attribute is represented by an instance identifier and an array of values in the visual presentation of the structured presentation, A system comprising a plurality of computers.

  Embodiment 30: One or more computers include a server operable to interact with a client device through a data communication network, and the client device is operable to interact with the server as a client. 30. The system of embodiment 29, wherein the client device comprises a personal computer executing a web browser, the personal computer comprising a display device.

  Embodiment 31: A machine-implemented method, the method comprising receiving descriptive data describing a pre-existing structured presentation, wherein the visual presentation of the pre-existing structured presentation is Visually present information in a system layout that conforms to the structured design, and the structured presentation provides an association between the instance and the values that characterize the attributes of the instance in the visual presentation of the structured presentation. A new representation associated with a pre-existing structured presentation by comparing the characteristics of the pre-existing structured presentation with the contents of the electronic document in an unstructured collection of electronic documents, represented by an array of Identifying the electronic document that identifies the attribute and the identifier of the new attribute Machine implemented method comprising the steps of forming additional to extend structured presentation structured presentation are, and an extended structured presentation step of outputting an instruction for presentation on a display screen.

  Embodiment 32: Adding an identifier of a new attribute includes configuring a collection of attribute proposals, providing the attribute proposal collection to a user, and receiving a user selection of the new attribute, wherein the new attribute is 32. The method of embodiment 31 included in a collection of instance proposals.

  Embodiment 33: The step of configuring the attribute proposal collection identifies a first document in the electronic document collection that is related to the instance identified in the pre-existing structured presentation and arranged according to the template. 33. The method of embodiment 32 comprising the steps of: adding attributes used in the first document to characterize instances in the attribute proposal collection.

  Embodiment 34: The step of comparing the characteristics of the preexisting structured presentation with the contents of the electronic document includes an electronic document that includes a structured component related to the instance identified in the preexisting structured presentation. 32. The method of embodiment 31, comprising identifying documents in the collection.

  Embodiment 35: The step of comparing pre-existing structured presentation characteristics with the content of the electronic document includes information about one or more instances in the pre-existing structured presentation. 32. The method of embodiment 31, comprising identifying a document within.

  Embodiment 36: Comparing pre-existing structured presentation characteristics with the content of an electronic document includes identifying new attributes in a stored collection of instance-attribute associations The method of form 31.

  Embodiment 37: The step of comparing the characteristics of the preexisting structured presentation with the contents of the electronic document compares the instance characterized in the preexisting structured presentation with the contents of the electronic document. The method of embodiment 31, comprising the steps.

  Embodiment 38: The step of comparing the characteristics of a preexisting structured presentation with the content of the electronic document identifies additional instances related to the instances identified in the preexisting structured presentation The method of embodiment 31, comprising the steps.

  Embodiment 39: The step of comparing the characteristics of a preexisting structured presentation with the content of the electronic document is to use an attribute or attribute value used to characterize an instance in the preexisting structured presentation 32. The method of embodiment 31, comprising comparing the to the content of the electronic document.

  Embodiment 40: The step of comparing the characteristics of a pre-existing structured presentation with the content of the electronic document includes comparing the category of the instance including the instance in the pre-existing structured presentation with the content of the electronic document. 32. The method of embodiment 31, comprising the step of:

  Embodiment 41: The method of Embodiment 31, wherein the collection of electronic documents includes electronic documents available on the Internet, and the electronic documents include web pages.

  Embodiment 42: The method of Embodiment 31, wherein the extended structured presentation includes a table.

  Embodiment 43: The method of Embodiment 31, wherein the extended structured presentation includes a collection of cards.

  Embodiment 44: The method of Embodiment 31, further comprising visually presenting the augmented structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Embodiment 45: An apparatus comprising one or more machine-readable data storage media storing instructions operable to cause one or more data processors to perform operations, wherein the operations are unstructured Configure a collection of attribute proposals based on the content of two or more documents in a structured electronic document collection, where the electronic document collection is permanent even if the electronic document format in the electronic document collection is limited A stage that is an unstructured collection, a stage that provides the user with an attribute proposal collection, a stage that receives a user selection of the first attribute in the collection of attribute proposals, and Adding an identifier to a structured presentation presented on a display screen, the structured The visual presentation of the presentation visually presents the information in an organized manner, and the structured presentation provides an association between the instance identifier and the value that characterizes the instance attributes in the structured presentation presentation. A device comprising steps represented by an array of values.

  Embodiment 46: The apparatus of Embodiment 45, wherein configuring the collection of attribute proposals includes comparing pre-existing structured presentation characteristics with the content of the electronic document in the electronic document collection.

  Embodiment 47: Configuring a collection of attribute proposals includes identifying documents in an electronic document collection that includes a structured component that characterizes an instance identified in a pre-existing structured presentation. Embodiment 46. The apparatus of embodiment 46.

  Embodiment 48: The step of constructing the attribute proposal collection identifies the first document in the electronic document collection that is related to the identified instance in the pre-existing structured presentation and arranged according to the template. 47. The apparatus of embodiment 46 comprising the steps and including attributes used to characterize instances in the attribute proposal collection.

  Embodiment 49: An embodiment wherein configuring a collection of attribute proposals includes identifying documents in an electronic document collection that includes information about one or more instances in a pre-existing structured presentation. 46 devices.

  Embodiment 50: The step of comparing the characteristics of a pre-existing structured presentation with the content of the electronic document is comparing the instance identified in the pre-existing structured presentation with the content of the electronic document. 47. The apparatus of embodiment 46, comprising:

  Embodiment 51: The step of comparing the characteristics of a pre-existing structured presentation with the content of the electronic document is to use an attribute or attribute value used to characterize an instance in the pre-existing structured presentation 47. The apparatus of embodiment 46, comprising comparing the to the content of the electronic document.

  Embodiment 52: The step of constructing a collection of attribute proposals is used to identify a category of instances that include instances in a pre-existing structured presentation and to characterize instances within that category of instances. 47. The apparatus of embodiment 46, comprising: constructing a collection of attribute proposals from the attributes.

  Embodiment 53: The apparatus of Embodiment 45, wherein configuring the collection of attribute proposals includes identifying an attribute proposal in the collection in which the association of instances and attributes is stored.

  Embodiment 54: The apparatus of Embodiment 45, wherein the collection of electronic documents includes electronic documents available on the Internet, and the electronic documents include web pages.

  Embodiment 55: The apparatus of Embodiment 45, wherein the structured presentation includes a table.

  Embodiment 56: The apparatus of Embodiment 45, wherein the structured presentation includes a collection of cards.

  Embodiment 57: A client device comprising a display screen and one or more computers programmed to operate the client device interactively and perform operations, the operations are pre-existing At the stage of receiving descriptive data describing the structured presentation, the visual presentation of the pre-existing structured presentation visually presents information in a system layout that conforms to the structured design, and the structured presentation The electronic representation of the stage and the characteristics of a pre-existing structured presentation, where the visual presentation of the structured presentation represents the association between the instance and the values characterizing the attributes of the instance by an instance identifier and an array of values. Unstructured document Identifying an electronic document that identifies a new attribute associated with a pre-existing structured presentation compared to the content of the electronic document in the collection, and a structured presentation pre-existing with an identifier for the new attribute A system comprising one or more computers, including: adding to form an extended structured presentation; and outputting instructions for presenting the extended structured presentation on a display screen.

  Embodiment 58: A client device comprising a display screen and one or more computers programmed to interact with the client device and perform operations, the operations comprising an unstructured electronic document collection Constructing a collection of attribute proposals based on the content of two or more documents in an electronic document collection, where the format of the electronic documents in the electronic document collection is neither restrictive nor permanent An unstructured collection, providing an attribute proposal collection to the client device, receiving a selection of the first attribute in the collection of attribute proposals from the client device, and an identifier for the first attribute proposal On the display screen In addition to the structured presentation shown, the visual presentation of the structured presentation visually presents the information in an organized form, and the structured presentation is instantiated with the visual presentation of the structured presentation. A system comprising one or more computers including a stage, wherein an association between values characterizing an instance attribute is represented by an instance identifier and an array of values.

  Embodiment 59: A machine-implemented method, the method comprising receiving descriptive data describing a pre-existing structured presentation, wherein the visual presentation of the pre-existing structured presentation is Visually present information in a system layout that conforms to the structured design, and the structured presentation provides an association between the instance and the values that characterize the attributes of the instance in the visual presentation of the structured presentation. A new representation associated with a pre-existing structured presentation by comparing the characteristics of the pre-existing structured presentation with the contents of the electronic document in an unstructured collection of electronic documents, represented by an array of Identifying the electronic document that identifies the value and structuring the new value pre-existing Comprising forming a new structured presentation in addition to the presentation, and the action of outputting an instruction for visually presenting the new structured presentation, the machine mounting method.

  Embodiment 60: The step of comparing the characteristics of the pre-existing structured presentation with the content of the electronic document includes identifying an identifier of the first instance that appears in the structured presentation in the first electronic document. 60. The method of embodiment 59, further comprising extracting the new value from the first electronic document.

  Embodiment 61: The step of adding a new value comprises identifying a collection of values of a first attribute of a first instance, a subset of one or more of the identified values, a first 60. The method of embodiment 59, comprising: determining the first attribute of the instance of as appropriately characterizing.

  Embodiment 62: The method of embodiment 61, wherein determining the subset of values as appropriate comprises dividing the values in the collection into groups.

  Embodiment 63: The method of Embodiment 61, wherein determining the subset of values as appropriate includes selecting the subset based at least in part on the number of values in the subset.

  Embodiment 64: The method of Embodiment 61, wherein determining the subset of values as appropriate includes selecting the subset based at least in part on values in the subset that satisfy a user-specified constraint condition.

  Embodiment 65: The method of embodiment 61, wherein determining the subset of values as appropriate includes selecting the subset based at least in part on the values in the subset being derived from the high quality document.

  Embodiment 66: Determining a subset of values as appropriate is based at least in part on steps derived from documents associated with other instances in the structured presentation for which the values in the subset already exist. 62. The method of embodiment 61, comprising selecting a subset.

  Embodiment 67: Determining a subset of values as appropriate is based at least in part on being derived from documents associated with other attributes in the structured presentation for which the values in the subset are pre-existing. 62. The method of embodiment 61, comprising selecting a subset.

  Embodiment 68: The method of embodiment 59, wherein the collection of electronic documents includes electronic documents available on the Internet, and the electronic documents include web pages.

  Embodiment 69 The method of embodiment 59, wherein the pre-existing structured presentation includes a table.

  Embodiment 70: The method of embodiment 59, wherein the pre-existing structured presentation comprises a collection of cards.

  Embodiment 71 The method of embodiment 59, further comprising visually presenting the new structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Embodiment 72: An apparatus comprising one or more machine-readable data storage media storing instructions operable to cause one or more data processors to perform an operation, the operations comprising: Receiving descriptive data describing the first instance, second instance, and first attribute, and the value of the first attribute of the first instance from two or more documents in the unstructured electronic document collection Extracting a first collection; extracting a second collection of first attribute values of a second instance from two or more documents of an unstructured electronic document collection; and Establishing a first subset of a collection of 1 as appropriately characterizing a first attribute of a first instance, and a second subset of a second collection of values. Display a structured presentation including the first value of the first subset and the second value of the second subset. Generating a machine readable instruction for the structured presentation, wherein the structured presentation represents an association between the instance and the value characterizing the attribute of the instance by an instance identifier and an array of values.

  Embodiment 73: Determining the first subset of values as appropriate includes dividing the values in the first collection into groups, each group being a subset of the first collection of values The device of embodiment 72, comprising:

  Embodiment 74: The embodiment of the embodiment 72 wherein determining the first subset of values as appropriate comprises selecting the first subset based at least in part on the number of values in the first subset. apparatus.

  Embodiment 75: The apparatus of embodiment 72, wherein determining the first subset of values as appropriate includes comparing the values in the first subset with user-specified constraints on those values.

  Embodiment 76: The apparatus of embodiment 72, wherein determining the first subset of values as appropriate includes determining that the values in the first subset have been derived from a high quality document.

  Embodiment 77: The method of Embodiment 72 wherein determining the first subset of values as appropriate includes determining that the values in the first subset are derived from a document associated with the second instance. apparatus.

  Embodiment 78: Determining a first subset of values as appropriate is derived from documents whose values in the first subset are related to other attributes that characterize both the first instance and the second instance. 73. The apparatus of embodiment 72, comprising determining that it has been.

  Embodiment 79: The description of the first instance includes the identifier of the first instance that appears in the pre-existing structured presentation, and the description of the second instance is the pre-existing structured [00117] 73. The apparatus of embodiment 72 comprising an identifier of a second instance that appears in the presentation.

  Embodiment 80: The apparatus of Embodiment 72, wherein the description of the first attribute includes a description of the new attribute that is added to the pre-existing structured presentation.

  Embodiment 81 The apparatus of Embodiment 72, wherein the unstructured electronic document collection includes electronic documents available on the Internet.

  Embodiment 82: The apparatus of Embodiment 72, wherein the structured presentation includes a table.

  Embodiment 83: The apparatus of Embodiment 72, wherein the structured presentation includes a collection of cards.

  Embodiment 84: The apparatus of embodiment 72 further comprising the step of visually presenting the structured presentation on the display screen, including physically transforming one or more elements of the display screen.

  Embodiment 85: A device and one or more computers programmed to interact with the device and perform operations, which describe pre-existing structured presentations At the stage of receiving descriptive data, a pre-existing visual presentation of a structured presentation visually presents information in a system layout that conforms to the structured design. An unstructured collection of electronic documents that describes the stage and characteristics of pre-existing structured presentations, representing the association between instances and values characterizing instance attributes in a visual presentation by an instance identifier and an array of values In advance compared to the content of the electronic document in Identifying an electronic document that identifies a new value associated with an existing structured presentation, adding a new value to the pre-existing structured presentation to form a new structured presentation, Outputting one or more computers including outputting instructions for visually presenting the new structured presentation on the device.

  Embodiment 86: a device and one or more computers programmed to interact with the device and perform operations, the operations comprising a first instance, a second instance, and a second instance Receiving descriptive data describing one attribute; extracting a first collection of first attribute values of a first instance from two or more documents of an unstructured electronic document collection; Extracting a second collection of values of the first attribute of the second instance from two or more documents of the unstructured electronic document collection, and a first subset of the first collection of values, Determining the first attribute of the first instance as appropriately characterizing, and the second subset of the second collection of values from the second instance. Machine-readable instructions for displaying a structured presentation including a first value of a first subset and a second value of a second subset; Generating a structured presentation comprising: representing an association between an instance and a value characterizing the attribute of the instance by an instance identifier and an array of values; and sending machine-readable instructions to the device. A system comprising one or more computers.

  Embodiment 87: A machine-implemented method, wherein the method is a step of displaying a structured presentation on a display device, wherein the structured presentation is an organized structured array that conforms to the structured design. Visually presenting the information, the structured presentation is displayed with steps that represent the association between the instance and the values characterizing the attributes of the instance in the visual presentation of the structured presentation by an instance identifier and an array of values. Receiving data including a first instance of the structured presentation and a first attribute specification that characterizes user interaction with the structured presentation, and previously hidden on the display device in response to receiving the data. Search interface that had been And wherein the search interface includes information or interactive elements that identify a position of a first value that characterizes the first attribute of the first instance in the electronic document collection. .

  Embodiment 88: Receiving data characterizing user interaction with a displayed structured presentation comprises receiving a manual user designation of a first instance and a first attribute associated with a cell in the structured presentation. The method of embodiment 87, comprising.

  Embodiment 89: Receiving data characterizing a user interaction comprises receiving data characterizing user interaction for a cell in a structured presentation, wherein the cell is an identifier of a first instance and a first attribute in the structured presentation. 90. The method of embodiment 87 wherein the array of cells is associated with the first instance and the first attribute.

  Embodiment 90: The method of embodiment 89, wherein receiving data characterizing user interaction with a cell comprises receiving data characterizing user interaction with an empty cell.

  Embodiment 91: Displaying a previously hidden search interface includes triggering a search of an electronic document collection and displaying an interactive element that can be selected by a user to identify a first value. The method of embodiment 87.

  Embodiment 92: Displaying a previously hidden search interface comprises displaying an interactive value input element that can be selected by a user to specify a value that characterizes the first attribute of the first instance. The method of embodiment 87, comprising.

  Embodiment 93: The method of embodiment 87, wherein displaying the previously hidden search interface comprises displaying a snippet characterizing the context of the first value in the first document of the electronic document collection. .

  Embodiment 94: The method of embodiment 87, wherein displaying the previously hidden search interface includes displaying a result of a pre-search of the electronic document collection to identify the first value.

  Embodiment 95: The first value appears in the structured presentation as a value characterizing the first attribute of the first instance, and displaying the previously hidden search interface is within the electronic document collection. 90. The method of embodiment 87, comprising: displaying an identifier of the first electronic document, wherein the first value is derived from the first electronic document.

  Embodiment 96: further comprising determining that the first electronic document does not function and cannot provide the first value, and displaying a visual indication indicating that the first document does not function The method of embodiment 87.

  Embodiment 97: The step of displaying a search interface that has been previously hidden is consistent with the first document regardless of changes in the relationship between the first document and the first instance and the first attribute. 90. The method of embodiment 87 comprising presenting the user with an option to select a value of 1.

  Embodiment 98: The step of displaying a previously hidden search interface comprises providing the user with an option to select a first value from the first document that is most relevant to the first instance and the first attribute. 90. The method of embodiment 87 comprising the step of presenting.

  Embodiment 99: further comprising searching an unstructured collection of electronic documents to identify the first value in response to user interaction with the search interface; and adding the first value to the structured presentation. The method of embodiment 87, comprising.

  Embodiment 100: The method of embodiment 87, wherein receiving a specification of a first instance and a first attribute comprises receiving a specification of a collection of attributes or a collection of instances.

  Embodiment 101: The method of embodiment 87, further comprising updating the display of the structured presentation over time.

  Embodiment 102: One or more computers programmed to interact with a client device and perform operations, wherein these operations are the first of a structured presentation displayed on a display device. Receiving data characterizing user interaction specifying a cell, where the structured presentation visually presents information in an organized structured array that conforms to the structured design, and the structured presentation is an instance Representing an association between the attribute and the value characterizing the instance attribute by an array of values in the cell; determining that a pre-search has been performed to write the first value to the first cell; In response to determining that the search has been performed, And a step of displaying the information characterizing the advance search on the device, provided with one or more computers, systems.

  Embodiment 103: Receiving data characterizing user interaction designating a first cell is receiving data characterizing manual user designation of a first instance and a first attribute associated with the first cell 102. The system of embodiment 102, comprising:

  Embodiment 104: The system of embodiment 102, wherein displaying information characterizing the pre-search includes displaying information identifying the electronic document from which the first value is derived.

  Embodiment 105: The system of embodiment 102, wherein displaying the information characterizing the pre-search includes displaying information identifying a collection of electronic documents that may have been derived from the first value.

  Embodiment 106: The system of embodiment 102, wherein displaying information characterizing the pre-search includes displaying information in a display element of a search interface that was previously hidden.

  Embodiment 107: The system of embodiment 102, wherein displaying the information characterizing the pre-search includes displaying information identifying the first electronic document in the electronic document collection from which the first value is derived. .

  Embodiment 108: These operations comprise determining that the first electronic document is inoperable and cannot provide the first value, and a visual indication indicating that the first document is inoperable 108. The system of embodiment 107, further comprising:

  Embodiment 109: The system of Embodiment 102, wherein these operations further comprise updating the display of values in the first cell of the structured presentation in response to user interaction.

  Embodiment 110: The system of embodiment 102, wherein displaying the information characterizing the pre-search comprises displaying a snippet characterizing the context of the first value in the first document of the electronic document collection.

  Embodiment 111: The system of embodiment 110, wherein the collection of electronic documents includes electronic documents available on the Internet, and the electronic documents include web pages.

  Embodiment 112 The system of embodiment 102, wherein the structured presentation includes a collection of cards.

  Embodiment 113: One or more computers programmed to interactively operate and perform operations on a client device comprising a display device, the operations displaying a structured presentation on the display device A structured presentation is a visual presentation of information in an organized, structured array that conforms to a structured design, and a structured presentation is a visual presentation of structured presentations, The first instance of the structured presentation that characterizes the stage and the user interaction with the displayed structured presentation, where the association between the values that characterize the attribute is represented by an instance identifier and an array of values And receiving data including the designation of the first attribute and displaying a search interface that was previously hidden on the display device in response to receiving the data, the search interface comprising: A system comprising: one or more computers comprising: an information or interactive element that identifies a location of a first value that characterizes a first attribute of a first instance in a document collection.

  These and other embodiments of the subject matter and functional operations described herein may be digital electronic circuits or computers including the structures disclosed herein and their structural equivalents. It can be implemented in software, firmware, or hardware, or a combination of one or more of these. Embodiments of the subject matter described herein are encoded on a computer storage medium for execution by one or more computer programs, ie, data processing devices, or for controlling the operation of data processing devices. Can be implemented as one or more modules of computerized computer program instructions. Alternatively, or in addition, the program instructions are propagated signals that are artificially generated signals that are generated to encode information for transmission to a suitable receiver device for execution by the data processor. For example, it can be encoded into a machine-generated electrical, optical, or electromagnetic signal. The computer storage medium may be a machine readable storage device, a machine readable storage substrate, a random or serial access memory device, or a combination of one or more of these.

  The term “data processing apparatus” encompasses any type of apparatus, device, and machine for processing data, including, for example, a programmable processor, a computer, or multiple processors or computers. The device may include dedicated logic, for example, an FPGA (Field Programmable Gate Array), or an ASIC (Application Specific Integrated Circuit). In addition to hardware, the device is code that creates an execution environment for the computer program of interest, eg, processor firmware, protocol stack, database management system, operating system, or one or more of these It is good also as what contains the code which constitutes a combination.

  A computer program (also called a program, software, software application, script, or code) is written in any form of programming language, including a compiled or interpreted language, or a declarative or procedural language, a stand-alone program, or It can be deployed in any form including modules, components, subroutines, or other units suitable for use in a computing environment. A computer program may correspond to a file in a file system, but need not be. A program can be part of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), a single file dedicated to the program of interest, or multiple It can be stored in a conditioned file (eg, a file that stores one or more modules, subprograms, or portions of code). A computer program is deployed to be executed on one computer, on one site, or on multiple computers distributed across multiple sites and interconnected by a communications network. sell.

  The process and logic flows described herein are one or more programmable that execute one or more computer programs to perform functions by manipulating input data and generating output. Can be executed by any processor. Processes and logic flows are also performed by dedicated logic circuits, such as FPGAs (Field Programmable Gate Arrays), or ASICs (Application Specific Integrated Circuits), and devices are also dedicated logic circuits such as FPGAs (Field Programmable Gate Arrays). ) Or ASIC (application specific integrated circuit).

  Processors suitable for the execution of computer programs include, for example, general purpose microprocessors, special purpose microprocessors, and any one or more processors of any type of digital computer. Generally, a processor will receive instructions and data from a read-only memory or a random access memory or both. The essential elements of a computer are a processor for executing or executing instructions and one or more memory devices for storing instructions and data. In general, a computer also includes one or more mass storage devices for storing data, eg, magnetic disks, magneto-optical disks, or optical disks, that receive data from or transfer data to them Or operably coupled to do both. However, the computer need not include such a device. In addition, a computer can be another device, such as a mobile phone, personal digital assistant (PDA), portable audio or video player, game console, global positioning system (GPS) receiver, or portable to name a few. It can be embedded in a storage device (eg, Universal Serial Bus (USB) flash drive).

  Computer readable media suitable for storing computer program instructions and data include, for example, semiconductor memory devices such as EPROM, EEPROM, and flash memory devices, magnetic disks such as internal hard disk or removable disks, magneto-optical disks, And all forms of non-volatile memory, media, and memory devices, including CD-ROM and DVD-ROM discs. The processor and the memory can be supplemented by, or incorporated in, dedicated logic circuitry.

  In order to exchange information with a user, the inventive embodiments described herein provide a display device for displaying information to the user, such as a CRT (cathode ray tube) or LCD (liquid crystal display) monitor. , And a keyboard and pointing device that a user can use to send input to the computer, such as a mouse or trackball. Other types of devices can also be used to interact with the user, for example, the feedback returned to the user can be any form of sensory feedback, such as visual feedback, audio feedback, or tactile feedback, and The input from the user can be received in any form, including acoustic, spoken, or tactile input. In addition, the computer responds to requests received from the web browser, for example by sending a document to the device that the user uses and receiving the document from that device, for example, a web page to the web browser on the user's client device. Information can be exchanged with the user.

  This specification contains many implementation-specific details, but these should not be construed as limitations on the scope of any invention or claim, but rather on a particular embodiment of a particular invention. It should be construed as an explanation of features that appear to be unique. Certain features that are described in this specification in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination. Furthermore, some features may be described as operating in a particular combination, and may be initially claimed as such, but one or more resulting from the claimed combination Features may be cut from the combination in some cases, and the claimed combination may be directed to a partial combination or modification of a partial combination.

  Similarly, operations are shown in a particular order in the drawings, as long as such operations are performed or performed in the particular order shown or in sequence. Should not be construed as requiring execution to achieve the desired result. Further, although various system components are separated in the above-described embodiments, it should not be understood that all such embodiments require such separation, and the program components and systems described It will be appreciated that generally can be integrated into a single software product or packaged into multiple software products.

  Particular embodiments of the subject matter described in this specification have been described. Other embodiments are within the scope of the following claims. For example, the actions recited in the claims can be performed in a different order and still achieve desirable results. By way of example, the processes shown in the accompanying drawings do not necessarily require the particular order shown, or order, to achieve the desired result.

100 system
102 Unstructured electronic document collection
104 display screen
106 structured presentation
108 Data communication infrastructure
110, 112, 114 Electronic documents

Claims (31)

A machine mounting method,
Receiving a machine-readable search query from a user;
Responding to the search query with instructions for presenting a structured presentation of instances associated with the search query to the user;
The visual presentation of the structured presentation represents an association between the instance and a value characterizing the instance attribute by an identifier of the instance and an array of the values;
The identifier and value of the instance are derived from two or more documents in an unstructured electronic document collection;
The machine-implemented method, wherein the electronic document collection is an unstructured collection in that the format of the electronic document in the electronic document collection is neither restrictive nor permanent. Responding to the search query comprises:
Identifying a first collection of electronic documents in the unstructured electronic document collection related to the instance;
Extracting the value of the attribute of the instance from the first collection of electronic documents;
2. The method of claim 1, comprising writing values extracted from two or more electronic documents to the structured presentation. Responding to the search query comprises:
Extracting a first value of a first attribute of a first instance from a first electronic document;
Extracting a second value of a second attribute of the first instance from a second electronic document;
Associating the first value and the second value with the first instance in a single record in the structured presentation;
2. The method according to claim 1, wherein the first attribute is different from the second attribute, and the first electronic document is different from the second electronic document. Responding to the search query comprises:
Extracting a first value of an attribute of a first instance from a first electronic document;
Extracting a second value of an attribute of a second instance from the first electronic document;
Associating the first value with the first instance in a first record;
Associating the second value with the second instance in a second record;
The method of claim 1, wherein the first instance is different from the second instance.   The method of claim 1, wherein the structured presentation includes a table.   The method of claim 1, wherein the structured presentation includes a collection of cards. Receiving a trigger to add a new instance to the structured presentation;
The method of claim 1, further comprising proposing a new instance to add to the structured presentation in response to the trigger. Further comprising receiving a constraint specification from the user;
The method of claim 7, wherein proposing a new instance includes proposing a new instance that satisfies the user-specified constraint condition. Receiving a trigger to add a new attribute to the structured presentation;
The method of claim 1, further comprising adding a new attribute to the structured presentation in response to the trigger. Receiving a user designation of the attributes of the new attribute;
2. The method of claim 1, further comprising writing the value of the attribute to the structured presentation based on the user-specified attribute.   The method of claim 1, wherein the unstructured electronic document collection includes electronic documents available on the Internet.   The method of claim 1, further comprising visually presenting the structured presentation on the display screen, including physically transforming one or more elements of the display screen. An apparatus comprising one or more machine-readable data storage media storing instructions operable to cause one or more data processors to perform an operation, the operation comprising:
Means for receiving descriptive data describing a pre-existing structured presentation;
Means for deriving the identifier of the first instance from the first website;
Means for deriving a first value of a first attribute of the first instance from a second website;
Means for adding the identifier of the first instance and the new value to the pre-existing structured presentation to form a new record in the new structured presentation;
Means for outputting instructions for visually presenting the new structured presentation;
The visual presentation of the pre-existing structured presentation visually presents information in a system arrangement that conforms to the structured design, the structured presentation includes a collection of records, each of which is the structure An apparatus for representing an association between an instance and a value characterizing an attribute of the instance in a visual presentation of a structured presentation by an identifier of the instance and an array of the values   The means for retrieving the identifier of the first instance from the first website compares characteristics of the pre-existing structured presentation with content of the pre-existing structured presentation 14. An apparatus according to claim 13, comprising means. The operation further comprises means for receiving an identifier of a second instance from the user;
14. The apparatus of claim 13, wherein the new structured presentation includes a second new record that presents the second instance in relation to a second value of the first attribute of the second instance. .   16. The apparatus of claim 15, wherein the operation further comprises means for receiving the second value from the user. Said operation is
Means for presenting a collection of candidate values to the user; and means for receiving a selection of the second value from the user;
The apparatus of claim 15, wherein the collection includes the second value. Said operation is
Means for identifying a collection of candidate values for the first attribute of the second instance;
16. The apparatus of claim 15, further comprising means for determining a certainty that the candidate value is correct for each of the candidate values.   14. The apparatus of claim 13, wherein the operation further comprises means for proposing a collection of new instances to add to the structured presentation.   20. The means for proposing the collection of new instances comprises means for comparing characteristics of the pre-existing structured presentation with content of the first website and the second website. The device described in 1.   20. The apparatus of claim 19, wherein means for proposing the collection of new instances comprises means for comparing machine readable search queries with content of the first website and the second website.   14. The apparatus of claim 13, wherein the means for deriving the first value from the second website comprises means for identifying that the second website includes a review.   14. The apparatus of claim 13, wherein the means for extracting the identifier from the first website comprises means for extracting the identifier directly from the first website.   14. The apparatus of claim 13, wherein the means for retrieving the identifier from the first website comprises means for extracting the identifier from a machine readable database that includes information extracted from the first website. The pre-existing structured presentation includes a table;
14. The apparatus of claim 13, wherein the record includes a row or column of the table. The pre-existing structured presentation includes a collection of cards;
14. The apparatus of claim 13, wherein the record includes individual cards in the collection.   14. The operation of claim 13, further comprising means for visually presenting the new structured presentation on the display screen, including means for physically transforming one or more elements of the display screen. Equipment. A system comprising a client device and one or more computers programmed to operate the client device interactively and perform operations,
Said operation is
Means for receiving descriptive data describing a pre-existing structured presentation;
Means for deriving the identifier of the first instance from the first website;
Means for deriving a first value of a first attribute of the first instance from a second website;
Means for adding the identifier of the first instance and the new value to the pre-existing structured presentation to form a new record in the new structured presentation;
Means for outputting instructions for visually presenting the new structured presentation to the client device;
The visual presentation of the pre-existing structured presentation visually presents information in a system arrangement that conforms to the structured design, the structured presentation includes a collection of records, each of which is the structure A system for representing an association between an instance and a value characterizing an attribute of the instance in a visual presentation of a structured presentation by an identifier of the instance and an array of the values. The one or more computers comprise a server operable to interactively operate the client device through a data communication network;
30. The system of claim 28, wherein the client device is operable to interact with the server as a client. A system comprising a client device and one or more computers programmed to operate the client device interactively and perform operations,
Said operation is
Means for receiving a machine readable search query from the client device;
Responding to the search query by sending instructions to the client device to present a structured presentation of instances associated with the search query;
The visual presentation of the structured presentation represents an association between the instance and a value characterizing the instance's attributes by an identifier of the instance and an array of values, the identifier of the instance and the value being non- Derived from two or more documents in a structured electronic document collection, the electronic document collection being an unstructured collection in that the format of the electronic document in the electronic document collection is neither restrictive nor permanent System. The one or more computers comprise a server operable to interactively operate the client device through a data communication network;
32. The system of claim 30, wherein the client device is operable to interact with the server as a client.