WO2021244657A1

WO2021244657A1 - Information stream extraction method, apparatus and device

Info

Publication number: WO2021244657A1
Application number: PCT/CN2021/098541
Authority: WO
Inventors: 德斯潘德S; 库玛庞卡; 乌玛尚卡尔希夫尚卡尔; 汉斯马库斯
Original assignee: 智慧芽信息科技(苏州)有限公司
Priority date: 2020-06-05
Filing date: 2021-06-07
Publication date: 2021-12-09
Also published as: CN113761214A

Abstract

An information stream extraction method, comprising: acquiring target entity information; extracting, from a data source, at least one piece of associated entity information corresponding to the target entity information and association relationship information between the target entity information and the associated entity information, the association relationship information comprising information describing an association direction and an association type between the entity information; and associating the target entity information with the associated entity information by using the association relationship information, so as to obtain an information stream corresponding to the target entity information. The various embodiments of the present description can greatly improve the accuracy and comprehensiveness of information extraction.

Description

Information flow extraction method, device and equipment

Technical field

This specification relates to the technical field of computer data processing, in particular, to an information flow extraction method, device and equipment.

Background of the invention

At present, there are some methods of information extraction and search, such as using keywords for information query. However, in view of the complex and changeable descriptions of many specific information in many data sources, and there are problems of inaccuracy and incompleteness, it is difficult to accurately and comprehensively extract the information that users need.

Summary of the invention

The purpose of the embodiments of this specification is to provide an information flow extraction method, device, and equipment, which can greatly improve the accuracy and comprehensiveness of information query.

This specification provides an information flow extraction method, device, and equipment that are implemented in the following ways:

An information flow display method, including:

The client sends an information flow acquisition request to the server, where the information flow acquisition request includes the input information acquired by the client;

The server receives the information flow acquisition request, and extracts one or more entity information from the input information as target entity information;

The server extracts at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes the direction of the association between the entity information and Information describing the type of association;

The server uses the association relationship information to associate the target entity information and the associated entity information, obtains the information flow corresponding to the target entity information, and sends the information flow to the client for the Display on the client.

On the other hand, the embodiment of this specification also provides an information flow extraction method, including:

Obtain target entity information;

At least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information are extracted from the data source, and the association relationship information includes an association direction and an association between the entity information Type description information;

Use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information.

In another embodiment of the method provided in this specification, the using the association relationship information to associate the target entity information with the associated entity information includes:

Obtain at least one associated entity information extracted from the data source and the association relationship information between the target entity information and the corresponding associated entity information, and use the association relationship information between the target entity information and the corresponding associated entity information to The target entity information is associated with the corresponding associated entity information to obtain the sub-information stream of the target entity information;

Using the target entity information as the reference information, link sub-information streams obtained from the same or different data sources to obtain the information stream corresponding to the target entity information.

In another embodiment of the method provided in this specification, when the association relationship information includes co-located entity description information, the associated entity information corresponding to the co-located entity description information is added to the target entity information, and updated The target entity information; wherein, the co-located entity description information includes description information in other expression forms that describe the associated entity information as target entity information;

Correspondingly, at least one associated entity information corresponding to the updated target entity information and the association relationship information between the target entity information and the associated entity information are extracted from the data source.

In another embodiment of the method provided in this specification, the method further includes:

Use the associated entity information extracted from the data source as the target entity information to extract the information flow, and obtain the information flow corresponding to the multiple target entity information;

Link the information streams corresponding to multiple target entity information with the corresponding target entity information as the reference information to obtain the information map.

In another embodiment of the method provided in this specification, the extracting at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, include:

Retrieve the data source where the target entity information is located;

Locate the context information where the target entity information is located in the data source;

When the context information includes a material process entity, a material name entity, and/or a material application entity, extract the associated entity information corresponding to the target entity information and the target entity information and the associated entity information according to the material process information association method The relationship information between.

Extracting the target entity information and the commodity name of the associated entity information from the data source;

The manufacturer or supplier corresponding to the target entity information or associated entity information is extracted according to the product name, and the manufacturer or supplier is associated with the target entity information or associated entity information.

Extract the target entity information and the parameter entity information of the associated entity information from the data source. The parameter entity information includes at least one of the material structure type entity, the process method entity, the material attribute entity, the unit entity or the measurement entity corresponding to one Entity information

The extracted parameter entity information is associated with the corresponding target entity information and associated entity information.

In another embodiment of the method provided in this specification, the method further includes: associating the sub-information stream of the target entity information with the corresponding data source.

Use the entity information in the information flow as an information node for interaction with the user, and use a visualization method to visualize the information flow;

Sending the processed information stream to the client, so that the user can view the information stream and trigger the information node through the client;

Based on the user's triggering operation on the information node, feedback other information associated with the information node to the client.

On the other hand, an embodiment of this specification also provides an information stream extraction device, including:

The first obtaining module is used to obtain target entity information;

The first extraction module is configured to extract at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes pair entity information Information describing the direction of the association and the type of association;

The first association module is configured to use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information.

In another embodiment of the device provided in this specification, the first association module includes:

The first associating unit is used to obtain at least one associated entity information extracted from the data source and the association relationship information between the target entity information and the corresponding associated entity information, and use the relationship between the target entity information and the corresponding associated entity information The association relationship information between each other associates the target entity information with the corresponding associated entity information, and obtains the sub-information stream of the target entity information;

The second associating unit is used to link the sub-information streams obtained from the same or different data sources using the target entity information as the reference information to obtain the information stream corresponding to the target entity information.

In another embodiment of the device provided in this specification, when the association relationship information includes colocation entity description information, the device further includes:

The update module is used to add the associated entity information corresponding to the co-located entity description information to the target entity information, and update the target entity information; wherein the co-located entity description information includes information describing the associated entity as the target entity Descriptive information in other forms of information;

The first extraction module is further configured to extract at least one associated entity information corresponding to the updated target entity information and the association relationship information between the target entity information and the associated entity information from the data source.

In another embodiment of the device provided in this specification, the device further includes:

The second extraction module is used to extract the information flow of each associated entity information extracted from the data source as the target entity information, and obtain the information flow corresponding to the multiple target entity information;

The second association module is used to link the information streams corresponding to multiple target entity information with the corresponding target entity information as the reference information to obtain the information graph.

In another embodiment of the device provided in this specification, the first extraction module includes:

The retrieval unit is used to retrieve the data source where the target entity information is located;

The positioning unit is used to locate the context information where the target entity information is located in the data source;

The extraction unit is configured to extract the associated entity information corresponding to the target entity information and the target entity according to the material process information association method when the context information includes material process entities, material name entities, and/or material application entities Information about the association relationship between information and associated entity information.

The third extraction module is used to extract the target entity information and the commodity name of the associated entity information from the data source;

The third association module is used to extract the manufacturer or supplier corresponding to the target entity information or the associated entity information according to the product name, and compare the manufacturer or supplier with the target entity information or the associated entity information Associated.

The fourth extraction module is used to extract the target entity information and the parameter entity information of the associated entity information from the data source. The parameter entity information includes at least a material structure type entity, a process method entity, a material attribute entity, a unit entity, or a measurement entity Entity information corresponding to one of the types;

The fourth association module is used to associate the extracted parameter entity information with corresponding target entity information and associated entity information.

In another embodiment of the device provided in this specification, the first association module further includes:

The third associating unit is used for associating the sub-information stream of the target entity information with the corresponding data source.

The visualization processing module is used to treat each entity information in the information flow as an information node for interacting with the user, and use a visualization method to visualize the information flow;

The first sending module is configured to send the processed information stream to the client, so that the user can view the information stream and trigger the information node through the client;

The second sending module is configured to feed back other information associated with the information node to the client based on the user's triggering operation on the information node.

On the other hand, the embodiments of this specification also provide an information flow extraction device. The device includes a processor and a memory for storing processor-executable instructions. When the instructions are executed by the processor, any one of or Multiple steps of the described method.

On the other hand, the embodiment of this specification also provides an information flow display method, which is applied to a server, and includes:

Receiving an information flow acquisition request sent by the client, where the information flow acquisition request includes input information acquired by the client;

Extract one or more entity information from the input information as target entity information;

Using the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information;

The information stream is sent to the client for display on the client.

On the other hand, the embodiment of this specification also provides a server, including:

The first receiving module is configured to receive an information flow acquisition request sent by the client, where the information flow acquisition request includes input information acquired by the client;

The fifth extraction module is used to extract one or more entity information from the input information as target entity information;

The sixth extraction module is configured to extract at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes pair entity information Information describing the direction of the association and the type of association;

A fifth association module, configured to use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information;

The third sending module is configured to send the information stream to the client for display on the client.

On the other hand, the embodiment of this specification also provides an information flow display method, which is applied to the client, including:

Send an information flow acquisition request to the server, where the information flow acquisition request includes the input information obtained by the client; so that the server receives the information flow acquisition request and extracts one or more entity information from the input information as the target entity Information; and, extracting at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes the relationship between the entity information Information describing the association direction and association type; and, using the association relationship information to associate the target entity information and the associated entity information, obtain the information flow corresponding to the target entity information, and send the information flow To the client;

Receive the information stream sent by the server and display it.

On the other hand, the embodiment of this specification also provides a client, including:

The fourth sending module is configured to send an information flow acquisition request to the server, where the information flow acquisition request includes the input information obtained by the client; so that the server receives the information flow acquisition request and extracts one or more information from the input information. Entity information as target entity information; and extracting at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, the association relationship information Including information describing the direction and type of association between entity information; and using the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information, And send the information stream to the client;

The second receiving module is used to receive the information stream sent by the server;

The first display module is used to display the information stream.

On the other hand, the embodiment of this specification also provides an information retrieval method applied to a server, including:

Receiving a search request sent by the client, where the search request includes input information obtained by the client;

Extracting entity information from the input information, where the entity information includes an entity type and an entity value;

The data source where the entity information is located is determined according to the associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information, and the association relationship information includes the direction of association between the entity information And information describing the type of association;

The search result is sent to the client, so that the client can display it.

The third receiving module is configured to receive a search request sent by the client, where the search request includes input information obtained by the client;

The seventh extraction module is used to extract entity information from the input information, where the entity information includes entity type and entity value;

The data source determining module is configured to determine the data source where the entity information is located according to the associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information, and the association relationship information includes Information describing the direction of association between entity information and the type of association;

The fifth sending module is configured to send the search result to the client, so that the client can display it.

On the other hand, the embodiment of this specification also provides an information retrieval method, which is applied to the client, including:

Send a search request to the server, the search request includes the input information obtained by the client; so that the server extracts the entity information from the input information, the entity information includes the entity type and the entity value; according to the entity information in the data source The associated entity information in and the association relationship information between the entity information and the associated entity information determine the data source where the entity information is located, and the association relationship information includes information describing the association direction and association type between the entity information; And sending the search result to the client;

Receive the retrieval result sent by the server and display it.

In another embodiment of the method provided in this specification, the information input interface displayed by the client includes an information input area, a first selection list, and/or a second selection list; the information input area is used to perform information Input; the first selection list and the second selection list are used for information selection, the first selection list includes material application or material structure category information; the second selection list includes entity type category information;

Correspondingly, the client terminal obtains input information based on the information input area, the first selection list, and/or the second selection list.

In another embodiment of the method provided in this specification, the first selection list and/or the second selection list adopts an interactive visualization format to display the information to be selected; wherein, the interactive visualization format means that the information to be selected is displayed through visualization Display the entity type or material structure, the category information of the material application in the form of, and determine the corresponding category and subcategory list selection format by receiving the trigger operation of each category information.

The sixth sending module is configured to send a search request to the server, the search request includes input information obtained by the client; so that the server extracts entity information from the input information, and the entity information includes entity type and entity value; The associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information determine the data source where the entity information is located, and the association relationship information includes the association direction between the entity information and Information describing the type of association; and sending the search result to the client;

The fourth receiving module is configured to receive the retrieval result sent by the server;

The second display module is used to display the retrieval result.

In another embodiment of the client provided in this specification, the client further includes:

The input interface display module is used to display an information input interface. The information input interface includes an information input area, a first selection list and/or a second selection list; the information input area is used for information input; the first selection The list and the second selection list are used for information selection, the first selection list includes category information of material application or material structure; the second selection list includes category information of entity type;

The input information obtaining module is configured to obtain input information based on the information input area, the first selection list and/or the second selection list.

On the other hand, the embodiment of this specification also provides a method for generating summary information, which is applied to a server, and includes:

Receiving a summary information generation request sent by the client, where the summary information generation request includes the data source for which the summary is to be generated;

The entity information and the association relationship information between the entity information are extracted from the data source, and the corresponding entity information is associated according to the extracted association relationship information; wherein the association relationship information includes the association direction between the entity information and Information describing the type of association;

The summary information is generated according to the associated entity information, and the generated summary information is sent to the client, so that the client can display it.

The fifth receiving module is configured to receive a summary information generation request sent by the client, where the summary information generation request includes the data source for which the summary is to be generated;

The eighth extraction module is used to extract entity information and the association relationship information between the entity information from the data source; wherein the association relationship information includes information describing the association direction and the association type between the entity information;

The sixth association module is used to associate corresponding entity information according to the extracted association relationship information;

The generation module is used to generate summary information based on the associated entity information;

The seventh sending module is configured to send the generated summary information to the client, so that the client can display it.

On the other hand, the embodiment of this specification also provides a method for generating summary information, which is applied to the client, and includes:

Send a summary information generation request to the server, where the summary information generation request includes the data source for which the summary is to be generated; so that the server extracts the entity information and the association relationship information between the entity information from the data source, based on the extracted association relationship information Associate the corresponding entity information; wherein, the association relationship information includes information describing the direction of association between the entity information and the type of association; and, according to the associated entity information, summary information is generated, and the generated summary information is sent The client;

Receive and display the summary information sent by the server.

The eighth sending module is configured to send a summary information generation request to the server, where the summary information generation request includes the data source for which the summary is to be generated; so that the server extracts entity information and the association relationship information between the entity information from the data source , Associate the corresponding entity information according to the extracted association relationship information; wherein the association relationship information includes information describing the association direction and association type between the entity information; and, generate summary information based on the associated entity information , Sending the generated summary information to the client;

The fifth receiving module is configured to receive the summary information sent by the server;

The third display module is used to display the summary information.

The information flow extraction method, device, and equipment provided by one or more embodiments of this specification can extract entity information and the association relationship information between entity information at the same time, so that the association relationship information can be effectively used to further verify and verify the extracted entity information. Screening to determine whether the extracted entity information is entity information related to the target entity information, thereby further improving the accuracy of extracting entity information associated with the target entity information, and effectively filtering noise. At the same time, using the association relationship information can also associate the extracted entity information with the target entity information, effectively displaying the association relationship between the extracted entity information and the target entity information, which is convenient for users to view and sort out, so that users can accurately and efficiently obtain what they need Useful information, find the needed or new solutions, and improve the user experience.

Brief description of the drawings

Figure 1 is a schematic flow diagram of an information flow display method provided in this specification.

Figure 2 is a schematic diagram of the information flow in an embodiment provided in this specification.

Figure 3 is a schematic diagram of the information flow in another embodiment provided in this specification.

Figure 4 is a schematic diagram of the information map in another embodiment provided in this specification.

Fig. 5 is a schematic diagram of the information map in another embodiment provided in this specification.

Fig. 6 is a schematic diagram of an information map in another embodiment provided in this specification.

Fig. 7 is a schematic diagram of an information map in another embodiment provided in this specification.

Fig. 8 is a schematic diagram of an information map in another embodiment provided in this specification.

Fig. 9 is a schematic diagram of an information map in another embodiment provided in this specification.

FIG. 10 is a schematic diagram of the information map in another embodiment provided in this specification.

FIG. 11 is a schematic diagram of the information map in another embodiment provided in this specification.

FIG. 12 is a schematic diagram of the information map in another embodiment provided in this specification.

FIG. 13 is a schematic diagram of the information map in another embodiment provided in this specification.

FIG. 14 is a schematic diagram of the visual display of the information graph in another embodiment provided in this specification.

Fig. 15 is a schematic diagram of material attribute classification in another embodiment provided in this specification.

Figure 16 is a schematic diagram of using units and metrics to link materials to attribute lists from different attribute classes in another embodiment provided in this specification.

FIG. 17 is a schematic diagram of extracting attributes, units, and metrics in another embodiment provided in this specification.

FIG. 18 is a schematic diagram of linking the extracted attributes, units, and metrics to corresponding materials in another embodiment provided in this specification.

Figure 19 is a schematic diagram of extracting attributes, units, and metrics in another embodiment provided in this specification.

FIG. 20 is a schematic diagram of extracting attributes, units, and metrics in another embodiment provided in this specification.

FIG. 21 is a schematic diagram of extracting attributes, units, and metrics in another embodiment provided in this specification.

Figure 22 is a schematic diagram of linking extracted attributes, units, and metrics to corresponding materials in another embodiment provided in this specification.

Figure 23 is a schematic flow chart of an information flow extraction method provided in this specification.

Figure 24 is a schematic diagram of a model structure of an information flow extraction device provided in this specification.

Figure 25 is a schematic flow chart of an information retrieval method provided in this specification.

FIG. 26 is a schematic diagram of a search interface and search results in another embodiment provided in this specification.

Figure 27 is a schematic diagram of a retrieval interface in another embodiment provided in this specification.

FIG. 28 is a schematic diagram of retrieval results in another embodiment provided in this specification.

FIG. 29 is a schematic diagram of application classification in another embodiment provided in this specification.

FIG. 30 is a schematic diagram of the material structure classification interactive interface in another embodiment provided in this specification.

FIG. 31 is a schematic diagram of the classification of polymers in another embodiment provided in this specification according to the material structure.

FIG. 32 is a schematic diagram of an application classification interaction interface in another embodiment provided in this specification.

FIG. 33 is a further detailed classification diagram of smart materials in another embodiment provided in this specification.

FIG. 34 is a schematic diagram of the retrieval result in another embodiment provided in this specification.

FIG. 35 is a schematic diagram of a retrieval interface in another embodiment provided in this specification.

FIG. 36 is a schematic diagram of retrieval results in another embodiment provided in this specification.

Fig. 37 is a schematic diagram of an image of the protein structure, DNA plasmid, and material microstructure in another embodiment provided in this specification.

FIG. 38 is an image schematic diagram of a circuit diagram and a flowchart in another embodiment provided in this specification.

Fig. 39 is a schematic diagram of material attribute classification in another embodiment provided in this specification.

FIG. 40 is a schematic flowchart of a method for generating an abstract provided in this specification.

Ways to implement the invention

In order to enable those skilled in the art to better understand the technical solutions in this specification, the following will make clear and complete the technical solutions in one or more embodiments of this specification in conjunction with the drawings in one or more embodiments of this specification. It is obvious that the described embodiments are only a part of the embodiments in the specification, rather than all the embodiments. Based on one or more embodiments of the specification, all other embodiments obtained by a person of ordinary skill in the art without creative work shall fall within the protection scope of the embodiment scheme of this specification.

In an example scenario, the user can search for information through the client. The client can be a mobile device. For example, the client can be a smart phone, a tablet electronic device, a portable computer, a personal digital assistant (PDA), a vehicle-mounted device, or a smart wearable device, etc. The client can also be a desktop device. For example, the client can be a server, an industrial computer (industrial control computer), a personal computer (PC), an all-in-one machine, or an intelligent self-service terminal (kiosk), etc.

The user can enter information in the page displayed on the client. The input information is reference information that the user wants to search for. If the user wants to inquire about alumina ceramics-related information, the user can enter the information "alumina ceramics" in the page displayed on the client, and the client can send an information search request to the server, and the search request can be accompanied by the above input information " Alumina ceramics". The server may extract the input information "alumina ceramics" from the information search request, and use the input information "alumina ceramics" as reference information to extract entity information or data sources associated with the input information "alumina ceramics".

The server can be provided with storage or information linking with the database. After the server receives the information search request sent by the client, it can use the data source in the server memory or the data source in the database linked to it to search for information. After the server completes the information search, it can feed back the searched information to the client, so that the client can display it.

However, there may be many different description forms of information in the same data source or different data sources, and even a large amount of incomplete and non-descriptive information, which leads to the problem of inaccuracy and incompleteness of the searched information. Moreover, the amount of information searched may be relatively large, irrelevant or noisy information, and the information is usually not displayed concisely to users, which makes users need to spend a lot of energy to filter and sort out when viewing, which makes users easy to miss what they want. Information, and the experience is poor.

In some embodiments, the associated entity information of the input information and the association type and direction between the input information and the associated entity information can be extracted from the data source, and the extracted associated entity information can be combined with the associated entity information by using the associated relationship information. The input information is associated to form an information flow corresponding to the input information. Then, the information flow can be shown to the user, so as to visually view the required information and the relationship between the information through the information flow. A specific embodiment is shown in Fig. 1. In an embodiment of the information flow display method provided in this specification, the method may include:

S20: The client sends an information flow acquisition request to the server, where the information flow acquisition request includes input information acquired by the client.

The client can send an information flow acquisition request to the server. The information flow acquisition request may include user input information acquired by the client. For example, there may be an information input box corresponding to the client, and the user may input information into the client by manually inputting or selecting a drop-down selection box of the information displayed by the client. Then, the client may further respond to the user's operation to generate an information flow acquisition request, and the information flow acquisition request may be accompanied by the aforementioned input information. Then, the client can send the information flow acquisition request to the server. For example, the display interface of the client can be provided with a search button or an information flow generation button. The user can trigger the search instruction by clicking the button or hitting the enter key on the keyboard, and then the client can generate information in response to the user's operation The stream gets the request and sends it to the server. Of course, the specific implementation manners of the foregoing information input, user operation, and information flow acquisition request are only examples, and the specific implementation is not limited to the foregoing manners.

In an example scenario, the client can show the user an information input box and an information flow generation button. The user can enter the information "polyvinyl alcohol" in the input box, and then click the information flow generation button, and the client can respond to the user's click operation , To generate an information flow acquisition request to obtain all or part of the information about the material. Or, the user enters the information "organic electroluminescence display coating" in the input box to obtain all or part of the information about the application of the material from the raw material to the material application. Alternatively, other types of information can also be input to obtain information related to the input information.

In another scenario example, the user can also define the extraction direction of the required information flow or the start node and end node of the information flow. For example, the process information flow of a certain material is limited to obtain information related to the material process of the material. The user can enter "polyvinyl alcohol material technology" or "polyvinyl alcohol and processing technology" on the page displayed on the client. Or, the page displayed by the client is also provided with a drop-down selection box, and the drop-down selection box is set with one or more information extraction direction selection items, such as material technology, material application and other options. After entering "polyvinyl alcohol", the user can further select the information extraction direction from the drop-down selection box to obtain the corresponding information. Correspondingly, the "polyvinyl alcohol" and the extraction direction "processing technology" can be attached to the retrieval request together and sent to the server.

S22: The server receives the information flow acquisition request, and extracts one or more entity information from the input information as target entity information.

After the server receives the information flow acquisition request, it may extract one or more entity information from the input information in the information flow acquisition request as target entity information to extract the information flow.

The entity information may include information such as a proper name or a meaningful quantitative phrase recognized from the data source. The proprietary name may include, for example, material name, specific material application, manufacturing method, material attribute, trade name, and so on. The meaning of quantitative phrases such as measurement values of material properties, dates, and so on. The entity information may include entity type and entity value. The entity type may refer to information that describes the category to which the information with common characteristics belongs after categorizing the information. The entity value may include specific information corresponding to each entity type in the data source. The data source may include relatively closed text information such as patent documents, dissertation documents, and text information in open source databases. The information in the data source is huge and the type of information is complex, and the workload of information analysis and extraction is usually large. By combining entity information to extract information, the accuracy and efficiency of information extraction can be greatly improved.

For the material technology field, the entity type may include, for example, material name entity, material structure type entity, material application entity, material process entity, material attribute entity, material trade name entity, material supplier/manufacturer entity, unit entity, and measurement Entities and so on. Wherein, the material process entity may also include a process method entity between materials, a manufacturing method entity between a material and a material application, an intermediate material entity in a processing or manufacturing process, and the like. The intermediate material entity may include, for example, an additive entity, a catalyst entity, and the like.

The material name entity may include the entity type corresponding to the description information of the material name to determine what kind of material. The entity value corresponding to the material name entity can be, for example, chemical, physical or biological description information with the name of the material, such as polyvinyl alcohol, zirconia, martensitic stainless steel, etc.

The material structure type entity may include an entity type corresponding to the description information of the type to which each material belongs, determined according to the material structure. For example, the entity value of the material structure type entity can be metal, ceramic, biology and so on. The material application entity may include an entity type corresponding to the description information of the application corresponding to the material. The entity value of the material application entity may be, for example, the application field of each material or specific application information. Among them, the application fields may include building materials, energy materials, 3D printing materials, optoelectronic materials, and so on. Specific applications may include organic electric display coatings, packaging bags, and so on.

The material process entity may include the entity type corresponding to the description information of the process/manufacturing method of obtaining another one or more materials or material application from one or more materials. The physical value of the material process entity may be, for example, description information of the process/manufacturing method, such as polymerization, hydrolysis, alloying, metal refining, stretching, finishing, etc., as well as additives, catalysts, etc. in the material processing or manufacturing process .

The material property entity may include the entity type corresponding to the description information that characterizes the internal characteristics such as the structure and performance of the material. The physical value of the material process entity may include, for example, melting point, thermal conductivity, specific heat capacity, yield strength, elastic modulus, and so on.

The material commodity name entity may include the entity type corresponding to the commodity name description information used in the actual production and sales of the material. The material supplier/manufacturer entity may include the entity type corresponding to the supplier/manufacturer description information of each material.

The unit entity may include the entity type corresponding to the measurement unit description information of the specific value of the attribute information of each material. The measurement entity may include the entity type corresponding to the specific value description information of the attribute information of each material. If the unit of melting point can include degrees Celsius, Kelvin, etc., the corresponding degrees Celsius and Kelvin are the entity values corresponding to the unit entity. For example, the melting point of tungsten is 3410 degrees Celsius, where the melting point is the physical value of the material properties, 3410 is the physical value of the measurement entity, and degrees Celsius is the physical value of the unit entity.

For the above input information "polyvinyl alcohol", the entity information that can be extracted is "material name entity-polyvinyl alcohol". For A-B in the entity information expression form, A is the entity type, B is the entity value, that is, the material name entity is the entity type, and polyvinyl alcohol is the entity value. The foregoing entity type and corresponding entity value expression methods are only examples for illustration, and methods such as association or table storage may also be used for specific data processing. In order to facilitate a unified description, the entity information in the embodiments of this specification uniformly adopts the form of "A-B". For example, the entity information that can be extracted for the aforementioned "melting point of tungsten is 3410 degrees Celsius" is "material name entity-tungsten", "material attribute entity-melting point", "unit entity-degrees Celsius", and "metric entity-3410".

In some embodiments, the method for extracting entity information, such as named entity extraction, can be obtained by learning and training a large amount of sample data. In some embodiments, for example, a natural language processing (Natural Langunge Possns, NLP) algorithm may be used to process and tokenize the input information. Then, based on a pre-built entity information dictionary or entity information database, a named entity recognition algorithm (Named Entity Recongnition, NER) can be used to extract the entity information in the input information. The entity information dictionary may include common professional word information in a designated field constructed in advance based on a large amount of sample data, and the entity information dictionary may for example be a database including professional word information such as material names, material attributes, and material applications in the material field. The entity information database may include a database pre-built by the platform that includes common professional word information in a specified field, such as an entity information database that includes professional word information such as material names, material attributes, and material applications in the material field. By extracting entity information based on a pre-built entity information dictionary or entity information database, the efficiency and accuracy of entity information extraction can be greatly improved. The server can use the extracted entity information as the target entity information to extract the information flow. The target entity information is reference information extracted from the information flow, and the server may use the target entity information as a reference to extract entity information associated with the target entity information and the association relationship information between the target entity information and the extracted entity information, And then construct the information flow to obtain the corresponding target entity information. For example, for the input information mentioned in the above example, the entity information mentioned is "material name entity-polyvinyl alcohol", the entity information can be used as the target entity information, and the data source can be extracted from the data source with "material name entity-polyvinyl alcohol". The entity information associated with the "polyvinyl alcohol" and the association relationship information between the "material name entity-polyvinyl alcohol" and the extracted entity information are then constructed to obtain the information flow of the "material name entity-polyvinyl alcohol".

In some embodiments, if there is more than one entity information extracted from the input information. The server can further analyze the information flow extraction requirements corresponding to the entity information extracted from the input information. If the extraction requirement is to separately extract the information flow corresponding to each entity information, the server may use each entity information as the target entity information, and extract the information flow corresponding to each entity information. For example, the input information is "polyvinyl alcohol OR polycarbonate", and the input information adopts the form of Boolean logic expression. If the logical operator "OR" means or is defined in advance, then the server can use the related word " OR" determines the information flow extraction requirements of the two entity information to be extracted separately. Then the server can use "material name entity-polyvinyl alcohol" and "material name entity-polycarbonate" as the target entity information respectively, that is, there are two target entity information, and the content of the two target entity information is "material name entity" -Polyvinyl alcohol" and "material name entity-polycarbonate".

Then, the server can extract the information flow of the two target entity information "material name entity-polyvinyl alcohol" and "material name entity-polycarbonate" to obtain the information flow of "material name entity-polyvinyl alcohol" and "material Name entity-polycarbonate" information flow. During specific implementation, the server may respectively extract the information streams of each target entity information in sequence, or may extract multiple target entity information information streams in parallel through parallel processing.

If the extraction requirement is a direction-restricted extraction, the input information mentioned in step S20 includes the direction-restricted extraction. The user enters "polyvinyl alcohol" in the input box, and selects or inputs the extraction direction limitation information "organic electroluminescent display coating" in the extraction direction limitation selection box or the input box. Then the server extracts the two entity information "material name entity-polyvinyl alcohol" and "material application entity-organic electroluminescent display coating" from the input information, and extracts the material from the material polyvinyl alcohol to Material application The information flow constituted by the material processing technology between the coatings of the organic electroluminescence display, other materials involved, and material suppliers.

Of course, the above-mentioned information flow extraction requirements, extraction requirements selection methods, input methods, or interface display methods are preferred examples. During specific implementation, other methods can also be used for implementation, which are not limited here.

S24: The server extracts at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, where the association relationship information includes the association between the entity information Information describing the direction and type of association.

For any piece of target entity information, the server may extract at least one associated entity information corresponding to the target entity information from the data source, and extract the association relationship information between the target entity information and the associated entity information. The associated entity information may refer to entity information that has a certain association relationship with the target entity information. The association relationship information may include information describing the direction and type of association between entity information.

The association direction may indicate the flow direction of the entity information in the information flow, and intuitively identifies the flow direction of the information flow, which is convenient for combing and viewing the entity information. If the material monomer obtains the material polymer through the material technology, the correlation direction among the material monomer, the material technology, and the material polymer is that the material monomer obtains the material polymer through the material technology. In some embodiments, when the information flow is displayed, arrows may also be used to indicate the direction of association. As shown in FIG. 2, the arrows in FIG. 2 identify the direction of association between entity information in the information flow.

The association type may include a feature description of the association relationship between entity information with different association relationship characteristics, so as to better sort out and verify whether there is an association between the entity information and what kind of association there is. For example, when materials are related by processing technology, the corresponding type of association can be the process between materials and materials. The type of association between material and material type information is material and material type.

The direction and type of association between entity information can be determined by analyzing the text information where the entity information is located. For example, a large amount of sample data can be trained, and the algorithm model obtained by training can be used to extract the association direction and type of entity information from the text information. In some embodiments, the association direction and association type between the entity information can also be determined by comprehensively considering the entity types between the entity information, so as to determine the association direction and association type more accurately and conveniently.

The association relationship information may include information describing an association direction and an association type. The information description method can be a plain text description, for example, the type of association is "material and material type", and the direction of association is a material type with "synthetic resin" as "polyvinyl alcohol". Alternatively, the method of information linking can also be adopted, for example, the server establishes a link for two or two entity information through symbols, graphics, etc., and then marks the link type and link direction on the link. Of course, the association type and the association direction can also be represented by numbers, symbols, graphics, etc. Of course, other information description methods can also be used.

The association relationship information may include the description information of the association direction and the association type, or may only include the description information of the association direction or the association type. For example, for the entity information between the name variants of the material name, there may be no information describing the association direction, and only the association type to indicate that the two are co-location information. For the entity information corresponding to the material and the material type, there may also be only the description information of the associated type. As for the entity information corresponding to the process between the material and the material, there may only be description information of the associated direction. For entity information that only has an association type, the association type can be marked when the information flow is displayed, or the entity value and entity type can be marked in the display entity information, and the entity type is used to characterize the association relationship between the entity information. At the same time, entity information can be associated with brackets or solid lines without arrows. For the association relationship information including the association direction, a straight line with arrows may be used to represent the information flow direction between entity information according to the association direction in the association relationship information, so as to facilitate visual display. Of course, the above-mentioned association direction and the display manner of the association relationship are only examples, and other methods can also be used in specific implementation.

In some embodiments, the associated entity information and the associated relationship information corresponding to the target entity information can be determined by analyzing the context information of the target entity information in the data source.

The context information may include the sentence where the target entity information is located and one or more sentences before and after the sentence where the target entity information is located. Alternatively, if there is a strong correlation between the paragraph where the target entity information is located and the paragraph before and after the paragraph through analysis, the context information may also include the paragraph where the target entity information is located and one or more paragraphs before and after the paragraph.

In some implementations, for example, the sentence in which the target entity information is located may be initially used as the context information in which the target entity information is located to extract the associated entity information and the association relationship information with the associated entity information. Then, one or more sentences before and after the sentence where it is located can be further added to the context information, and one or more sentences before and after the sentence where it is located can be used to perform the related entity information extraction and the association relationship information with the associated entity information. Fix. Alternatively, one or more paragraphs before and after the paragraph where it is located can be further added to the context information, and the associated entity information extracted by one or more paragraphs before and after the paragraph where it is located and the relationship information between the associated entity information and the associated entity information Make corrections.

Or, you can initially use the sentence where the target entity information is located as the context information where the target entity information is located to extract the associated entity information and the associated relationship information with the associated entity information. When the associated entity information and the associated entity information cannot be extracted In the case of the association relationship information between the information, one or more sentences before and after the sentence where it is located can be further added to the context information, and one or more sentences before and after the sentence where it is used are used to further extract the associated entity information and the related entity information of the target entity information. Association information between associated entity information. Of course, one or more paragraphs before and after the paragraph where it is located can be further added to the context information, and one or more paragraphs before and after the paragraph where it is located can be used to further extract or modify the associated entity information of the target entity information and the relationship between the associated entity information and the associated entity information. Relationship information between.

Usually, the entity information with strong association relationship is described in a sentence. By initially extracting the sentence where the target entity information is located as context information, the interference of other information can be avoided, and the target entity information can be extracted more accurately and efficiently. Corresponding associated entity information and association relationship information with associated entity information. Then, the association relationship information between the extracted entity information is further modified by combining the preceding and following sentences or paragraphs, or further extraction can be achieved when the entity information or the association relationship information between the entity information is not extracted. Improve the accuracy and comprehensiveness of the extraction of entity information and the relationship between entity information.

In some implementations, the data source where the target entity information is located can be retrieved first, and then the context information of the target entity information in the data source can be located. Then, natural language processing algorithms and the like can be used to analyze the context information where the target entity information is located, and extract the associated entity information corresponding to the target entity information and the association relationship information with the associated entity information.

The data source where the target entity information is located can be preliminarily retrieved through methods such as information matching. Then, the method provided in the foregoing embodiment can be used to locate the context information of the target entity information in the data source. You can first use the named entity extraction method to extract the entity information from the context information where the target entity information is located, and then use semantic role tagging, dependency semantic analysis, part-of-speech tagging and other methods to compare the extracted entity information and the sentence where the target entity information is located, or The sentence between the extracted entity information and the sentence where the target entity information is located is analyzed, and description information that can describe the direction and type of association between the extracted entity information and the target entity information is extracted. The extracted description information can be analyzed, or the extracted description information, the extracted entity information, and the target entity information can be comprehensively analyzed. If there is a certain association direction and/or association type between the extracted entity information and the target entity information, Then the extracted description information can be used as the associated entity information, and the extracted description information can be used as the associated relationship information.

For example, for the sentence "polyvinyl alcohol" in the data source where the target entity information "polyvinyl alcohol" is located, "polyvinyl alcohol is a synthetic resin, usually prepared by alcoholysis (usually called hydrolysis or saponification) of polyvinyl acetate", The named entity extraction algorithm can be used to extract the entity information "material name entity-polyvinyl alcohol", "material type entity-synthetic resin", "material name entity-polyvinyl acetate", "material process entity-alcoholysis", " Material process entity-hydrolysis", "material process entity-saponification".

The entity information in the above sentences and the words between the entity information can be further processed by natural language processing methods such as semantic role labeling, dependency semantic analysis, part-of-speech labeling, etc., to extract the association relationship between "polyvinyl alcohol" and "synthetic resin" The description information "X is a kind of Y", further combining the entity types of "polyvinyl alcohol" and "synthetic resin" can determine the type of association between the two as "material and material type", and the direction of association is "synthetic resin". "Polyvinyl alcohol" type of material.

The description information of the relationship between "polyvinyl alcohol", "polyvinyl acetate" and "alcolysis" is "X is prepared by Y and Z", through further analysis of the entity types of the three entities: "alcolysis" corresponds to the material Process entity, the entity type corresponding to "polyvinyl alcohol" and "polyvinyl acetate" is the material name entity. It can be further determined that the association type between the three entities is "material and material process", and the association direction is "poly "Vinyl acetate" obtains "polyvinyl alcohol" through "alcolysis".

At the same time, it can also extract the description information of the relationship between "alcolysis", "hydrolysis", and "saponification", "alcolysis (usually called hydrolysis or saponification)", through the description information and "alcolysis" and "hydrolysis" "" and "saponified" entity type, the association type between the three can be determined as "co-entity", and the direction of association is co-location information, and the co-location entity or co-location information represents each of two or more entity information Entity information is other representations of other entity information.

Natural language processing methods such as named entity extraction algorithm, semantic role tagging, dependency semantic analysis, part-of-speech tagging, etc. used in the process of extracting entity information and association relationship information can extract corresponding entity information and association relationship information according to the information flow Features are obtained after deep learning of a large amount of sample data, which can improve the accuracy of entity information and association information extraction.

In other embodiments, the associated entity information and association relationship information corresponding to the target entity information may be determined in the following manner:

For the context information of the target entity information determined above, the context information can also be processed using the NLP algorithm, for example, the data source can be processed and tokenized using the NLP algorithm. Then, based on the entity information dictionary or entity information database, NER can be used to identify the entity information in the context information. For example, the construction of material information flow can be based on the entity information dictionary or entity information database corresponding to the material, or the entity information dictionary or entity information database corresponding to the material structure type corresponding to the target entity information, and use the NER algorithm to obtain the context information Extract the entity information contained in the context information.

For example, if the target entity information is "polyvinyl alcohol" and the material structure type corresponding to "polyvinyl alcohol" is polymer, then the entity information dictionary or entity information database corresponding to the polymer is obtained. The NER algorithm is used to extract the entity information from the context information of the target entity information, and to compare the information with the information in the entity information dictionary or entity information database to more accurately and efficiently identify the entity information in the context information. For example, the corresponding description information of monomer, polymerization reaction, polymerization medium, polymer, polymer type, polymer type, manufacturing process, manufacturer, supplier, chemical modification, physical modification, application can be identified.

Then, the NLP algorithm filtering method can also be used to remove the interference information in the extracted entity information. The interference information may include stop words, common words, non-scientific words, words unrelated to the material context, and the like. Then, the extracted entity information and target entity information are used as objects, and for each object, the NLP algorithm is further used to analyze the context information to determine the metadata corresponding to each object. The metadata may refer to information describing each object. Then, the metadata corresponding to the target entity information and the metadata corresponding to the extracted entity information can be compared with each other to determine whether there is an association relationship between the target entity information and the extracted entity information and specific association relationship information. If there is an association relationship between the target entity information and the extracted entity information, the entity information with the association relationship can be used as the associated entity information of the target entity information, and the association relationship information between the two can be determined based on the comparison result to compare The two are related.

S26: The server uses the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information.

The server can use the extracted association type and association direction to associate the corresponding target entity information and the associated entity information, and obtain the information flow corresponding to the target entity information. The implementation of the association can be carried out with reference to the content described in step S24, which will not be repeated here.

S28: The server sends the information stream to the client for display on the client.

The server may send the information stream to the client for display on the client. For example, a visualization tool can be used to visualize the information flow, and then the visualized information flow can be sent to the client for visual display on the client. Figure 2 shows the information flow established in the above example.

By extracting entity information and the association relationship information between entity information at the same time, the association relationship information can be effectively used to further verify and filter the extracted entity information to determine whether the extracted entity information is entity information related to the target entity information. Thereby, the accuracy of extracting entity information associated with the target entity information is further improved, and noise is effectively filtered. At the same time, using the association relationship information can also associate the extracted entity information with the target entity information, effectively displaying the association relationship between the extracted entity information and the target entity information, which is convenient for users to view and sort out, so that users can accurately and efficiently obtain what they need Useful information, find the needed or new solutions, and improve the user experience.

For example, the user is trying to find some or all of the processing technology or manufacturing method, raw materials, additives and other information of a specific material or material application, and the scheme provided by the above embodiment will associate the specific material or material application in the same or different data sources. Entity information is extracted and associated to form an information flow and displayed to users. Using the information flow corresponding to the specific material, users can easily and conveniently find the information they need, avoiding the process of selecting and sorting out multiple different data sources by themselves, saving time. At the same time, through the above information flow method, the information of each entity can also be displayed to users in an associated manner, which can avoid the omission of information and the neglect of information due to the large amount of data or the lack of understanding of the industry when users perform information screening on their own. The relevance of, enables users to more accurately obtain the solutions they need and the raw materials and processing technology needed in the solutions. In addition, through the above method of simultaneously extracting entity information and the association relationship information between entity information, the extraction of effective information from noise information in the data source can also be effectively removed, and the accuracy of the information obtained by the user can be improved.

The information in the data source is complex and changeable, and entity information and the association relationship information between entity information are usually difficult to accurately extract. In some embodiments, the algorithm model of the associated entity information corresponding to the target entity information and the association relationship information with the associated entity information may be trained based on the entity type of the target entity information, so as to further improve the association between the associated entity information and the entity information. Accuracy of relational information extraction. For the entity information in the field of material technology, the conversion between materials and materials or between materials and material applications is usually through a certain processing technology or manufacturing method, and a material or material application is obtained from which material processing, And, what kind of processing technology or manufacturing method is used is usually what users want to know. Based on this application scenario, in other embodiments of this specification, the extracting at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, Can include:

Retrieve the data source where the target entity information is located;

The material process information association manner may be determined according to the information association characteristics between materials or between materials and material applications. When the context information of the target entity information includes material process entities, material name entities, and/or material application entities, the context information may usually contain information about the processing technology conversion of one or more materials, or the material What kind of manufacturing method is used to manufacture the application? When extracting the material name entity, material application entity, and material process entity, the corresponding information association method can usually be the way that the material name entity obtains the material name entity through the material process entity, or the material name entity passes the material process entity The way to get the material application entity. At this time, the material process entity can be used as the associated entity information of the material name entity or the material application entity, or part of the association relationship information between multiple material name entities or between the material name entity and the material application entity.

In some embodiments, algorithms such as deep learning can be used to learn a large number of data sources including material process entities, material name entities, and/or material application entities, so as to accurately associate material name entities or material name entities through material process entities. With the material application entity, the entity information and the relationship between entity information can be accurately extracted.

For example, the user enters the information as polyvinyl alcohol in the input box of the client and clicks the information flow extraction button. The client can generate an information flow acquisition request based on polyvinyl alcohol and send it to the server. After the server receives the information flow acquisition request, it can extract the input information from the information flow acquisition request, and then can further perform information extraction on the input information. The server performs entity information extraction on the input information, and the extracted entity information may include "material name entity-polyvinyl alcohol". Then, the server can extract information from the data sources in the database, and extract one or more data sources containing polyvinyl alcohol. Then, you can locate the contextual information where the polyvinyl alcohol is located.

It is assumed that the contextual information located is as follows: "1,2-ethylene glycol polyvinyl alcohol can be produced according to known or conventional production methods. For example, 1,2-ethylene glycol polyvinyl alcohol can be produced using the following production method: Polymerization of vinyl acetate under pressure (load) higher than under conventional conditions to prepare polyvinyl alcohol; and copolymerization of vinyl acetate and ethylene carbonate to obtain 1,2-ethylene glycol polyvinyl alcohol with the aforementioned content".

Using the named entity extraction method, you can extract "material name entity-vinyl acetate", "material process entity-polymerization"; and "material name entity-vinyl acetate", "material name entity-vinyl carbonate", Entity information such as "material process entity-copolymerization". Then the entity information corresponding to the material name entity can be initially used as the associated entity information, and the entity information corresponding to the material process entity can be used as the type of association between the material name entities. At the same time, analyze the sentence where the above information is located to further determine the direction of the association. Analyzing the description information "polymerization of vinyl acetate to prepare polyvinyl alcohol" and "copolymerization of vinyl acetate and ethylene carbonate to obtain the above content of 1,2-ethylene glycol polyvinyl alcohol", the connection direction can be obtained as "acetic acid "Vinyl ester" is "polymerized" to obtain "polyvinyl alcohol", and "vinyl acetate" and "vinyl carbonate" are "copolymerized" to obtain "polyvinyl alcohol".

Then, the "polyvinyl alcohol" and "vinyl acetate" can be associated according to the extracted association type, direction and other association information, and the "polyvinyl alcohol" and "acetic acid" can be "copolymerized" according to the extracted association information. "Vinyl ester and vinyl carbonate" are associated. For example, the "arrow" can be used to indicate the direction of association, and the information display algorithm can be used to display the information flow formed after the association, as shown in Figure 3 (a) and (b) respectively.

Based on the information association characteristics of the material process, the extracted information including the material name entity, the material application entity, and the material process entity can be associated with information, which can further improve the accuracy of the extraction of the material entity information and the association relationship information between the material entity information.

In other embodiments of the present specification, the association relationship information may further include colocation entity description information, and the colocation entity description information may include description information in other expression forms that describe the associated entity information as target entity information. When the association relationship information includes the co-located entity description information, the server may add the associated entity information corresponding to the co-located entity description information to the target entity information, and update the target entity information. Correspondingly, the server may extract at least one associated entity information corresponding to the updated target entity information and the association relationship information between the target entity information and the associated entity information from the data source. And, using the association relationship information to associate the updated target entity information and the associated entity information to obtain the information flow corresponding to the updated target entity information.

For example, for the polyvinyl alcohol mentioned in the above example, when analyzing the data source, extract the description information of "polyvinyl alcohol can also be expressed as "poly(vinyl alcohol)" or "PVA", you can use the description The information is used as the association relationship information to associate poly(vinyl alcohol), polyvinyl alcohol and PVA. You can also add poly(vinyl alcohol) and PVA to the target entity information, and update the target entity information. Correspondingly, the updated The target entity information can include polyvinyl alcohol, poly(vinyl alcohol), and PVA. Then, you can use poly(vinyl alcohol), polyvinyl alcohol, and PVA to simultaneously perform data source retrieval and use poly(vinyl alcohol) in the retrieved data source. Poly(vinyl alcohol), polyvinyl alcohol, and PVA for the extraction of related entity information and related relationship information. After obtaining the sub-information streams corresponding to poly(vinyl alcohol), polyvinyl alcohol, and PVA, you can also use poly(vinyl alcohol) at the same time. ), polyvinyl alcohol and PVA as the reference information, associate each sub-information stream extracted based on poly(vinyl alcohol), polyvinyl alcohol, and PVA to obtain the information stream corresponding to the updated target entity information.

Of course, for other entity types, the above method can also be used to associate different expressions of the same entity information, and then further entity information extraction, such as "alcoholysis", "hydrolysis", and "saponification" in the above examples "It belongs to the co-location information corresponding to the extracted material process entity.

Or, in other embodiments, the metadata of each entity information may be extracted based on the solution provided in the above-mentioned embodiment, and the metadata corresponding to each entity information may be compared with each other. If the metadata corresponding to the two entity information is determined If there is a certain similarity or it is determined through metadata comparison to explain that two entity information is each other, then the two entity information can be co-located entities with each other. The same material, material application, process, production method, etc. may have many different description forms in the data source. For example, polyvinyl alcohol can also be described in English or other language forms. Complex and large data sources usually lack specific and standardized descriptions of material names, applications, or processes. When users search for information, they usually enter only one or more forms that are familiar to users. The information finally queried is not accurate and comprehensive. For example, due to the different ways of expressing material names in patents, users searching for certain materials such as polyvinyl alcohol may not find certain materials that use PVOH, PVA, poly(vinyl alcohol), hydrolyzed polyvinyl acetate, etc. These patents lead to incomplete analysis of user infringement, which leads to the risk of infringement.

Using the solution of the above embodiment, when extracting entity information, when other expression forms of entity information are extracted, the other expression forms are updated to the reference information relied on during information search in time, and the updated reference information is used to perform The search and association of information can greatly improve the accuracy and comprehensiveness of information extraction. For specific display, you can also provide or select a common name for a variety of material name variants, and link with each material name to identify the material name, which is convenient for users to sort out and read the information.

In other embodiments, the method may further include: the server extracts the target entity information or the product name of the associated entity information from the data source; and extracts the target entity information or the product name corresponding to the associated entity information according to the product name. The manufacturer or supplier associates the manufacturer or supplier with the target entity information or associated entity information. By further extracting the product name corresponding to each material, and extracting the supplier or manufacturer information corresponding to each material from open source databases and other resources through the product name, associating the supplier or manufacturer information with the material, it is convenient for users to follow The information flow directly obtains the manufacturing source of the material and improves the user experience.

Multiple associated entity information of the target entity information may be extracted from the same data source or multiple data sources, and the association relationship information between the target entity information and different associated entity information. In some embodiments, the server may also use the information flow formed by associating the extracted one or more associated entity information and the corresponding association relationship information as the sub-information flow. Then, using the target entity information as the reference information, the information flow from the same Data sources or different sub-information streams extracted from different data sources are linked to obtain the information stream corresponding to the target entity information, and then the final information stream is displayed to the user, which further improves the correlation between the information and facilitates the comprehensive analysis by the user Check.

Suppose that the user is looking for a solution from beginning to end for a specific material, such as from monomer to application. Through the above information flow, users can stitch together entities obtained from different sources (databases, patents, etc.), so that in order to obtain a complete solution at a single location, there is no need to analyze multiple different data sources, thereby saving time.

Or, the user needs to search for the manufacturing process of a new material, but cannot search because the manufacturing process patent has not been published or submitted. However, there may already be journal articles, news articles, or other types of data sources describing the material's process. By extracting information from multiple data sources and performing correlation and fusion, users can obtain the existing manufacturing process more accurately and intuitively.

For example, the information streams shown in (a) and (b) in FIG. 3 can be used as sub-information streams, and (c) in FIG. 3 is an information stream obtained after combining two sub-information streams. Based on the foregoing solution, in some embodiments, the use of the association relationship information to associate the target entity information with the associated entity information may include: acquiring at least one associated entity information extracted from a data source and the target entity information The association relationship information with the corresponding associated entity information, the association relationship information between the target entity information and the corresponding associated entity information is used to associate the target entity information with the corresponding associated entity information to obtain the target A sub-information stream of entity information; using the target entity information as the reference information to link the sub-information streams obtained from the same or different data sources to obtain the information stream corresponding to the target entity information.

In other embodiments, the method further includes: the server may also extract the information flow of each associated entity information extracted from the data source as the target entity information, and obtain the information flow corresponding to the multiple target entity information; and Link the information streams corresponding to multiple target entity information with the corresponding target entity information as the reference information to obtain the information map.

The server may further use the extracted associated entity information as target entity information, extract the information flow, and further extract the information flow corresponding to each associated entity information, so as to obtain a series of information flows. Then, each associated entity information is used as the reference information to link each information stream, so that the information graph can be obtained. Figure 4 is an information map formed by the entire workflow from vinyl acetate (monomer) to organic electroluminescent display coating (application), which is related to polyvinyl alcohol by users searching for polyvinyl alcohol. Among them, all parties in Figure 4 The content in the box is entity information, and the arrow indicates the direction of association.

Use the related entity information to further extract the information flow, and then further link to construct the information map, you can get all the information from the material monomer to the final application, and it is convenient for users to understand the materials, applications, and applications involved in the entire life cycle of a material. A comprehensive, accurate and intuitive understanding of process and manufacturing.

For example, using the solution of the above embodiment, users can easily identify the entire process information flow from monomer to chemically modified polymer (including catalyst, polymerization reaction, chemical modification, physical modification and additives) in the data source with the help of information flow. , Which can also help users discover alternative polymerization processes, additives and catalysts.

For example, users who usually search for polyethylene may not get documents that use PE, polyethylene HDPE, or industry polyethylene trade names to label polyethylene. He may get incomplete information about different aspects of the material details. For example, some polyethylene patents may contain monomers, but there is no information about the polymerization process, the polymer processing of the polymer, or even the different applications that the polyethylene can be used for. . Users may need to search different patents to search for different details of polyethylene. Through the above solution, users can get all the information of a specific material through a single search, including name variants, suppliers, material applications, manufacturing processes, etc.

At the same time, by using the information flow of the above-mentioned embodiment, the user can also easily and conveniently find suitable materials for a specific application. By using the information flow, users can discover new materials, suppliers, and material manufacturing technologies for specific applications. For example, users looking for materials suitable for medical stents may find materials such as Bioflow-V and Orsio (trade name). However, the material behind the trade name cannot be obtained from the same source. By using the information map formed by the interlinking of the above information, users can understand that Orsio uses cobalt-chromium alloy as the base metal and has the knowledge of active poly-L-lactide (PLLA) polymer coating. Users can further explore other materials used to make stents, including stainless steel (CYPHER stent) with poly(ethylene-vinyl acetate copolymer) or poly(n-butyl methacrylate) coating, and cobalt with carbon coating Chrome alloy (CRE8 inner stent) and so on. Then, the user can also use the information atlas to obtain the entire manufacturing process of each material used in the stent.

Figures 4 to 13 are schematic diagrams of information maps of several common material types given in this specification. Among them, FIG. 4 and FIG. 5 are schematic diagrams of information maps of polymer materials. Figures 6 and 7 are schematic diagrams of information maps of metal materials. Figures 8 and 9 are schematic diagrams of information maps of ceramic materials. Figures 10 and 11 are schematic diagrams of information maps of biological materials. Figure 12 and Figure 13 are schematic diagrams of material information maps. Among them, Figure 4, Figure 6, Figure 8, Figure 10, Figure 12 are based on the entity information extracted from the data source and the entity information formed after association. Figure 5, Figure 7, Figure 9, Figure 11, Figure 13 are schematic diagrams of the general information flow of the above five materials.

Comparing five kinds of materials to obtain the information map and general information flow diagram after extracting the actual information from the data source, it can be seen that through the solution of the embodiment of this specification, the user can input the entity information corresponding to any one or more nodes in the information map , Such as a single material, material application, material technology, etc., and then part or all of the information related to the input information and the flow relationship between the information can be obtained, and finally the information flow can be obtained by association. This allows users to accurately obtain part or the entire information flow of each material from the material monomer to the material application. At the same time, it also effectively associates information such as name variants, processing techniques, manufacturing methods, manufacturers, and suppliers. It effectively realizes that the user can accurately and comprehensively obtain part or all of the information related to the input information and the circulation relationship between the information through less input information.

Of course, you can also further associate the data sources of each information node in the information stream, as shown in Figure 4, Figure 6, Figure 8, Figure 10, and Figure 12. The different gray levels in the figure represent entities from different data sources. information. In practical applications, you can also use different colors to represent different data sources, or click each information node to view which data source each information node comes from. By further associating and displaying data sources, information from different data sources can be effectively distinguished at a single node, which is beneficial for users to effectively control the technologies owned or used by different patentees, and make infringements easier and more convenient. , Patent rights, etc. Based on this application scenario, in other embodiments, the sub-information stream of the target entity information may also be associated with the corresponding data source. By associating the extracted sub-information stream with the corresponding data source, the user can conveniently view the source of the corresponding sub-information stream, and effectively realize the tracking of various information in the information stream.

In other embodiments, based on the solutions provided in the above embodiments, the page displayed by the client may also include the start node and end node options of the information flow. The user can click the corresponding option and enter the corresponding corresponding in the input box. Start node information and end node information. After obtaining the user's click and input information, the client can generate the input information of "start node information + end node information", attach it to the search request, and send it to the server. The server can generate and display the information flow based on the input information.

For example, the entity information can be extracted from the start node information and the end node information separately, the entity information extracted from either the start node information or the end node information is initially used as the target entity information, and the other corresponding entity information is used as the ending entity information . Then, you can extract the information flow of the target entity information, and further use the associated entity information in the extracted information flow as the target entity information, and further extract the information flow, and so on, if a certain associated entity information is terminating entity information , Then no more information extraction is performed on the associated entity information. In this way, the information graph between the start node information and the end node information can be obtained.

For example, if the starting node information is a single material, the ending node information is a material application. Using the above method, the information flow can be gradually extracted from the material monomer to the material application, and the information atlas from the material monomer to the material application can be obtained. Alternatively, it is also possible to gradually extract the information flow from the material application to the material monomer, and obtain the information map from the material monomer to the material application.

In other embodiments, the method may further include: the server uses the entity information in the information stream as an information node for interacting with the user, and uses a visualization method to visualize the information stream; and the processed information stream Sent to the client, so that the user can view the information flow and trigger the information node through the client; based on the user's triggering operation on the information node, feedback other information associated with the information node to the client. The information node is the information corresponding to each entity information in the information flow. For example, the trigger operation may include operation modes such as tap, long press, and slide. The other information associated with the information node may include, for example, data sources or other information streams that are not shown. For example, the visualization method may adopt the Neo4j graphical method. As shown in Figure 14, Figure 14 shows a schematic diagram of an interactive visualized information map. The circles represent different information nodes, the lines between the circles indicate that the information nodes are related, and the text on the line can indicate the type of association between the information nodes. Describe information, and use arrows to indicate the direction of association between information. In some embodiments, different colors may also be used to indicate the data source from which the information comes.

For example, the user can view the data source of the information node by clicking on the information node. Or, by sliding left or right based on the information node, to view the undisplayed lower information or upper information corresponding to the information node, where the lower information may refer to the material corresponding to the current information node to the material. The information between applications, the higher-level information may refer to the information from the material corresponding to the current information node to the single material. Or, swipe up or down based on the information node to view the undisplayed parameter entity information and supplier/manufacturer entity information corresponding to the information node. The parameter entity information can be, for example, information such as material type, process type, material attribute, unit, or measurement. By designing the display in a form that can interact with the user, the user experience can be further improved.

Based on the above application scenarios, in other embodiments, the method may further include: the server extracts the target entity information and the parameter entity information of the associated entity information from the data source, and the parameter entity information includes at least the material structure type entity, the process The entity information corresponding to one of the method entity, the material attribute entity, the unit entity or the measurement entity; the extracted parameter entity information is associated with the corresponding target entity information and associated entity information. In some implementations, for example, deep learning technology can be used to extract different parameter entity information and associate it with corresponding entity information.

The data source usually also contains further specific description information of the entity information. For example, for the process entity information, there may also be description information such as process method type, process parameters, and manufacturer information. For example, for the polymerization process, the data source may also contain more specific description information of the polymerization process, such as solution polymerization, bulk polymerization, suspension polymerization, and emulsion polymerization. At the same time, it may also provide a large number of polymerization processes. The parameter condition information used at the time, such as temperature, pressure, surrounding environment, production environment and other information. The above-mentioned information can be further extracted, and the material attribute entity, unit entity or measurement entity of each material involved in the above process method, or the parameter entity information of the surrounding environment, unit entity, measurement entity and other parameter entity information and corresponding entity information Association can enable users to obtain specific parameter information more accurately and comprehensively.

The main attribute classes are shown in Figure 15. Each attribute class has different attribute subcategories. For example, the spectroscopy category can be divided into subcategories including Raman spectroscopy, X-ray diffraction, and the like. Figure 16 is a schematic diagram of using units and measures to link materials to attribute lists from different attribute classes. Figure 17 is a schematic diagram of extracting attributes, units, and metrics from a data source. Figure 18 is a schematic diagram of linking the extracted attributes, units, and metrics to the corresponding materials. Figure 19, Figure 20, and Figure 21 show different representations of attributes, units, and metrics in the data source and schematic diagrams of information extraction. Figure 22 is a schematic diagram of linking the extracted attributes, units, and metrics to the corresponding materials.

Through the above-mentioned unit/metric and attribute extraction, different attributes associated with materials and changes in attributes when subjected to different treatments can be visualized, enabling users to select the best materials and treatments to meet their needs. At the same time, by extracting information such as material properties and units/measures, and associating them into the information flow, it is also possible to associate material applications with material properties, and further realize the application of recommending materials based on the material properties.

By extracting the information flow by using the solutions provided by one or more of the above embodiments, users can accurately and comprehensively understand the specific applications, process methods and other specific solutions that currently exist for the materials. At the same time, they can also use the associated data sources. Know the patentee who discloses the above-mentioned application and process method, so as to effectively avoid the user's risk of infringement.

For example, for users who conduct patent infringement analysis, the patent claims usually include a description of the material. If the patent claims a new material manufacturing process, the applicant usually needs to add all the details of the manufacturing process to the claims. However, If the patent involves a modification of the existing manufacturing process, the applicant does not have to add all the details of the manufacturing process. If an application-material link is newly discovered, the applicant can also add the application of the material in the patent claims. However, the applicant generally prefers to describe a broader application field in the patent claims, and more specific Applications are usually not described in the claims. In addition, the description of material properties, attributes, etc. in the claims will usually be described in general terms to obtain a better scope of protection.

In view of the lack of specific descriptions of materials, manufacturing processes and applications in many patent claims, it is difficult to accurately and comprehensively retrieve a lot of information. If a third party manufactures or uses the material, some modifications are made to the basic material that has been protected by the patent, or the manufacturing process is slightly modified, but the modification may have been covered by the broad claims of other manufacturing patents, or, The applications claimed by the patent application are within the scope of the application of the previously published patent. If the third party fails to conduct a comprehensive and accurate search, the third party may face a higher risk of infringement.

Based on the methods provided in the foregoing embodiments, some other embodiments of this specification also provide an information flow extraction method, which is applied to a server. As shown in FIG. 23, the method may include:

S40: Obtain target entity information;

S42: Extract at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, where the association relationship information includes an association direction between the entity information And information describing the type of association;

S44: Use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information.

In other embodiments, the using the association relationship information to associate the target entity information with the associated entity information may include:

In other embodiments, when the association relationship information includes colocation entity description information, the associated entity information corresponding to the colocation entity description information is added to the target entity information, and the target entity information is updated; wherein, The colocation entity description information includes description information in other expression forms that describe the associated entity information as target entity information;

In other embodiments, the method may further include:

In other embodiments, the extracting from the data source at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information may include:

Retrieve the data source where the target entity information is located;

In other embodiments, the method may further include:

In other embodiments, the method may further include: associating the sub-information stream of the target entity information with a corresponding data source.

In other embodiments, the method may further include:

For the specific implementation of the information flow extraction method provided in the foregoing embodiment, reference may be made to the implementation of the foregoing method, and the repetition will not be repeated.

The information flow extraction device provided by one or more embodiments of this specification simultaneously extracts entity information and the association relationship information between the entity information, and can effectively use the association relationship information to further verify and filter the extracted entity information to determine Whether the extracted entity information is entity information related to the target entity information, thereby further improving the accuracy of extracting entity information associated with the target entity information and effectively filtering noise. At the same time, using the association relationship information can also associate the extracted entity information with the target entity information, effectively displaying the association relationship between the extracted entity information and the target entity information, which is convenient for users to view and sort out, so that users can accurately and efficiently obtain what they need Useful information, find the needed or new solutions, and improve the user experience.

Based on the above-mentioned information flow extraction method, one or more embodiments of this specification also provide an information flow extraction device. The described devices may include systems, software (applications), modules, components, servers, etc., which use the methods described in the embodiments of this specification, combined with necessary implementation hardware devices. Since the implementation scheme of the device to solve the problem is similar to the method, the implementation of the specific device in the embodiment of this specification can refer to the implementation of the foregoing method, and the repetition will not be repeated. Specifically, FIG. 24 shows a schematic diagram of the module structure of an embodiment of an information flow extraction device provided in the specification. As shown in FIG. 24, the device may include:

The first obtaining module 402 may be used to obtain target entity information;

The first extraction module 404 may be used to extract at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes a pair of Information describing the direction of association between entity information and the type of association;

The first association module 406 may be configured to use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information.

In other embodiments, the first association module 406 may include:

The first associating unit may be used to obtain at least one associated entity information extracted from the data source and the association relationship information between the target entity information and the corresponding associated entity information, and use the target entity information and the corresponding associated entity information The association relationship information between associates the target entity information with the corresponding associated entity information, and obtains the sub-information stream of the target entity information;

The second associating unit may be used to link the sub-information streams obtained from the same or different data sources with the target entity information as the reference information, to obtain the information stream corresponding to the target entity information.

In other embodiments, when the association relationship information includes colocation entity description information, the apparatus may further include:

The update module may be used to add the associated entity information corresponding to the co-located entity description information to the target entity information, and update the target entity information; wherein the co-located entity description information includes information describing the associated entity as a target Descriptive information in other forms of entity information;

The first extraction module 404 may also be used to extract at least one associated entity information corresponding to the updated target entity information and the association relationship information between the target entity information and the associated entity information from the data source.

In other embodiments, the device may further include:

The second extraction module can be used to extract the information flow of each associated entity information extracted from the data source as the target entity information, and obtain the information flow corresponding to the multiple target entity information;

The second association module can be used to link the information streams corresponding to multiple target entity information with the corresponding target entity information as the reference information to obtain the information graph.

In other embodiments, the first extraction module 404 may include:

The retrieval unit can be used to retrieve the data source where the target entity information is located;

The locating unit may be used to locate the context information where the target entity information is located in the data source;

The extracting unit may be used to extract the associated entity information corresponding to the target entity information and the target according to the material process information association method when the context information includes the material process entity, material name entity, and/or material application entity The association relationship information between the entity information and the associated entity information.

In other embodiments, the device may further include:

The third extraction module can be used to extract the target entity information and the commodity name of the associated entity information from the data source;

The third association module may be used to extract the manufacturer or supplier corresponding to the target entity information or associated entity information according to the product name, and associate the manufacturer or supplier with the target entity information or associated entity information Make an association.

In other embodiments, the device may further include:

The fourth extraction module may be used to extract the target entity information and the parameter entity information of the associated entity information from the data source. The parameter entity information includes at least a material structure type entity, a process method entity, a material attribute entity, a unit entity, or a metric. Entity information corresponding to one of the entities;

The fourth association module can be used to associate the extracted parameter entity information with corresponding target entity information and associated entity information.

In other embodiments, the first association module 406 may further include:

The third association unit may be used to associate the sub-information stream of the target entity information with the corresponding data source.

In other embodiments, the device may further include:

The visualization processing module can be used to treat each entity information in the information flow as an information node for interacting with the user, and use a visualization method to visualize the information flow;

The first sending module may be used to send the processed information stream to the client, so that the user can view the information stream and trigger the information node through the client;

The second sending module may be configured to feed back other information associated with the information node to the client based on a user's triggering operation on the information node.

It should be noted that the above-mentioned device may also include other implementation manners according to the description of the method embodiment. For specific implementation manners, reference may be made to the description of the related method embodiments, which will not be repeated here.

The method or device described in the foregoing embodiment provided in this specification can implement business logic through a computer program and be recorded on a storage medium, and the storage medium can be read and executed by a computer to achieve the effects of the solution described in the embodiment of this specification. Therefore, this specification also provides an information flow extraction device that includes a processor and a memory storing processor-executable instructions. When the instructions are executed by the processor, the steps include the method described in any one of the foregoing embodiments.

The storage medium may include a physical device for storing information, and usually the information is stored in an electric, magnetic, or optical medium after digitizing the information. The storage medium may include: devices that use electric energy to store information, such as various types of memory, such as RAM, ROM, etc.; devices that use magnetic energy to store information, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, bubble memory, U disk; a device that uses optical means to store information, such as CD or DVD. Of course, there are other ways of readable storage media, such as quantum memory, graphene memory, and so on.

The information flow extraction device described in the above embodiment simultaneously extracts the entity information and the association relationship information between the entity information, and can effectively use the association relationship information to further verify and filter the extracted entity information to determine the extracted entity information Whether it is the entity information related to the target entity information, so as to further improve the accuracy of extracting the entity information related to the target entity information, and effectively filter the noise. At the same time, using the association relationship information can also associate the extracted entity information with the target entity information, effectively displaying the association relationship between the extracted entity information and the target entity information, which is convenient for users to view and sort out, so that users can accurately and efficiently obtain what they need Useful information, find the needed or new solutions, and improve the user experience.

Based on the method provided in the foregoing embodiment, the embodiment of this specification also provides an information flow display method, which is applied to a server, and the method may include:

The information stream is sent to the client for display on the client.

Based on the method provided in the foregoing embodiment, an embodiment of this specification also provides a server, which may include:

The first receiving module may be used to receive an information flow acquisition request sent by the client, where the information flow acquisition request includes input information acquired by the client;

The fifth extraction module may be used to extract one or more entity information from the input information as target entity information;

The sixth extraction module may be used to extract at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes Information describing the direction of association between information and the type of association;

The fifth association module may be configured to use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information;

The third sending module may be used to send the information stream to the client for display on the client.

Based on the method provided in the foregoing embodiment, the embodiment of this specification also provides an information flow display method, which is applied to the client and may include:

Receive the information stream sent by the server and display it.

Based on the method provided by the foregoing embodiment, an embodiment of this specification also provides a client, which may include:

The fourth sending module may be used to send an information flow acquisition request to the server, where the information flow acquisition request includes the input information obtained by the client; so that the server receives the information flow acquisition request and extracts one or the other from the input information Multiple entity information as target entity information; and extracting at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, the association relationship The information includes information describing the direction and type of association between entity information; and, using the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information , And send the information stream to the client;

The second receiving module can be used to receive the information stream sent by the server;

The first display module can be used to display the information stream.

The method or device described in the foregoing embodiment provided in this specification can implement business logic through a computer program and be recorded on a storage medium, and the storage medium can be read and executed by a computer to achieve the effects of the solution described in the embodiment of this specification. Therefore, this specification also provides an information flow display device that includes a processor and a memory storing executable instructions of the processor. When the instructions are executed by the processor, the steps including the method described in any one of the foregoing embodiments are implemented.

In other implementation scenarios, users often use information retrieval methods when performing technical queries, infringement analysis, etc., to find out the data source or the patentee behind the data source. However, there may be many different description forms of information in the same data source or different data sources, and even a large amount of incomplete and non-descriptive information, which leads to the problem of inaccuracy and incompleteness of the searched information. In addition, users need to spend a lot of energy to filter and sort out when viewing, and the experience is poor. Correspondingly, as shown in FIG. 25, in other embodiments of this specification, an information retrieval method may also be provided, and the method may include:

S60: The client sends a search request, where the search request includes input information obtained by the client;

S62: The server receives the search request sent by the client, and extracts entity information from the input information, where the entity information includes an entity type and an entity value;

S64: The server determines the data source where the entity information is located according to the associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information, and the association relationship information includes a pair of entity information. Information describing the direction of the association and the type of association;

S66: The server sends the search result to the client, so that the client can display it.

The input information may be reference information for information retrieval, and the server uses the input information as a reference to retrieve a data source that directly describes or indirectly describes the input information. The associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information can be implemented with reference to the embodiments in the above-mentioned information flow extraction method, and will not be repeated here.

Fig. 26 shows a search interface for searching for "Fullerene as polymer additives" using keywords and a schematic diagram of the search results. The darker gray scale shown in FIG. 26 is the non-correlated search, and the weaker gray scale is the relevant search result. As shown in Figure 26, only 3 related patents were obtained through keyword search. Fig. 27 is a schematic diagram of a retrieval interface using the solution of the embodiment of the present specification to perform information retrieval based on material information flow. FIG. 28 is a schematic diagram of retrieval results using the solution of the embodiment of this specification, and all related searches are accurately extracted by using the solution of the embodiment of this specification.

Many data sources may not have specific information about the expression of "Fullerene as (as) polymer additive", or even if there is "Fullerene (fullerene) as (as) polymer additive" The expression of is not necessarily the same as the expression entered by the user, resulting in many data sources using keyword matching methods, which cannot be effectively retrieved. Or, when the user inputs a keyword, he can also use an "or" method to separate different keyword information, but in this way, more noise information may be retrieved.

Using the solution of the embodiment of this specification, by searching the associated entity information of the entity information in the input information and the association relationship information with the associated entity information from the data source, even if the data source is not clear or specifically exists "Fullerene" as The description information of "additive" can still use entity information: Additive entity-Fullerene, through the associated entity information and the association relationship information with the associated entity information, accurately locate the data source describing the entity information, and then effectively extract the information about the entity. All data sources that describe the role of "Fullerene" as the "additive", therefore, can effectively separate the signal (related search) from the noise (irrelevant search), greatly improving the search accuracy. Further restricting the condition "polymer" can eliminate the data source describing "Fullerene" as an "additive" in the information flow corresponding to other material types, and further improve the accuracy of information retrieval.

In some embodiments, the name variant of "Fullerene" can be further extracted, that is, the parity information, such as the Chinese description of "Fullerene". By further updating the parity information to the reference information corresponding to the search, the accuracy and comprehensiveness of the search results can be further improved.

In some implementations, it is also possible to perform classified retrieval according to material structure or material application during information retrieval. For example, it is possible to add restrictions according to the material structure or material application in the input information during information retrieval.

The material structure may refer to the way in which the atoms (or ions, molecules) constituting the material are combined with each other or the form of formation (ie, structural elements), and the combination, arrangement, and various connections between the structural elements in a certain order. Different materials have different structural elements. For example, various phases, structures, defects, monomers, and macromolecular chains of materials are all structural elements of materials. Materials are classified according to material structure, such as polymer, metal, ceramic, biological, composite and other material classifications. For the above classifications, there can be further detailed classifications, as shown in Figure 31. Figure 31 is a schematic diagram showing the classification of polymers further refined according to the material structure.

The material application may include the application field or specific application corresponding to the material. The application field classification can include building materials, energy materials, printing materials, optoelectronic materials, and so on. Different application fields can be further divided up to specific applications, such as organic electric display coatings in optoelectronic materials. As shown in Figure 33, Figure 33 shows a further detailed classification diagram of smart materials.

There may be big differences in the entity types corresponding to materials of different material structures or the association relationship between each entity information, and the information flow characteristics from the material monomer to the material application. By pre-classifying the material structure, based on Retrieving the material structure category to which the material to be retrieved, material application, or material technology belongs to can further improve the accuracy of entity information and the association relationship information between entity information, so that the retrieved data source is more in line with the needs of users.

Corresponding to different application fields or corresponding to different specific applications, the involved raw materials, processing technology, manufacturing methods, etc. may also be quite different. Through further refinement of material applications, entity information and entity information are based on material applications. Information extraction of the association relationship between information can also further improve the accuracy of information extraction, reduce noise interference, and make the retrieved data source more in line with the needs of users. At the same time, for some data sources, such as patents, papers, etc., many retrieval systems can also be classified according to application fields, as shown in Figure 29. The left figure of Fig. 29 shows a larger application category, and the right figure shows an application category that further embodies a certain application category in the left figure. Therefore, by considering the material application when searching, it is possible to search from the data source corresponding to the corresponding application field in a targeted manner, which greatly improves the retrieval efficiency and retrieval accuracy.

Especially for the retrieval of image information, classification and retrieval according to the material structure and material application can reduce the difficulty of detecting material graphics at the fine-grained level, which helps to detect different materials and material patterns from the image, and improve the retrieval accuracy .

For example, as shown in Figure 34 to Figure 36, Figure 34 shows a schematic diagram of the search results retrieved using keywords; Figure 35 shows a schematic diagram of the retrieval interface provided by an embodiment of this specification; Schematic diagram of search results. It can be seen from Figure 36 that the embodiments of this specification further limit the application field when searching, and the server can further consider the patents to which the material is applied, and then extract entity information and association relationship information between entity information. Improve the accuracy of image search, image extraction, image object detection and image classifier results.

For example, as shown in FIG. 37 and FIG. 38, FIG. 37 shows a schematic diagram of a protein structure, a DNA plasmid and a microstructure of a material, and FIG. 38 shows a schematic diagram of a circuit diagram and a flow chart. When searching for protein structure, DNA plasmid and material microstructure, by considering the classification of material structure, patents corresponding to noise images with specific image types (such as circuit diagrams, flowcharts, etc.) can be effectively eliminated, and the protein structure, Image of DNA plasmid and material microstructure.

Based on the foregoing application scenario example, in some embodiments, the information input interface displayed by the client may further include an information input area, a first selection list, and/or a second selection list. The information input area may be used for the user to input information, and the first selection list and the second selection list may be used for the user to select information. The first selection list may include material application or material structure category information. The second selection list may include category information of the entity type. Correspondingly, the client terminal obtains input information based on the information input area, the first selection list, and/or the second selection list.

As shown in Figure 27, the user can perform information retrieval by entering information in the input box and limiting the entity type of the entered information. Information retrieval in this way can facilitate accurate extraction of entity types and entities from the entered retrieval information. Value to improve the accuracy of information retrieval.

As shown in Figure 27 and Figure 35, the application field or structure category can also be limited. By limiting the application field or structure category, it is convenient to determine the entity information corresponding to the corresponding application field or structure category and the extraction of the association relationship information between the entity information Features to improve the accuracy of information extraction. At the same time, the preset material structure and material application classification selection method can also make the input information more standardized and improve the accuracy of information retrieval.

In some embodiments, the first selection list or the second selection list may also adopt an interactive visualization format to display the information to be selected. The interactive visualization format means that the entity type or material structure, the category information of the material application is displayed in a visual form, and the corresponding category and subcategory list selection format is determined by receiving the trigger operation of each category information, as shown in Figure 30 to Figure. 33 shown.

Figure 30 shows several larger categories of material structure display diagrams. The user can click on the material structure, and the client can determine that the user has selected the material structure as part of the input information based on the user's click operation. The client can further display several larger categories. The user can further click on any of these categories, such as clicking on polymers. The client can determine that the user has further selected the polymer as part of the input information based on the user's click operation. The client can further display the detailed classification of polymers, as shown in Figure 31. The user can further click on any of the categories in FIG. 31, and the client can determine the detailed category further selected by the user as part of the input information based on the user's click operation. Of course, the user's operation method can also adopt other methods. Figure 32 shows the display diagrams of several larger categories of material applications, and Figure 33 shows the detailed classification diagram of smart materials. The triggering of material application and the display implementation can be the same as the material structure, which will not be repeated here.

By adopting an interactive visualization format as the form of the first selection list or the second selection list, the display of the material structure and the detailed classification of the material application can also be more convenient, and at the same time, the interactivity can be improved, and the user experience can be improved.

In other embodiments, it is also possible to link material applications and material properties, and display properties related to specific applications, as shown in Figure 39, which shows the 10 properties of smart materials. The internal structure of the material can be changed with changes in chemical composition and external conditions, thereby changing the properties of the material. For example, low carbon steel with carbon mass fraction below 0.25% usually has good plasticity and toughness, but low strength and hardness; high carbon steel with carbon mass fraction in the range of 0.6% to 1.4% has higher strength and hardness. High, but poor plasticity and toughness. Different properties of materials have a greater impact on the actual application of materials. For example, materials with different yield strengths have different application fields. By establishing an association between material application and material properties, and by describing different units/measurements of material properties in the data source, the extraction of material application information can be effectively determined and verified, and the accuracy of information extraction and retrieval can be improved.

Based on the method provided in the foregoing embodiment, the embodiment of this specification also provides an information retrieval method, which is applied to a server and may include:

The search result is sent to the client, so that the client can display it.

The third receiving module may be used to receive a search request sent by the client, where the search request includes input information obtained by the client;

The seventh extraction module can be used to extract entity information from the input information, where the entity information includes an entity type and an entity value;

The data source determining module may be used to determine the data source where the entity information is located according to the associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information, and the association relationship information includes Information describing the direction and type of association between entity information;

The fifth sending module may be used to send the search result to the client, so that the client can display it.

Based on the method provided in the foregoing embodiment, the embodiment of this specification also provides an information retrieval method, which is applied to the client and may include:

Send a search request to the server, the search request includes the input information obtained by the client; so that the server extracts the entity information from the input information, the entity information includes the entity type and the entity value; according to the entity information in the data source The associated entity information in and the association relationship information between the entity information and the associated entity information determine the data source where the entity information is located, and the association relationship information includes information describing the association direction and association type between the entity information; And sending the retrieval result to the client;

The retrieval result sent by the server is received and displayed.

In other embodiments, the information input interface displayed by the client may include an information input area, a first selection list, and/or a second selection list; the information input area may be used for information input; the first The selection list and the second selection list may be used for information selection, the first selection list may include category information of material applications or material structure; the second selection list may include category information of entity types;

Correspondingly, the client can obtain input information based on the information input area, the first selection list, and/or the second selection list.

In other embodiments, the first selection list and/or the second selection list may use an interactive visualization format to display the information to be selected; wherein, the interactive visualization format may indicate that the entity type is displayed in a visualization form or The material structure, the category information of the material application, and the list selection format of the corresponding category and subcategory are determined by receiving the trigger operation of each category information.

The sixth sending module can be used to send a search request to the server, where the search request includes input information obtained by the client; so that the server extracts entity information from the input information, and the entity information includes entity type and entity value; The data source where the entity information is located is determined according to the associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information, and the association relationship information includes the direction of association between the entity information And information describing the associated type; and sending the search result to the client;

The fourth receiving module may be used to receive the retrieval result sent by the server;

The second display module can be used to display the search results.

In other embodiments, the client may further include:

The input interface display module may be used to display an information input interface. The information input interface may include an information input area, a first selection list and/or a second selection list; the information input area is used for information input; A selection list and a second selection list can be used for information selection, the first selection list can include category information of material applications or material structure; the second selection list can include category information of entity types;

The input information obtaining module may be used to obtain input information based on the information input area, the first selection list and/or the second selection list.

The method or device described in the foregoing embodiment provided in this specification can implement business logic through a computer program and be recorded on a storage medium, and the storage medium can be read and executed by a computer to achieve the effects of the solution described in the embodiment of this specification. Therefore, this specification also provides an information retrieval device that includes a processor and a memory storing processor-executable instructions. When the instructions are executed by the processor, the steps including the method described in any one of the foregoing embodiments are implemented.

Generally, a patent or dissertation document may describe a large amount of information, and the abstract information given by the author or applicant cannot effectively reflect the main information of the entire patent or dissertation document. When making technical inquiries, due to the increase in the amount of information, users are usually accustomed to preliminarily determining whether the required information exists in the current literature only through abstracts, which leads to missing information. However, if the omission is avoided, the user will need to spend a lot of energy to consult and analyze the entire content of the literature, which will lead to time-consuming and labor-intensive. Correspondingly, as shown in FIG. 40, in other embodiments of this specification, a method for generating summary information may also be provided, and the method may include:

S80: The client sends a summary information generation request, where the summary information generation request includes the data source for which the summary is to be generated;

S82: The server receives the summary information generation request sent by the client;

S84: The server extracts the entity information and the association relationship information between the entity information from the data source, and associates the corresponding entity information according to the extracted association relationship information; wherein the association relationship information includes the relationship between the entity information Information describing the direction of the association and the type of association;

S86: The server generates summary information according to the associated entity information, and sends the generated summary information to the client, so that the client can display it.

The data source may include the patent application text, thesis or other types of information text for which abstract information is to be generated. After the server obtains the corresponding information text, it can use NLP to process and tokenize the original information text. After marking, based on a pre-built entity information dictionary or entity information database, NER can be used to extract entity information in the information text, and extract the material name entity or material application entity. And take the extracted entity information as the object to be analyzed.

Then, the server can use the NLP algorithm to obtain the metadata of the object to be analyzed. For example, the entity information in the context information of the object to be analyzed can be extracted, and a filtering algorithm can be used to delete noise words. The noise words include stop words, common words, non-scientific words and words that have nothing to do with the material context. Then, the filtered entity information can be attached to these objects as metadata.

Then, the metadata of the object to be analyzed can be compared to establish an association relationship between the two objects to be analyzed. Then, you can use the association relationship to associate the object to be analyzed. In this way, the association relationship between the important entity information in the information text is established, the information summary is formed, and the information is fed back to the client for display. The presentation form of the information summary may be a piece of text information generated based on the entity information and the association relationship information between the entity information, or may be an information flow formed based on the entity information and the association relationship information between the entity information.

For multiple information texts, it is also possible to compare the entity information and the association relationship information extracted from the two information texts to find the similarity between the entity information and the association relationship information between the two information texts. Entity similarity can be considered by comparing multiple attributes related to different contexts and metadata. Each entity may have a unique threshold for entity matching. For example, it is possible to find similarities/matches between entities based on entity thresholds, and combine entities and relationships to form important workflow links between two documents.

The extraction of the entity information and the association relationship information between the entity information can refer to the implementation in the above-mentioned information flow extraction method, which will not be repeated here. The summary information generated by the above method can accurately and comprehensively extract the entity information and the association relationship between the entity information in the data source information, avoid user screening and sorting, and enable users to find the information they need more quickly and accurately .

The generated summary information may be displayed in the form of text or tables, or may be displayed in the form of information flow. The extraction method of the corresponding information flow can be performed with reference to the above-mentioned embodiments. The display of summary information in the manner of information flow can make it easier for users to view and improve the user experience.

Based on the method provided in the foregoing embodiment, the embodiment of this specification also provides a method for generating summary information, which is applied to a server and may include:

The fifth receiving module may be used to receive a summary information generation request sent by the client, where the summary information generation request includes the data source for which the summary is to be generated;

The eighth extraction module can be used to extract entity information and association relationship information between entity information from the data source; wherein, the association relationship information includes information describing the association direction and association type between entity information ；

The sixth association module can be used to associate corresponding entity information according to the extracted association relationship information;

The generating module is used to generate summary information based on the associated entity information;

The seventh sending module may be used to send the generated summary information to the client, so that the client can display it.

Based on the method provided in the foregoing embodiment, the embodiment of this specification also provides a method for generating summary information, which is applied to the client and may include:

Send a summary information generation request to the server, the summary information generation request including the data source for which the summary is to be generated; so that the server extracts the entity information and the association relationship information between the entity information from the data source, based on the extracted association relationship information Associate corresponding entity information; wherein, the association relationship information includes information describing the direction of association and type of association between entity information; and, generating summary information based on the associated entity information, and sending the generated summary information The client;

Receive the summary information sent by the server and display it.

The eighth sending module may be used to send a summary information generation request to the server, where the summary information generation request includes the data source for which the summary is to be generated; so that the server performs entity information and the association relationship information between the entity information on the data source Extracting, associating corresponding entity information according to the extracted association relationship information; wherein the association relationship information includes information describing the association direction and association type between entity information; and, generating a summary based on the associated entity information Information, sending the generated summary information to the client;

The fifth receiving module may be used to receive the summary information sent by the server;

The third display module can be used to display the summary information.

The method or device described in the foregoing embodiment provided in this specification can implement business logic through a computer program and be recorded on a storage medium, and the storage medium can be read and executed by a computer to achieve the effects of the solution described in the embodiment of this specification. Therefore, this specification also provides a summary information generating device, including a processor and a memory storing processor-executable instructions, which, when executed by the processor, implement the steps including the method described in any one of the foregoing embodiments.

The storage medium may include a physical device for storing information, and usually the information is stored in an electric, magnetic, or optical medium after digitizing the information. The storage medium may include: devices that use electrical energy to store information, such as various types of memory, such as RAM, ROM, etc.; devices that use magnetic energy to store information, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, bubble memory, U disk; a device that uses optical means to store information, such as CD or DVD. Of course, there are other ways of readable storage media, such as quantum memory, graphene memory, and so on.

This specification also provides a system, which can be a separate information flow extraction system, or an information flow display system, or an information retrieval system, or an abstract generation system, and can also be applied to a variety of information extraction systems. The system can be a single server, or it can include server clusters, systems (including distributed systems), software (applications), The actual operation device, logic gate circuit device, quantum computer, etc., combined with the terminal device necessary to implement the hardware. The information retrieval system may include at least one processor and a memory storing computer-executable instructions, and the processor implements the steps of the method in any one or more of the foregoing embodiments when the processor executes the instructions.

It should be noted that the above-mentioned system may also include other implementation manners according to the description of the method or device embodiment. For the specific implementation manner, refer to the description of the related method embodiment, and details are not repeated here.

The embodiments of this specification are not limited to the conditions described in the embodiments of this specification that must conform to the standard data model/template. Certain industry standards or implementations described in custom methods or examples with slight modifications can also achieve the same, equivalent or similar implementation effects of the foregoing examples, or predictable implementation effects after modification. The examples obtained by applying these modified or deformed data acquisition, storage, judgment, processing methods, etc., can still fall within the scope of the optional implementation solutions of this specification.

This specification is described with reference to flowcharts and/or block diagrams of methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and/or block in the flowchart and/or block diagram, and the combination of processes and/or blocks in the flowchart and/or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing equipment to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing equipment are used to generate It is a device that realizes the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

It should also be noted that the terms "include", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or equipment including a series of elements includes not only those elements, but also Other elements that are not explicitly listed, or include elements inherent to such processes, methods, commodities, or equipment. If there are no more restrictions, the element defined by the sentence "including a..." does not exclude the existence of other same elements in the process, method, or device that includes the element.

The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, as for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for related parts, please refer to the part of the description of the method embodiment. In the description of this specification, descriptions with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" etc. mean specific features described in conjunction with the embodiment or example , Structure, materials or features are included in at least one embodiment or example in this specification. In this specification, the schematic representation of the above-mentioned terms does not necessarily refer to the same embodiment or example. Moreover, the described specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine the different embodiments or examples and the features of the different embodiments or examples described in this specification without contradicting each other.

The above descriptions are only examples of this specification, and are not intended to limit this specification. For those skilled in the art, this specification can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this specification shall be included in the scope of the claims of this specification.

Claims

An information flow display method, including:

The client sends an information flow acquisition request to the server, where the information flow acquisition request includes the input information acquired by the client;

The server receives the information flow acquisition request, and extracts one or more entity information from the input information as target entity information;

The server extracts at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes the direction of the association between the entity information and Information describing the type of association;

The server uses the association relationship information to associate the target entity information and the associated entity information, obtains the information flow corresponding to the target entity information, and sends the information flow to the client for the Display on the client.
An information flow extraction method includes:

Obtain target entity information;

At least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information are extracted from the data source, and the association relationship information includes an association direction and an association between the entity information Type description information;

Use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information.
The method according to claim 2, wherein the using the association relationship information to associate the target entity information with the associated entity information comprises:

Obtain at least one associated entity information extracted from the data source and the association relationship information between the target entity information and the corresponding associated entity information, and use the association relationship information between the target entity information and the corresponding associated entity information to The target entity information is associated with the corresponding associated entity information to obtain the sub-information stream of the target entity information;

Using the target entity information as the reference information, link sub-information streams obtained from the same or different data sources to obtain the information stream corresponding to the target entity information.
The method according to claim 2 or 3, wherein, when the association relationship information includes co-located entity description information, the associated entity information corresponding to the co-located entity description information is added to the target entity information, and all information is updated. The target entity information; wherein, the co-located entity description information includes description information in other expression forms that describe the associated entity information as target entity information;

Correspondingly, at least one associated entity information corresponding to the updated target entity information and the association relationship information between the target entity information and the associated entity information are extracted from the data source.
The method according to any one of claims 2 to 4, wherein the method further comprises:

Use the associated entity information extracted from the data source as the target entity information to extract the information flow, and obtain the information flow corresponding to the multiple target entity information;

Link the information streams corresponding to multiple target entity information with the corresponding target entity information as the reference information to obtain the information map.
The method according to any one of claims 2 to 5, wherein the extracting at least one associated entity information corresponding to the target entity information and the association relationship between the target entity information and the associated entity information from the data source Information, including:

Retrieve the data source where the target entity information is located;

Locate the context information where the target entity information is located in the data source;

When the context information includes a material process entity, a material name entity, and/or a material application entity, extract the associated entity information corresponding to the target entity information and the target entity information and the associated entity information according to the material process information association method The relationship information between.
The method according to any one of claims 2 to 6, wherein the method further comprises:

Extracting the target entity information and the commodity name of the associated entity information from the data source;

The manufacturer or supplier corresponding to the target entity information or associated entity information is extracted according to the product name, and the manufacturer or supplier is associated with the target entity information or associated entity information.
The method according to any one of claims 2 to 7, wherein the method further comprises:

Extract the target entity information and the parameter entity information of the associated entity information from the data source. The parameter entity information includes at least one of the material structure type entity, the process method entity, the material attribute entity, the unit entity or the measurement entity corresponding to one Entity information

The extracted parameter entity information is associated with the corresponding target entity information and associated entity information.
The method according to any one of claims 3 to 8, wherein the method further comprises: associating a sub-information stream of the target entity information with a corresponding data source.
The method according to claim 2, wherein the method further comprises:

Use the entity information in the information flow as an information node for interaction with the user, and use a visualization method to visualize the information flow;

Sending the processed information stream to the client, so that the user can view the information stream and trigger the information node through the client;

Based on the user's triggering operation on the information node, feedback other information associated with the information node to the client.
An information flow extraction device includes:

The first obtaining module is used to obtain target entity information;

The first extraction module is configured to extract at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes pair entity information Information describing the direction of the association and the type of association;

The first association module is configured to use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information.
The device according to claim 11, wherein the first association module comprises:

The first associating unit is used to obtain at least one associated entity information extracted from the data source and the association relationship information between the target entity information and the corresponding associated entity information, and use the relationship between the target entity information and the corresponding associated entity information The association relationship information between each other associates the target entity information with the corresponding associated entity information, and obtains the sub-information stream of the target entity information;

The second associating unit is used to link the sub-information streams obtained from the same or different data sources using the target entity information as the reference information to obtain the information stream corresponding to the target entity information.
The apparatus according to claim 11 or 12, wherein, when the association relationship information includes colocation entity description information, the apparatus further comprises:

The update module is used to add the associated entity information corresponding to the colocation entity description information to the target entity information, and update the target entity information; wherein the colocation entity description information includes information describing the associated entity as the target entity Descriptive information in other forms of information;

The first extraction module is further configured to extract at least one associated entity information corresponding to the updated target entity information and the association relationship information between the target entity information and the associated entity information from the data source.
The device according to any one of claims 11 to 13, wherein the device further comprises:

The second extraction module is used to extract the information flow of each associated entity information extracted from the data source as the target entity information, and obtain the information flow corresponding to the multiple target entity information;

The second association module is used to link the information streams corresponding to multiple target entity information with the corresponding target entity information as the reference information to obtain the information graph.
The device according to any one of claims 11 to 14, wherein the first extraction module comprises:

The retrieval unit is used to retrieve the data source where the target entity information is located;

The positioning unit is used to locate the context information where the target entity information is located in the data source;

The extraction unit is used to extract the associated entity information corresponding to the target entity information and the target entity according to the material process information association method when the context information includes material process entities, material name entities, and/or material application entities Information about the association relationship between information and associated entity information.
The device according to any one of claims 11 to 15, wherein the device further comprises:

The third extraction module is used to extract the target entity information and the commodity name of the associated entity information from the data source;

The third associating module is used to extract the manufacturer or supplier corresponding to the target entity information or the associated entity information according to the product name, and perform the comparison between the manufacturer or supplier and the target entity information or associated entity information. Associated.
The device according to any one of claims 11 to 16, wherein the device further comprises:

The fourth extraction module is used to extract the target entity information and the parameter entity information of the associated entity information from the data source. The parameter entity information includes at least a material structure type entity, a process method entity, a material attribute entity, a unit entity, or a measurement entity Entity information corresponding to one of the types;

The fourth association module is used to associate the extracted parameter entity information with corresponding target entity information and associated entity information.
The device according to any one of claims 12 to 17, wherein the first association module further comprises:

The third associating unit is used for associating the sub-information stream of the target entity information with the corresponding data source.
The device according to any one of claims 11 to 18, wherein the device further comprises:

The visualization processing module is used to treat each entity information in the information flow as an information node for interacting with the user, and use a visualization method to visualize the information flow;

The first sending module is configured to send the processed information stream to the client, so that the user can view the information stream and trigger the information node through the client;

The second sending module is configured to feed back other information associated with the information node to the client based on the user's triggering operation on the information node.
An information flow extraction device, the device comprising a processor and a memory for storing executable instructions of the processor, and when the instructions are executed by the processor, the steps of the method according to any one of claims 2 to 10 are realized .
An information flow display method, applied to a server, including:

Receiving an information flow acquisition request sent by the client, where the information flow acquisition request includes input information acquired by the client;

Extract one or more entity information from the input information as target entity information;

At least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information are extracted from the data source, and the association relationship information includes the association direction and the association between the entity information Type description information;

Using the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information;

The information stream is sent to the client for display on the client.
A server that includes:

The first receiving module is configured to receive an information flow acquisition request sent by the client, where the information flow acquisition request includes input information acquired by the client;

The fifth extraction module is used to extract one or more entity information from the input information as target entity information;

The sixth extraction module is configured to extract at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes pair entity information Information describing the direction of the association and the type of association;

A fifth association module, configured to use the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information;

The third sending module is configured to send the information stream to the client for display on the client.
An information flow display method, applied to the client, includes:

Send an information flow acquisition request to the server, where the information flow acquisition request includes the input information obtained by the client; so that the server receives the information flow acquisition request and extracts one or more entity information from the input information as the target entity Information; and, extracting at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, and the association relationship information includes the relationship between the entity information Information describing the association direction and association type; and, using the association relationship information to associate the target entity information and the associated entity information, obtain the information flow corresponding to the target entity information, and send the information flow To the client;

Receive the information stream sent by the server and display it.
A client, including:

The fourth sending module is configured to send an information flow acquisition request to the server, where the information flow acquisition request includes the input information obtained by the client; so that the server receives the information flow acquisition request and extracts one or more information from the input information. Entity information as target entity information; and extracting at least one associated entity information corresponding to the target entity information and the association relationship information between the target entity information and the associated entity information from the data source, the association relationship information Including information describing the direction and type of association between entity information; and using the association relationship information to associate the target entity information and the associated entity information to obtain the information flow corresponding to the target entity information, And send the information stream to the client;

The second receiving module is used to receive the information stream sent by the server;

The first display module is used to display the information stream.
An information retrieval method applied to a server, including:

Receiving a search request sent by the client, where the search request includes input information obtained by the client;

Extracting entity information from the input information, where the entity information includes an entity type and an entity value;

The data source where the entity information is located is determined according to the associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information, and the association relationship information includes the direction of association between the entity information And information describing the type of association;

The search result is sent to the client, so that the client can display it.
A server that includes:

The third receiving module is configured to receive a search request sent by the client, where the search request includes input information obtained by the client;

The seventh extraction module is used to extract entity information from the input information, where the entity information includes entity type and entity value;

The data source determining module is configured to determine the data source where the entity information is located according to the associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information, and the association relationship information includes Information describing the direction of association between entity information and the type of association;

The fifth sending module is configured to send the search result to the client, so that the client can display it.
An information retrieval method, applied to the client, includes:

Send a search request to the server, the search request includes the input information obtained by the client; so that the server extracts the entity information from the input information, the entity information includes the entity type and the entity value; according to the entity information in the data source The associated entity information in and the association relationship information between the entity information and the associated entity information determine the data source where the entity information is located, and the association relationship information includes information describing the association direction and association type between the entity information; And sending the retrieval result to the client;

The retrieval result sent by the server is received and displayed.
The method according to claim 27, wherein the information input interface displayed by the client terminal includes an information input area, a first selection list and/or a second selection list; the information input area is used for information input; The first selection list and the second selection list are used for information selection, the first selection list includes category information of material applications or material structure; the second selection list includes category information of entity types;

Correspondingly, the client terminal obtains input information based on the information input area, the first selection list, and/or the second selection list.
The method according to claim 28, wherein the first selection list and/or the second selection list adopts an interactive visualization format to display the information to be selected; wherein, the interactive visualization format means that the information to be selected is displayed in a visualized form The entity type or material structure, the category information of the material application, and the corresponding category and subcategory list selection format are determined by receiving the trigger operation of each category information.
A client, including:

The sixth sending module is configured to send a search request to the server, the search request includes input information obtained by the client; so that the server extracts entity information from the input information, and the entity information includes entity type and entity value; The associated entity information of the entity information in the data source and the association relationship information between the entity information and the associated entity information determine the data source where the entity information is located, and the association relationship information includes the association direction between the entity information and Information describing the type of association; and sending the search result to the client;

The fourth receiving module is configured to receive the retrieval result sent by the server;

The second display module is used to display the retrieval result.
The client according to claim 30, wherein the client further comprises:

The input interface display module is used to display an information input interface. The information input interface includes an information input area, a first selection list and/or a second selection list; the information input area is used for information input; the first selection The list and the second selection list are used for information selection, the first selection list includes category information of material application or material structure; the second selection list includes category information of entity type;

The input information obtaining module is configured to obtain input information based on the information input area, the first selection list and/or the second selection list.
A method for generating summary information, applied to a server, includes:

Receiving a summary information generation request sent by the client, where the summary information generation request includes the data source for which the summary is to be generated;

The entity information and the association relationship information between the entity information are extracted from the data source, and the corresponding entity information is associated according to the extracted association relationship information; wherein the association relationship information includes the association direction between the entity information and Information describing the type of association;

The summary information is generated according to the associated entity information, and the generated summary information is sent to the client, so that the client can display it.
A server that includes:

The fifth receiving module is configured to receive a summary information generation request sent by the client, where the summary information generation request includes the data source for which the summary is to be generated;

The eighth extraction module is used to extract entity information and the association relationship information between the entity information from the data source; wherein the association relationship information includes information describing the association direction and the association type between the entity information;

The sixth association module is used to associate corresponding entity information according to the extracted association relationship information;

The generation module is used to generate summary information based on the associated entity information;

The seventh sending module is configured to send the generated summary information to the client, so that the client can display it.
A method for generating summary information, applied to a client, includes:

Send a summary information generation request to the server, the summary information generation request including the data source for which the summary is to be generated; so that the server extracts the entity information and the association relationship information between the entity information from the data source, based on the extracted association relationship information Associate corresponding entity information; wherein, the association relationship information includes information describing the direction of association and type of association between entity information; and, generating summary information based on the associated entity information, and sending the generated summary information The client;

Receive the summary information sent by the server and display it.
A client, including:

The eighth sending module is configured to send a summary information generation request to the server, where the summary information generation request includes the data source for which the summary is to be generated; so that the server extracts entity information and the association relationship information between the entity information from the data source , Associate the corresponding entity information according to the extracted association relationship information; wherein the association relationship information includes information describing the association direction and association type between the entity information; and, generate summary information based on the associated entity information , Sending the generated summary information to the client;

The fifth receiving module is configured to receive the summary information sent by the server;

The third display module is used to display the summary information.