CN114328875A - Question answering method, question answering model building method, electronic equipment and storage medium - Google Patents

Abstract

The invention discloses a question-answering method, a question-answering model construction method, electronic equipment and a storage medium. In the question answering method, user question information is obtained; determining answers corresponding to the user question information based on a question-answer model; the question-answering model is constructed based on a question-answering material table, the question-answering material table comprises at least one header attribute, each header attribute tag is provided with at least one corresponding attribute data, each header attribute is provided with corresponding question-answering marking information, and the question-answering marking information comprises question marking information or answer marking information. Therefore, the question-answer model based on the table document is quickly constructed, and the constructed question-answer model can accurately reply the aggregation calculation type question.

Description

Question answering method, question answering model building method, electronic equipment and storage medium

Technical Field

The invention belongs to the technical field of internet, and particularly relates to a question and answer method, a question and answer model construction method, electronic equipment and a storage medium.

Background

The current artificial intelligence is developed rapidly, the AI field can be deeply mined based on mass data provided by users, and knowledge question answering is provided for the users through various technologies such as knowledge inquiry, reasoning, reverse reasoning, algorithm and the like. Currently supported technologies are: QA question-answer pairs, knowledge maps and other technologies, and the technologies are used for providing multi-field and fine-granularity knowledge question-answers for users.

After the user uploads the domain-level original data, the underlying engine technology processing generally answers the user's questions after the engine processing through several links of knowledge construction, knowledge acquisition, knowledge fusion, knowledge inference and knowledge application. However, the current engine has weak question answering function and is complicated to construct and operate.

In addition, the answer technology in the market generally obtains answers based on a text similarity algorithm, a keyword algorithm, or a knowledge graph deep association analysis, but for question types such as "aggregate type", "data trend search", "statistics by attribute", and the like, for example, "how many employees a company has? "vehicle-mounted access amount equalization" and "access amount analysis by provinces" are not supported.

In view of the above problems, the industry has not provided a better solution for the moment.

Disclosure of Invention

The embodiment of the invention provides a question-answering method, a question-answering model construction method, electronic equipment and a storage medium, which are used for solving at least one of the technical problems.

In a first aspect, an embodiment of the present invention provides a question answering method, including: acquiring user question information; determining answers corresponding to the user question information based on a question-answer model; the question-answering model is constructed based on a question-answering material table, the question-answering material table comprises at least one header attribute, each header attribute tag is provided with at least one corresponding attribute data, each header attribute is provided with corresponding question-answering marking information, and the question-answering marking information comprises question marking information or answer marking information.

In a second aspect, an embodiment of the present invention provides a method for constructing a question-answer model, including: determining at least one header tag attribute and corresponding attribute data in the initial material table; acquiring question and answer labeling information aiming at each header label attribute; the question and answer labeling information comprises question labeling information or answer labeling information; determining a question and answer material table based on the corresponding header tag attributes, the question and answer marking information and the attribute data; and constructing a question-answer model based on the question-answer material table.

In a third aspect, an embodiment of the present invention provides an electronic device, including: the computer-readable medium includes at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the steps of the above-described method.

In a fourth aspect, the present invention provides a storage medium, in which one or more programs including execution instructions are stored, where the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to perform the steps of the above-mentioned method of the present invention.

In a fifth aspect, the present invention also provides a computer program product, which includes a computer program stored on a storage medium, the computer program including program instructions, which when executed by a computer, cause the computer to perform the steps of the above method.

The embodiment of the invention has the beneficial effects that:

by utilizing the question-answer model constructed based on the question-answer material table, accurate answer can be better realized aiming at the questions with different header attributes proposed by the user, and accurate reply of aggregated calculation question sentences such as aggregation type, data searching trend, attribute statistics and the like can be realized. In addition, developers can train and construct corresponding question-answer models only by labeling corresponding questions or answers with the attribute of each header label in the initial material table, and development difficulty of the question-answer models is reduced.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on the drawings without creative efforts.

FIG. 1 illustrates a flow diagram of one example of a question answering method in accordance with an embodiment of the present invention;

FIG. 2 is a flow chart illustrating an example of a method of constructing a question-answer model according to an embodiment of the present invention;

FIG. 3 is a flow chart illustrating an example of a method of constructing a question-answer model according to an embodiment of the present invention;

FIG. 4 is a diagram illustrating an example of an initial material table according to an embodiment of the present invention;

FIG. 5 is a diagram illustrating a display interface of an electronic device after a form is imported;

FIG. 6 illustrates a training publication production consumption flow diagram for a question-answering model, according to an embodiment of the invention;

FIG. 7 is a diagram illustrating test results for an example of a question-and-answer model test, according to an embodiment of the present invention;

fig. 8 is a schematic diagram of a hardware structure of an electronic device that executes a question-answering method or a question-answering model building method according to another embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict.

The invention may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

As used herein, a "module," "system," and the like are intended to refer to a computer-related entity, either hardware, a combination of hardware and software, or software in execution. In particular, for example, an element may be, but is not limited to being, a process running on a processor, an object, an executable, a thread of execution, a program, and/or a computer. Also, an application or script running on a server, or a server, may be an element. One or more elements may be in a process and/or thread of execution and an element may be localized on one computer and/or distributed between two or more computers and may be operated by various computer-readable media. The elements may also communicate by way of local and/or remote processes based on a signal having one or more data packets, e.g., from a data packet interacting with another element in a local system, distributed system, and/or across a network in the internet with other systems by way of the signal.

Finally, it should be further noted that the terms "comprises" and "comprising," when used herein, include not only those elements but also other elements not expressly listed or inherent to such processes, methods, articles, or devices. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

It should be noted that the current question-answer model is generally implemented by using a chat algorithm, an hsqa algorithm, or a key algorithm, such as "QA question-answer pair" for a question-answer dataset. However, if the user problem is an inference or statistical type problem, it may result in an inability to reply correctly due to algorithm limitations.

At present, competitive products on the market do not support accurate reply to question types such as 'aggregation type', 'data searching trend', 'statistics by attribute' and the like. For example, the "convergent" question-answer type belongs to a functional computational question, a technical underlying database is generally an elastic search, and a non-relational database is mainly used for a full-text search tool, so that accurate question-answering cannot be realized for the computational question-answer.

Fig. 1 is a flowchart illustrating an example of a question answering method according to an embodiment of the present invention. Regarding the execution subject of the method of the embodiment of the present invention, it may be any electronic device with processing capability, for example, an electronic device installed with a question-answering software program.

As shown in fig. 1, in step 110, user question information is acquired. Here, the user question information may be of various types, such as text question information input by the user, or user question information obtained by voice-recognizing user question audio.

In step 120, based on the question-answer model, an answer corresponding to the user question information is determined.

Here, the question-answering model is constructed based on a question-answering material table, the question-answering material table includes at least one header attribute, each header attribute tag has at least one corresponding attribute data, each header attribute has corresponding question-answering tagging information, and the question-answering tagging information includes question tagging information or answer tagging information.

In some cases, the user questioning information includes: a computational question associated with a header tag attribute having question marking information. In this way, by using the question-answer model constructed based on the question-answer material table, accurate reply to the calculation type question of the types of aggregation type, statistics by attribute type and the like can be realized.

Fig. 2 is a flowchart illustrating an example of a question-answering model constructing method according to an embodiment of the present invention. The execution subject of the method according to the embodiment of the present invention may be any electronic device with processing capability, for example, an electronic device with a model building environment or a skill development platform, so that developers can develop corresponding question and answer products.

As shown in fig. 2, at least one header tag attribute and corresponding attribute data in the initial material table is determined in step 210. Here, the initial material table may be any spreadsheet document type, such as xls format or xlsx format, and the like, and the table import may be performed through an interface opened by the table skill entry interface.

In step 220, the question-answer tagging information for each header tag attribute is obtained. Here, the question and answer labeling information includes question labeling information or answer labeling information. Specifically, on one hand, the developer can label each header tag attribute in the initial material table, and label whether the attribute belongs to a condition or an answer. On the other hand, the electronic device can display the attribute of each header tag in the initial material table so that the user can perform corresponding labeling operation.

In step 230, a question and answer material table is determined based on each corresponding header tag attribute, question and answer tagging information, and attribute data.

In step 240, a question-answer model is constructed based on the question-answer material table. Illustratively, the dialogue model is trained using the question and answer material sheet, so as to construct a question and answer model with customized forms, and the form skills can realize the question and answer service capability.

FIG. 3 is a flowchart illustrating an example of a question-answering model construction method according to an embodiment of the present invention.

As shown in fig. 3, in the initial material table import phase 310, the developer can import a pre-prepared initial material table in an open-skills platform (e.g., a DUI platform) by triggering a specific control (e.g., a table import control). Here, the question-answer pair material provided by the developer can be implemented by using an excel table.

Fig. 4 is a diagram showing an effect of an example of the initial material table according to the embodiment of the present invention. Referring to the example in fig. 4, there are a plurality of header tag attributes (title, director, genre, etc.) and corresponding attribute data in the initial material table.

In the training sample set preparation stage 320, a developer may perform a question and answer tagging operation on each of the header tag attributes.

Continuing with the example of FIG. 4 described above, a schematic view of a display interface of the electronic device after the form import is shown with reference to FIG. 5. As shown in fig. 5, the electronic device displays each of the header label attributes and the corresponding question-and-answer tagging control. Here, the question-answering marking control is used for receiving an operation instruction so as to switch or confirm between the question marking option and the answer marking option. Referring to the example in fig. 5, a "column name" may represent a tab attribute, a question and answer labeling control corresponding to each tab attribute is shown at a "column type", and a user may select between a "condition" or an "answer". Furthermore, based on each question-answer tagging control, question-answer tagging information corresponding to each header tag attribute is determined, for example, a developer can determine corresponding question-answer tagging information through switching operation of the column type control.

In some service scenarios, a plurality of questions are in coupling and progressive relation, and a single question-answer interaction may not meet the question-answer requirement of the user. In view of this, the electronic device may further obtain a corresponding question-chasing technique for the header tag attribute having the question labeling information; and then, determining a question and answer material table based on the corresponding header label attribute, question and answer marking information, attribute data and question and answer operation. Referring to the example in fig. 5, a "start question following" control is provided behind the corresponding "column name" of the "column type" as the "condition", and the developer can select whether to turn on this function and input a specific "question following operation" when turning on. The column type and the data type are set aiming at the column name, if the column type is the condition, the open question following can be set, the question following dialect can be set, and multiple rounds of question answering operation are supported.

In conjunction with the example of "providing a question-answering service process" in fig. 1, after step 110, the electronic device may further identify whether a target question-answering technique associated with the user question information exists based on the question-answering model, where the question-answering material table further includes a question-answering technique for the header tag attribute with the question marking information. When the target question calling technique exists, the electronic device may output an answer corresponding to the user question information and the target question calling technique. Therefore, multiple rounds of question-answer interaction processes can be provided based on the question-answer model developed by the form skill.

In the question-answer model construction phase 330, a user may generate a form resource when clicking for training or publishing, and then train the dialogue model to generate a customized question-answer model of the form.

FIG. 6 illustrates a training publication production consumption flow diagram for a question-answering model, according to an embodiment of the invention. As shown in FIG. 6, the form resource may be generated by: firstly, calling resource service to store config.json skill information and cdm dialogue service information; then, calling the structure and data of the resource service storage table of the mscp-table-resource table to generate a dynamic sql table, and setting indexes for condition columns; and then, informing the NLQ engine service to generate semantic resources, and the engine service can call the pasc configuration storage service to acquire the cdma conversation resources and then access the mscp-table-resource table resource service library to generate resources.

Further, with continued reference to the example in fig. 6, after the training or publication is successful, the question-answer pairs may also be tested against the question-answer model. The specific test steps are as follows: firstly, a user request is forwarded to a DM dispatch central control scheduling service through a DDS (Data Distribution service); then, the DM service calls a pasc configuration storage service to acquire product configuration information and forwards the product configuration information to a downstream cdma server dialogue service and an NLQ form engine service; and then, the NLQ form engine service calls a pac to obtain the cdma dialogue resources, obtains form configuration, analyzes the corresponding result of the query of the 'form resource storage database' if the semantic resources exist, and outputs the result as an answer.

Fig. 7 is a test result diagram showing an example of a test performed on a question-and-answer model according to an embodiment of the present invention. As shown in fig. 7, the question-answer model trained according to the form skills can accurately answer the user question, and can also complete accurate answer to the aggregated calculation type question, for example, for the question "score is greater than 9.4" and "chinese", the output result is "kawangbiji".

It should be noted that, in the process of performing the question and answer service by the user, the question posed by the user may be relatively random, for example, an alias is used, and in order to accurately recognize the intention of the user question, the question and answer model needs to have the capability of recognizing an expanded word (for example, a synonym or a similar synonym).

Specifically, when determining the question and answer material table, the electronic device may obtain semantic extension words related to the attribute and/or attribute data of the header tag, for example, a developer inputs the semantic extension words. And then, determining a question and answer material table based on the corresponding header label attributes, question and answer labeling information, attribute data and semantic expansion words. Illustratively, in the training sample preparation stage 320, the excel form imported by the developer supports three sheets, the first sheet describing form attributes and data, the second sheet setting synonyms for form column names, and the third sheet setting synonyms for form data. Therefore, training samples are enriched by utilizing semantic expansion words, and the question-answering model can have semantic recognition capability aiming at the problems with the same condition attributes.

With reference to the example of "providing a question-answering service process" in fig. 1, regarding implementation details of step 120, the electronic device determines, based on the question-answering model, user header tag attributes semantically matching with the user question information, and determines, based on each attribute data corresponding to the user header tag attributes, an answer corresponding to the user question.

In some service scenarios of the embodiment of the invention, the form skill can be provided to the customer as a single atomic ability overhead skill open platform (for example, a DUI platform), and the customer only needs to customize the form skill material data and purchase the product usage to realize question answering. Here, the form skill can also be used as one item in the full link solution, and a full link product is created through recognition, semantics, dialogue and synthesis technologies and delivered to the customer, and the customer can use the full link product after opening a box, so that the personalized customization process of the question and answer service is simplified.

It should be noted that for simplicity of explanation, the foregoing method embodiments are described as a series of acts or combination of acts, but those skilled in the art will appreciate that the present invention is not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the invention. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required by the invention. In the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In some embodiments, the present invention provides a non-transitory computer-readable storage medium, in which one or more programs including execution instructions are stored, where the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to perform any one of the above-described question answering methods or question answering model building methods of the present invention.

In some embodiments, the present invention further provides a computer program product including a computer program stored on a non-volatile computer-readable storage medium, the computer program including program instructions that, when executed by a computer, cause the computer to perform any one of the above-mentioned question-answering method or question-answering model construction method.

In some embodiments, an embodiment of the present invention further provides an electronic device, which includes: the system comprises at least one processor and a memory which is in communication connection with the at least one processor, wherein the memory stores instructions which can be executed by the at least one processor, and the instructions are executed by the at least one processor so as to enable the at least one processor to execute a question-answering method or a question-answering model building method.

Fig. 8 is a schematic diagram of a hardware structure of an electronic device that executes a question-answering method or a question-answering model building method according to another embodiment of the present invention, and as shown in fig. 8, the electronic device includes:

one or more processors 810 and a memory 820, with one processor 810 being an example in FIG. 8.

The apparatus for performing the question answering method or the question answering model building method may further include: an input device 830 and an output device 840.

The processor 810, the memory 820, the input device 830, and the output device 840 may be connected by a bus or other means, such as the bus connection in fig. 8.

The memory 820, which is a non-volatile computer-readable storage medium, may be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as program instructions/modules corresponding to the question-answering method or the question-answering model building method in the embodiments of the present invention. The processor 810 executes various functional applications of the server and data processing by executing nonvolatile software programs, instructions, and modules stored in the memory 820, that is, implementing the question-answering method or the question-answering model building method of the above-described method embodiments.

The memory 820 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function; the storage data area may store data created according to use of the voice interactive apparatus, and the like. Further, the memory 820 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some embodiments, the memory 820 may optionally include memory located remotely from the processor 810, which may be connected to the voice interaction device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The input device 830 may receive input numeric or character information and generate signals related to user settings and function control of the voice interactive apparatus. The output device 840 may include a display device such as a display screen.

The one or more modules are stored in the memory 820 and, when executed by the one or more processors 810, perform the question-answering method or the question-answering model construction method of any of the above-described method embodiments.

The product can execute the method provided by the embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method. For technical details that are not described in detail in this embodiment, reference may be made to the method provided by the embodiment of the present invention.

The electronic device of embodiments of the present invention exists in a variety of forms, including but not limited to:

(1) mobile communication devices, which are characterized by mobile communication capabilities and are primarily targeted at providing voice and data communications. Such terminals include smart phones, multimedia phones, functional phones, and low-end phones, among others.

(2) The ultra-mobile personal computer equipment belongs to the category of personal computers, has calculation and processing functions and generally has the characteristic of mobile internet access. Such terminals include PDA, MID, and UMPC devices, among others.

(3) Portable entertainment devices such devices may display and play multimedia content. The devices comprise audio and video players, handheld game consoles, electronic books, intelligent toys and portable vehicle-mounted navigation devices.

(4) Other onboard electronic devices with data interaction functions, such as a vehicle-mounted device mounted on a vehicle.

The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.

Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a general hardware platform, and certainly can also be implemented by hardware. Based on such understanding, the above technical solutions substantially or contributing to the related art may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments or some parts of the embodiments.

Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (10)

1. A question-answering method comprising:

acquiring user question information;

determining answers corresponding to the user question information based on a question-answer model; the question-answering model is constructed based on a question-answering material table, the question-answering material table comprises at least one header attribute, each header attribute tag is provided with at least one corresponding attribute data, each header attribute is provided with corresponding question-answering marking information, and the question-answering marking information comprises question marking information or answer marking information.

2. The method of claim 1, wherein the determining answers corresponding to the user quiz information based on a question-answer model comprises:

determining user header tag attributes semantically matched with the user question information based on a question-answering model, and determining answers corresponding to the user questions based on each attribute data corresponding to the user header tag attributes; wherein the question and answer material table further comprises semantic expansion words related to the header tag attributes and/or the attribute data.

3. The method of claim 1, wherein after the obtaining user questioning information, the method further comprises:

identifying whether a target question-chasing technique related to the user question information exists or not based on the question-answering model; the question-answering material table also comprises a question-chasing operation aiming at the attribute of the header label with question marking information;

wherein when the target question-line exists, the method further comprises:

and outputting the answer and the target question-chasing technique corresponding to the user question information.

4. The method of any of claims 1-3, wherein the user questioning information includes: a computational question associated with the header tag attribute having question marking information.

5. A question-answering model construction method comprises the following steps:

determining at least one header tag attribute and corresponding attribute data in the initial material table;

acquiring question and answer labeling information aiming at each header label attribute; the question and answer labeling information comprises question labeling information or answer labeling information;

determining a question and answer material table based on the corresponding header tag attributes, the question and answer marking information and the attribute data; and

and constructing a question-answer model based on the question-answer material table.

6. The method of claim 5, wherein the obtaining of the question-and-answer tagging information for each of the head tag attributes comprises:

displaying each header label attribute and a corresponding question-answer labeling control; the question-answer labeling control is used for receiving an operation instruction so as to switch or confirm between question labeling options and answer labeling options;

and determining the question-answer labeling information corresponding to the attribute of each header label based on each question-answer labeling control.

7. The method of claim 5, wherein determining a question and answer material table based on each corresponding header tag attribute, question and answer tagging information, and the attribute data comprises:

acquiring semantic expansion words related to the attribute of the header label and/or the attribute data;

and determining a question and answer material table based on the corresponding header tag attributes, the question and answer labeling information, the attribute data and the semantic expansion words.

8. The method of claim 5, after obtaining the question-and-answer tagging information for each of the headtag attributes, the method further comprising:

acquiring a call-following technique aiming at the attribute of a header label with question marking information;

wherein the determining a question and answer material table based on the corresponding header tag attributes, the question and answer tagging information, and the attribute data comprises:

and determining a question and answer material table based on the corresponding header tag attributes, the question and answer marking information, the attribute data and the question and answer operation.

9. An electronic device, comprising: at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the steps of the method of any one of claims 1-8.

10. A storage medium on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 8.

Images