CN118193696A

CN118193696A - Intelligent question answering method, device, equipment, storage medium and program product

Info

Publication number: CN118193696A
Application number: CN202410308979.2A
Authority: CN
Inventors: 杨涛; 田洁; 鲍丽霞; 郭元昊
Original assignee: Industrial and Commercial Bank of China Ltd ICBC
Current assignee: Industrial and Commercial Bank of China Ltd ICBC
Priority date: 2024-03-18
Filing date: 2024-03-18
Publication date: 2024-06-14

Abstract

The disclosure provides an intelligent question-answering method which can be applied to the technical field of artificial intelligence or the technical field of finance. The intelligent question-answering method comprises the following steps: responding to a question-answer task sent by a user, and analyzing information to be processed carried in the question-answer task to obtain an initial question of the user question; generating first feedback information with prompt property based on the initial problem under the condition that no answer matched with the initial problem exists in an enterprise knowledge base, wherein the enterprise knowledge base is obtained by integrating enterprise knowledge information with multiple dimensions; responding to the operation of the user on the first feedback information, and determining a target question of the user question; retrieving a target answer to the target question from the enterprise knowledge base based on the target question; the target answer is pushed to the user. The present disclosure also provides an intelligent question answering apparatus, device, storage medium and program product.

Description

Intelligent question answering method, device, equipment, storage medium and program product

Technical Field

The present disclosure relates to the field of artificial intelligence technology or financial technology, and in particular, to an intelligent question-answering method, apparatus, device, storage medium, and program product.

Background

With the continuous expansion of enterprise scale and the increasing business scope, users within an enterprise face more and more challenges in daily work. Based on the needs of the business, the users inside the business often need to solve various problems, for example, when processing the newly added business, the users inside the business generally need to query related information, for example, find suitable colleagues to assist in processing work, query data related to the new business, and query data necessary for the new business.

In the process of implementing the inventive concept of the present disclosure, the inventor found that the following problems generally exist in the related art: when an enterprise user inquires information, the enterprise user generally inquires from a plurality of data platforms, and inquires the information from the plurality of data platforms, the defect that the information is scattered exists, and therefore the efficiency of inquiry and reply is low. In addition, although enterprise users can also interact information through the question and answer system, the existing question and answer system is low in intelligent degree, the true intention of the user is difficult to understand accurately, and further accurate answer information is difficult to generate for the user, and the answer accuracy and user experience are reduced.

Disclosure of Invention

In view of the foregoing, the present disclosure provides an intelligent question-answering method, apparatus, device, storage medium, and program product that improve the degree of intellectualization of a question-answering system, query-answer efficiency, accuracy of generating answers, and user experience.

One aspect of the present disclosure provides an intelligent question-answering method, the method including: responding to a question-answer task sent by a user, and analyzing information to be processed carried in the question-answer task to obtain an initial question of the user question; generating first feedback information with prompt property based on the initial problem under the condition that no answer matched with the initial problem exists in an enterprise knowledge base, wherein the enterprise knowledge base is obtained by integrating enterprise knowledge information in multiple dimensions; responding to the operation of the user on the first feedback information, and determining a target question asked by the user; retrieving a target answer to the target question from the enterprise knowledge base based on the target question; pushing the target answer to the user.

According to an embodiment of the present disclosure, the above method further includes: acquiring second feedback information of the user replying to the target; labeling the target answer according to the second feedback information and the attribute information of the target answer to obtain a labeling result; generating an update notification for updating the target reply when the labeling result satisfies a predetermined condition; and pushing the update notification to the target object so that the target object updates the target reply.

According to an embodiment of the disclosure, the labeling of the target reply according to the second feedback information and the attribute information of the target reply to obtain a labeling result includes at least one of the following: labeling the target reply with an expiration identifier under the condition that the second feedback information comprises a preset field, and obtaining the labeling result; monitoring the pushed frequency of the target answer, and labeling the target answer with an expiration identifier under the condition that the pushed frequency is smaller than a first preset threshold value to obtain the labeling result; monitoring the update frequency of a target sub-knowledge base, and labeling an expiration identifier for the target reply to obtain the labeling result under the condition that the update frequency is smaller than a second preset threshold value, wherein the target sub-knowledge base is a sub-knowledge base for storing the target reply; and monitoring version information of the target sub-knowledge base, and labeling an expiration identifier for the target reply under the condition that the version information is lower than the preset version information to obtain the labeling result.

According to an embodiment of the present disclosure, the generating the first feedback information having the prompt property based on the initial problem includes: extracting a target field capable of representing the initial problem from the initial problem; according to the target field, a plurality of associated questions associated with the target field are retrieved from a history question record file; sorting the plurality of associated problems according to the frequency with which the associated problems are queried, so as to obtain a sorting result; and processing a predetermined number of associated questions selected from the sorting result to obtain the first feedback information with the prompting property.

According to an embodiment of the present disclosure, the processing the predetermined number of associated questions selected from the ranking result to obtain the first feedback information with prompt properties includes: invoking a template of the first feedback information; and splicing the initial problems and the preset number of associated problems with the templates of the first feedback information to obtain the first feedback information.

According to an embodiment of the present disclosure, the enterprise knowledge base includes a plurality of sub knowledge bases, and each sub knowledge base is obtained by integrating enterprise knowledge information of one dimension; the retrieving, based on the target question, a target answer to the target question from the enterprise knowledge base, includes: determining a target sub-knowledge base from the plurality of sub-knowledge bases based on field information of the target problem; the target answer is retrieved from the target sub-knowledge base based on the target question.

According to an embodiment of the present disclosure, the retrieving the target answer from the target sub-knowledge base based on the target question includes: searching the enterprise knowledge information set associated with the target problem from the target sub-knowledge base by using a preset searching method; calculating the similarity between the enterprise knowledge information in the enterprise knowledge information set and the target problem; and taking enterprise knowledge information corresponding to the similarity meeting the preset condition as the target answer.

According to an embodiment of the present disclosure, the sub-knowledge base includes at least one of: community data sub-knowledge base, expert consultation sub-knowledge base and platform application programming interface sub-knowledge base; each of the sub-knowledge bases is obtained by integrating knowledge information of enterprises in one dimension, and includes: the community knowledge base is obtained by integrating enterprise knowledge information for maintaining the working positions of enterprises; the expert consultation sub-knowledge base is obtained by integrating an expert information table corresponding to the responsibility of the enterprise working position; and integrating query application programming interfaces of the enterprise data query system to obtain the platform application programming interface sub-knowledge base.

According to an embodiment of the present disclosure, the analyzing the information to be processed carried in the question-answering task to obtain the initial question asked by the user includes: normalizing the information to be processed into initial text information; removing noise information in the initial text information to obtain intermediate text information; word segmentation is carried out on the intermediate text information to obtain target text information; and marking the parts of speech of the target file information to obtain the initial problem.

According to an embodiment of the present disclosure, the information to be processed includes an image to be processed and a text to be processed; the normalizing the information to be processed into the initial text information includes: identifying the object in the image to be processed to obtain an identification result, and taking the identification result as the label information of the image to be processed; converting the tag information into text information of the image to be processed by using a natural language processing technology; and integrating the text information of the image to be processed with the text to be processed to obtain the initial text information.

Another aspect of the present disclosure further provides an intelligent question answering device, where the device includes: the analysis module is used for responding to a question-answer task sent by a user and analyzing information to be processed carried in the question-answer task to obtain an initial question of the user question; the first generation module is used for generating first feedback information with prompt property based on the initial problems under the condition that no answer matched with the initial problems exists in an enterprise knowledge base, wherein the enterprise knowledge base is obtained by integrating enterprise knowledge information in multiple dimensions; the determining module is used for responding to the operation of the user on the first feedback information and determining a target question asked by the user; a search module for searching a target answer of the target question from the enterprise knowledge base based on the target question; and the first pushing module is used for pushing the target answer to the user.

Another aspect of the present disclosure also provides an electronic device, including: one or more processors; and the memory is used for storing one or more computer programs, and the one or more processors execute the one or more computer programs to realize the steps of the intelligent question-answering method.

Another aspect of the present disclosure also provides a computer-readable storage medium having stored thereon a computer program or instructions which, when executed by a processor, implement the steps of the intelligent question-answering method described above.

Another aspect of the present disclosure also provides a computer program product comprising a computer program or instructions which, when executed by a processor, implement the steps of the intelligent question-answering method described above.

According to the intelligent question-answering method, the device, the equipment, the storage medium and the program product provided by the embodiment of the disclosure, the information to be processed in the question-answering task is analyzed by responding to the question-answering task sent by the user, so that the initial question of the user question is obtained; generating first feedback information with promptness based on the initial question under the condition that no answer matched with the initial question exists in the enterprise knowledge base; responding to the operation of the user on the first feedback information, and determining a target question of the user question; based on the target questions, target answers to the target questions are retrieved from the enterprise knowledge base and pushed to the user. In the question-answering process, under the condition that the question-answering system does not find a answer matched with the initial question, first feedback information with prompting performance is generated based on the initial question, and the target question is determined according to the operation of the user on the first feedback information and the target answer is queried, so that the user can be helped to re-ask or provide more information, the question-answering system can better understand the requirement of the user and give an accurate answer, the intelligent degree of the question-answering system is improved, and the problem that the accurate answer is difficult to generate for the asking user due to the fact that the intelligent degree of the existing question-answering system is low in the related art is at least partially overcome. In addition, enterprise knowledge information with multiple dimensions is integrated in the enterprise knowledge base, and the problem that query response efficiency is low due to the fact that information existing in related technologies is scattered is at least partially solved, so that the technical effects of improving query response efficiency, generating response accuracy and user experience are achieved.

Drawings

The foregoing and other objects, features and advantages of the disclosure will be more apparent from the following description of embodiments of the disclosure with reference to the accompanying drawings, in which:

FIG. 1 schematically illustrates an application scenario diagram of intelligent question-answering methods, apparatuses, devices, storage media and program products according to embodiments of the present disclosure;

FIG. 2 schematically illustrates a flow chart of a method of intelligent question-answering according to an embodiment of the present disclosure;

FIG. 3 schematically illustrates a system architecture diagram of an intelligent question-answering system according to an embodiment of the present disclosure;

Fig. 4 schematically illustrates a block diagram of the intelligent question-answering apparatus according to an embodiment of the present disclosure; and

Fig. 5 schematically illustrates a block diagram of an electronic device adapted to implement an intelligent question-answering method according to an embodiment of the present disclosure.

Detailed Description

Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood that the description is only exemplary and is not intended to limit the scope of the present disclosure. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. It may be evident, however, that one or more embodiments may be practiced without these specific details. In addition, in the following description, descriptions of well-known structures and techniques are omitted so as not to unnecessarily obscure the concepts of the present disclosure.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The terms "comprises," "comprising," and/or the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.

All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It should be noted that the terms used herein should be construed to have meanings consistent with the context of the present specification and should not be construed in an idealized or overly formal manner.

Where a convention analogous to "at least one of A, B and C, etc." is used, in general such a convention should be interpreted in accordance with the meaning of one of skill in the art having generally understood the convention (e.g., "a system having at least one of A, B and C" would include, but not be limited to, systems having a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.).

In the technical solution of the present disclosure, the related user information (including, but not limited to, user personal information, user image information, user equipment information, such as location information, etc.) and data (including, but not limited to, data for analysis, stored data, displayed data, etc.) are information and data authorized by the user or sufficiently authorized by each party, and the related data is collected, stored, used, processed, transmitted, provided, disclosed, applied, etc. in compliance with relevant laws and regulations and standards, necessary security measures are taken, no prejudice to the public order colloquia is provided, and corresponding operation entries are provided for the user to select authorization or rejection.

In the scenario of using personal information to make an automated decision, the method, the device and the system provided by the embodiment of the disclosure provide corresponding operation inlets for users, so that the users can choose to agree or reject the automated decision result; if the user selects refusal, the expert decision flow is entered. The expression "automated decision" here refers to an activity of automatically analyzing, assessing the behavioral habits, hobbies or economic, health, credit status of an individual, etc. by means of a computer program, and making a decision. The expression "expert decision" here refers to an activity of making a decision by a person who is specializing in a certain field of work, has specialized experience, knowledge and skills and reaches a certain level of expertise.

With the continuous expansion of enterprise scale and the increasing business scope, users within an enterprise face more and more challenges in daily work. They often need to address a variety of issues, including finding appropriate colleagues to assist in the processing work, finding relevant data, and querying necessary data, etc. However, the existing working mode often has some defects, for example, when a user in an enterprise needs to find a proper colleague to assist in processing work, the user needs to ask for the proper person layer by layer, and thus, the problem cannot be solved in time, and the working efficiency and quality are affected. When searching data, the problems of information dispersion, incomplete, inaccuracy and the like are often encountered, so that the query efficiency is low, and time and energy are wasted. When users in enterprises query data, the users often need to query on a plurality of platforms, and the process is tedious and time-consuming, so that the working efficiency and the quality are directly affected. These problems not only waste time and effort for users within the enterprise, but also directly affect the efficiency and quality of the work. In addition, the existing question-answering system has low intelligent degree, so that the true intention of the user is difficult to be accurately understood, and further, answers required by the user are difficult to be provided for the user.

In view of this, embodiments of the present disclosure provide an intelligent question-answering method, apparatus, device, storage medium, and program product for improving the degree of intellectualization of a question-answering system, query-answer efficiency, accuracy of answer generation, and user experience. Specifically, the method comprises the following steps: responding to a question-answer task sent by a user, and analyzing information to be processed carried in the question-answer task to obtain an initial question of the user question; generating first feedback information with prompt property based on the initial problem under the condition that no answer matched with the initial problem exists in an enterprise knowledge base, wherein the enterprise knowledge base is obtained by integrating enterprise knowledge information with multiple dimensions; responding to the operation of the user on the first feedback information, and determining a target question of the user question; retrieving a target answer to the target question from the enterprise knowledge base based on the target question; the target answer is pushed to the user.

Fig. 1 schematically illustrates an application scenario diagram of an intelligent question-answering method, apparatus, device, storage medium and program product according to an embodiment of the present disclosure.

As shown in fig. 1, an application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, a server 105, and an enterprise knowledge base 106. The network 104 is a medium used to provide communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105, and between the server 105 and the enterprise knowledge base 106. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

The user may interact with the server 105 via the network 104 using the first terminal device 101, the second terminal device 102, the third terminal device 103 to receive or send messages or the like, e.g. to send question-answering tasks or questions, to receive answers to questions or the like. Various communication client applications, such as a smart question and answer type application, a shopping type application, a web browser application, a search type application, an instant messaging tool, a mailbox client, social platform software, etc. (by way of example only) may be installed on the first terminal device 101, the second terminal device 102, the third terminal device 103.

The first terminal device 101, the second terminal device 102, the third terminal device 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smartphones, tablets, laptop and desktop computers, and the like.

The enterprise knowledge base 106 may integrate knowledge and experience related to each working post maintained by the enterprise user in daily work, expert information tables corresponding to detailed responsibilities of each post of the enterprise, and data query APIs (Application Programming Interface ) call of each system in the enterprise, which may be a comprehensive enterprise knowledge base.

The server 105 may be a server providing various services, such as a background management server (merely an example) providing support for question-answer tasks transmitted by the user using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as the question-answer task, and feed back the processing result (such as a reply, a web page, information, or data obtained or generated from the enterprise knowledge base 106 according to the question-answer task) to the terminal device.

It should be noted that, the intelligent question-answering method provided by the embodiments of the present disclosure may be generally executed by the server 105. Accordingly, the intelligent question and answer apparatus provided by the embodiments of the present disclosure may be generally provided in the server 105. The intelligent question-answering method provided by the embodiments of the present disclosure may also be performed by a server or a server cluster that is different from the server 105 and is capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, the enterprise knowledge base 106, and/or the server 105. Accordingly, the intelligent question-answering apparatus provided by the embodiments of the present disclosure may also be provided in a server or a server cluster that is different from the server 105 and is capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, the enterprise knowledge base 106, and/or the server 105.

It should be understood that the number of terminal devices, networks, servers, and enterprise knowledge bases in fig. 1 are merely illustrative. Any number of terminal devices, networks, servers, and enterprise knowledge bases may be provided as desired for implementation.

The intelligent question-answering method of the embodiment of the present disclosure will be described in detail below with reference to the scenario described in fig. 1 through fig. 2 to 3.

Fig. 2 schematically illustrates a flow chart of an intelligent question-answering method according to an embodiment of the present disclosure.

As shown in fig. 2, the intelligent question-answering method of this embodiment includes operations S210 to S250.

In operation S210, in response to the question-answer task sent by the user, the information to be processed carried in the question-answer task is parsed, and an initial question asked by the user is obtained.

In operation S220, in the case where there is no answer matching the initial question in the enterprise knowledge base, first feedback information having a prompt property is generated based on the initial question, wherein the enterprise knowledge base is obtained by integrating enterprise knowledge information of various dimensions.

In operation S230, a target question asked by the user is determined in response to the user' S operation of the first feedback information.

In operation S240, a target answer to the target question is retrieved from the enterprise knowledge base based on the target question.

In operation S250, a target reply is pushed to the user.

Alternatively, a question-answer task may refer to a task initiated by a questioning user that requires a response to be obtained. The information to be processed carried in the question-answering task may be information that the user wants to ask questions. The information to be processed may be through at least one of an input text, an input voice, a transmitted image, and a transmitted video. And analyzing the information to be processed, and processing the information into a text form to obtain an initial question which the user wants to ask.

For example, in the case where the information to be processed has only text input by the user, the processing information may be directly used as an initial question; in the case where the information to be processed includes the inputted text and the transmitted image, for example, the user has uploaded a photograph including a mobile phone and a notebook computer, and asked a question: "what are the prices of these two products, respectively? ". The initial problem obtained by analyzing the information to be processed at this time may be "what are the prices of the mobile phone and the notebook computer, respectively? ".

It will be appreciated that in embodiments of the present disclosure, the question and answer tasks sent by the user are all performed with the user agreeing or authorizing. In the embodiment of the disclosure, a corresponding operation entrance can be provided for the user to choose to agree or reject the automated decision result. That is, before parsing the information to be processed in the question-answer task, an instruction to parse the information by agreeing or rejecting the user input through the corresponding operation entry may be obtained. If the user agrees to parse, the information to be processed is parsed, i.e. step S210 is performed. If the user refuses to analyze the information to be processed, the expert decision flow is entered.

Alternatively, based on the initial question, a reply to the initial question may be retrieved from the enterprise knowledge base, and in the event a matching reply is found, the matching reply may be directly targeted and pushed to the user. In the case where no matching answer is found, first feedback information having a prompt property generated based on the initial question may be fed back to the user at this time.

Illustratively, the user may present a vague problem: "I want to know information about artificial intelligence". In this case, it may be difficult to find a matching answer from the enterprise knowledge base because the question is unclear, at which time, the first feedback information with prompt properties generated by the question answering system according to the initial question may be artificial intelligence asking you to know what aspect? Such as natural language processing, computer vision, etc., for the user to operate. Therefore, the user is guided to ask questions again or provide more information, so that the system can better understand the demands of the user and give accurate answers, and the intelligent degree of the question-answering system is improved.

Alternatively, taking the first feedback information of the above example as an example, the user may further input or confirm information about the natural language processing aspect of the artificial intelligence based on the first feedback information, that is, determine that the target question of the user is "i want to know information about the natural language processing aspect of the artificial intelligence". Knowledge information in natural language processing of artificial intelligence can be retrieved from the enterprise knowledge base as a target answer based on the well-defined target question, and the target answer is pushed to the user.

Optionally, the enterprise knowledge base described above may integrate knowledge and experience related to each working post maintained by the enterprise user in daily work, an expert information table corresponding to detailed responsibilities of each post of the enterprise, and information such as data query API call of each system inside the enterprise, which may be an enterprise knowledge base with comprehensive enterprise knowledge information. And thereby at least partially overcomes the inefficiency of query replies caused by information dispersion.

Alternatively, the enterprise knowledge base may feed back in a collective form when feeding back the target replies. For example, information on natural language processing aspects of artificial intelligence may exist in a plurality of material files, and the enterprise knowledge base may act as a target answer to the question in the form of a file compression package, or a resource link, when feeding back the answer. It will be appreciated that multiple files may be included in the compressed package and multiple sub-links may be included in the resource links. Therefore, through integrating the information, the reply information is more comprehensive, the problem that the information is too scattered is further overcome, the asking user can efficiently determine the reply information, and the efficiency and the accuracy of inquiring and replying are improved.

Fig. 3 schematically illustrates a system architecture diagram of an intelligent question-answering system according to an embodiment of the present disclosure.

As shown in fig. 3, the intelligent question-answering system 300 can be used to implement the intelligent question-answering method provided by embodiments of the present disclosure. The intelligent question and answer system 300 may include a knowledge lookup subsystem 301, a question guide subsystem 302, and a knowledge base refinement subsystem 303.

The knowledge searching subsystem 301 may include an enterprise knowledge base, and the enterprise knowledge base may include a plurality of sub-knowledge bases, where each sub-knowledge base is obtained by integrating enterprise knowledge information in one dimension. It can be appreciated that in the process of inquiring the reply information by the user, the integrated enterprise knowledge information can help the user to inquire the reply information more efficiently, and the enterprise knowledge base of the embodiment of the disclosure can integrate more comprehensive enterprise knowledge information.

Alternatively, the sub-knowledge base may comprise at least one of: community data sub-knowledge base, expert consultation sub-knowledge base, platform application programming interface sub-knowledge base.

Each sub-knowledge base is obtained by integrating enterprise knowledge information of one dimension, and the method can comprise the following steps: the community knowledge base is obtained by integrating enterprise knowledge information for maintaining the working positions of enterprises; the expert consultation sub-knowledge base is obtained by integrating expert information tables corresponding to the responsibilities of the enterprise; and integrating query application programming interfaces of the enterprise data query system to obtain a platform application programming interface sub-knowledge base.

Specifically, the community repository sub-knowledge base may be that knowledge and experience related to each work post maintained by the enterprise user in daily work is integrated into the community repository. Such information may be in the form of text, pictures, video, etc., including common problem solutions, instruction manuals, best practices, etc. Problems and solutions in various operating scenarios are contemplated. Users can quickly query and reference these materials through the system to solve the problems encountered in the work. At least partially, the problems of low query efficiency, time waste and energy waste caused by the problems of scattered, incomplete, inaccurate information and the like frequently encountered in the related technology when searching data are overcome, and the query reply efficiency is further improved. It will be appreciated that these materials may be presented by professionals and uploaded to a community materials sub-knowledge base.

The expert consulting sub-knowledge base can be an expert information table corresponding to the detailed responsibilities of each post of the enterprise. When a user needs to seek help in a specific field, the intelligent question-answering system can quickly find out the expert in the related field and provide a communication channel so that the user can timely consult the expert and solve the problem. At least partially, the problems that the problems can not be solved in time and the working efficiency and the working quality are affected because the problems are difficult to find suitable persons quickly due to the fact that the users in the enterprises need to find suitable colleagues to assist in processing work and the users need to inquire and find the persons layer by layer in the related art are solved, and the technical effect of improving the inquiring efficiency is achieved.

The platform API sub-knowledge base can be a data query API call integrating all systems in the enterprise, and related data can be directly queried in the systems. At least partially, the problems that the process is tedious and time-consuming, and the working efficiency and the working quality are directly influenced when the related technology is used for inquiring a plurality of platforms when users in enterprises inquire data are at least partially overcome, and the technical effect of improving the inquiring efficiency is further achieved.

According to the embodiment of the disclosure, through integrating the community data sub-knowledge base, the expert consultation sub-knowledge base and the platform API sub-knowledge base, comprehensive knowledge support is provided for users, the users are helped to be asked to quickly solve problems encountered in work, and the working efficiency and quality are improved.

The problem guiding subsystem 302 may analyze the to-be-processed information in the question-answering task issued by the user by using the technology of the artificial intelligence algorithm, and then match the answers of the problems through the enterprise knowledge information in the enterprise knowledge base in the knowledge searching subsystem 301.

Specifically, the above-described operation S210 may include the following operations: the information to be processed is standardized into initial text information; removing noise information in the initial text information to obtain intermediate text information; word segmentation is carried out on the intermediate text information, and target text information is obtained; and marking the part of speech of the target file information to obtain an initial problem.

Optionally, the information to be processed includes an image to be processed and a text to be processed; the process of normalizing the information to be processed to the initial text information may include the operations of: identifying an object in the image to be processed to obtain an identification result, and taking the identification result as label information of the image to be processed; converting the tag information into text information of the image to be processed by using a natural language processing technology; and integrating the text information of the image to be processed with the text to be processed to obtain initial text information.

Alternatively, the feature information in the image to be processed may be extracted by an image processing technique in the related art to obtain a recognition result, such as an object, a scene, a color, or the like. Identifying an object or object in the image to be processed may generate a corresponding tag or description, optionally with the identification result as the tag or description. And the recognition result obtained by recognizing the image to be processed is converted into a text form by utilizing a natural language processing technology, so that the recognition result is convenient to match, analyze and integrate with the user question, and the initial question of the user question is obtained.

Illustratively, the user uploads a picture containing a cell phone and a notebook computer and asks questions: "what are the prices of these two products, respectively? The process of resolving the information to be processed to obtain the initial problem can be realized through the following steps.

And (5) image processing. The problem-guiding subsystem 302 extracts features in the picture by using an image processing technology, identifies the mobile phone and the notebook computer, and generates corresponding tag information, such as the mobile phone and the notebook computer.

Natural language processing. The question guide subsystem 302 converts the tag information obtained by recognizing the image to be processed into a text description, such as "mobile phone", "notebook", while understanding the intention of the user to ask a question.

Information integration and problem solving. The problem guiding subsystem 302 comprehensively considers the image recognition result and the text to be processed input by the user, and matches the information of the mobile phone and the notebook computer, and the obtained initial text information can be "what are the prices of the mobile phone and the notebook computer respectively? ". The issue guidance subsystem 302 may present price information for both products based on the issue.

By combining image processing techniques and natural language processing techniques, the problem-guiding subsystem 302 may extract useful information from pictures and problems provided by the user, implement a more intelligent problem-solving process, and improve user experience and accuracy of problem solutions.

In one embodiment, the initial text information may be further processed using natural language techniques. This process may include word segmentation, part-of-speech tagging, syntactic analysis, and the like.

Specifically, the process of removing noise information in the initial text information to obtain intermediate text information may be to remove stop words, punctuation marks and special characters in the initial text information, and perform word drying or word shape reduction.

The process of word segmentation of the intermediate text information to obtain the target text information may be using rule-based or statistical methods, such as a maximum matching method, a machine learning model, e.g., a hidden Markov model (Hidden Markov Model, HMM) model, etc.

Part-of-speech tagging of the target file information to obtain the initial question may be by using statistical-based algorithms such as HMM, probabilistic graphical models (Conditional Random Field, CRF), etc.

Optionally, syntactic analysis may be performed after part-of-speech tagging. Syntactic analysis may use rule-based Context-free grammar (Context-FREE GRAMMAR, CFG), statistical-based methods such as probabilistic Context-free grammar (Probabilistic Context FREE GRAMMAR, PCFG), and deep learning methods such as neural network models.

Illustratively, for example, the user proposes: "how does improve software performance? ". The noise information rejection process will remove the stop word "how" and punctuation "? The processed text is obtained to improve the software performance; the word segmentation process may be divided into "boost", "software", "performance", and the part-of-speech tagging process may tag the part of speech of each word, such as "boost (verb) software (noun) performance (noun)". For the text after word segmentation and part-of-speech tagging, the syntactic analysis algorithm analyzes the structure of the sentence, for example, "improvement" is a verb, and "software" and "performance" are nouns, so that the system can accurately understand the intention of the user and the proposed problem.

Alternatively, based on the enterprise knowledge base described above, text similarity calculation, semantic matching, information retrieval, and other techniques may be used to retrieve answers from the enterprise knowledge base based on the initial questions or the target questions.

In particular, the process of retrieving the target reply may include the following operations: determining a target sub-knowledge base from a plurality of sub-knowledge bases based on field information of the target problem; the target answer is retrieved from the target sub-knowledge base based on the target question.

Optionally, determining the target sub-knowledge base from the multiple sub-knowledge bases based on field information of the target problem, for example, in a case that fields such as expert information exist in the target problem, the expert consulting sub-knowledge base may be used as the target sub-knowledge base; under the condition that fields such as an API exist in the target problem, the platform API sub-knowledge base can be used as a target sub-knowledge base; in the case that there are non-expert information fields such as an operation manual field, a knowledge information field, and the like, and non-API fields in the target question, the community knowledge sub-knowledge base may be used as the target sub-knowledge base.

Alternatively, the process of retrieving the target answer from the target sub-knowledge base based on the target question may include the operations of: retrieving the enterprise knowledge information set associated with the target problem from the target sub-knowledge base using a predetermined retrieval method; calculating the similarity between the enterprise knowledge information in the enterprise knowledge information set and the target problem; and taking enterprise knowledge information corresponding to the similarity meeting the preset condition as a target answer.

Alternatively, information retrieval finds documents or information most relevant to the user query by retrieving text data in the system. The predetermined retrieval method is, for example, an inverted index, TF-IDF (Term Frequency-inverse document Frequency) algorithm, BM25 (Best Match 25, core algorithm for information retrieval), or the like. For example, the user enters the question "which popular movies are recently? The intelligent question-answering system can search the movie database through a preset search method, find the latest popular movie list or popular movie set and return to the user. For another example, the user enters the question "what are product introductions recently about a class a financial product? The intelligent question-answering system can search the community data sub-knowledge base through a preset search method, find the product introduction of the A-type financial products to obtain an enterprise knowledge information set, and return the enterprise knowledge information set to the user in the form of a link or a compressed package.

Alternatively, the enterprise knowledge information may be determined in the enterprise database in a similarity manner. The answer information for solving the questioning problem of the questioning user can be determined in a similarity mode in the enterprise knowledge set. And taking the enterprise knowledge information corresponding to the similarity meeting the preset condition as the target answer, for example, taking the enterprise knowledge information with the highest similarity as the target answer, or taking the enterprise knowledge information with the similarity being larger than a preset similarity threshold value as the target answer, and particularly, carrying out adaptability adjustment according to actual needs.

Alternatively, the text similarity calculation may be to determine the degree of association between two pieces of text by comparing their similarity. Common text similarity calculation methods include cosine similarity, jaccard similarity, edit distance, and the like. For example, the user input question "how to learn Python programming? The system can find the text related to the text through text similarity calculation, such as 'Python programming entry guide' or 'Python learning course', and the like.

Alternatively, the semantic matching may be to determine the degree of semantic association between two pieces of text by understanding the semantic information of the text. Common semantic matching methods may include similarity computation based on word vectors, semantic role labeling, semantic relationship extraction, etc. For example, the user enters the question "how to treat cold? The system can understand the user requirements through semantic matching technology and find relevant medical knowledge or treatment methods.

Optionally, on the basis of the above retrieval process, a deep learning technology may be added, and text representation learning may be performed using a pre-trained deep learning model such as BERT (Bidirectional Encoder Representations from Transformers, bi-directional encoder representation from Transformer), GPT (GENERATIVE PRE-Trained Transformer, a generated pre-trained transducer model), a natural language processing model, which may capture more complex semantic information between texts, and improve accuracy of text similarity calculation and semantic matching. Meanwhile, a pre-trained model can be finely tuned to a specific task by utilizing a transfer learning technology, and the text matching requirement of a specific field or context can be better adapted. And by combining the techniques of a attention mechanism, a bi-directional encoder and the like, the understanding and expressing capacity of the relevance between the models and the texts can be further improved, so that the system can find an answer which is matched with the user problem best.

According to embodiments of the present disclosure, an automated dialog model is built by the system using the histories of the enterprise internal chat system and the community knowledge base via an artificial intelligence algorithm. This enables the intelligent question-answering system to more intelligently understand questions posed by the user and to give more accurate and personalized answers.

Optionally, incremental learning may be used for the models involved in the above process to help the problem guiding subsystem 302 continuously update model parameters, adapt to new data and problems, and thereby improve the understanding accuracy of the intelligent question-answering system for user questions.

In particular, the problem guidance subsystem 302 may periodically collect new data from pictures and questions uploaded by the user, including new product pictures and related problems. Problem guidance subsystem 302 may combine the new data with the existing data to retrain the model. By training on new data, the model can continuously improve the understanding ability of images and text information. During the incremental learning process, problem guidance subsystem 302 may dynamically adjust model parameters to accommodate new data and problems. By adjusting the model parameters based on the characteristics of the new data, the accurate understanding of the user's question by the question guidance subsystem 302 may be improved. After model update, the problem guidance subsystem 302 needs to evaluate the new model to check its performance on the new data. According to the evaluation result, the system can obtain feedback information to guide the incremental learning process of the next round. By continuing to perform incremental learning, the problem guidance subsystem 302 can continually optimize the model, increasing the accuracy of asking the user. As the amount of data and the variety of questions increases, the comprehensiveness of the question guidance subsystem 302 will gradually increase.

According to the embodiment of the disclosure, the historical data of the user question and the system answer are learned and optimized through the machine learning algorithm, and the question guiding subsystem can improve the accuracy and efficiency of the user question, help the user to quickly solve the question and improve the working efficiency.

Optionally, the process of searching in the enterprise knowledge base based on the initial question and the target question is described above, and in the case that the initial question is searched in the enterprise knowledge base, and no answer matching the initial question exists in the enterprise knowledge base, at this time, first feedback information with prompt property may be generated based on the initial question, and the target question asked by the user is determined according to the operation of the user on the first feedback information, so that further searching in the enterprise knowledge base is performed according to the target question. The process of obtaining the first feedback information will be described below.

In particular, the process of generating the first feedback information having the prompt property based on the initial question may include the operations of: extracting a target field capable of representing the initial question from the initial question; according to the target field, a plurality of associated questions associated with the target field are retrieved from the history question record file; according to the queried frequency of the associated problems, sequencing a plurality of associated problems to obtain a sequencing result; and processing a predetermined number of associated problems selected from the sorting result to obtain first feedback information with prompt property.

Optionally, the process of processing the predetermined number of associated questions selected from the ranking result to obtain the first feedback information with the prompting property may include the following operations: invoking a template of the first feedback information; and splicing the initial problems and the preset number of associated problems with templates of the first feedback information to obtain the first feedback information.

Alternatively, the target field that can characterize the initial question is extracted from the initial question, for example, the target field extracted from the initial question, "i want to know information about artificial intelligence" may be "artificial intelligence".

Optionally, the history problem record file may record information such as the answer pair of the problem, the frequency of the problem being queried, the frequency of the answer being pushed, and the like. According to the target field of the artificial intelligence, a plurality of queried problems related to the artificial intelligence can be retrieved from the history problem record file to obtain related problems, such as the problems of understanding natural language processing of the artificial intelligence, understanding computer vision information of the artificial intelligence and the like.

Each associated question may be labeled with a frequency of being queried, and the associated questions may be ranked according to the order of the queried frequency from high to low, resulting in a ranking result. And obtaining first feedback information with prompt property according to the preset number of associated problems in the sequencing result. For example, the first 2 associated questions are selected from the sorting result, namely, the natural language processing of the artificial intelligence is known, and the computer vision information of the artificial intelligence is known. The first feedback information derived from these associated questions may be "please ask you about artificial intelligence of what aspect? Such as natural language processing, computer vision? ".

Alternatively, the first feedback information may have a template, for example, "please ask you for xxx (target field extracted from initial questions) of which aspect, such as xxx (field extracted from the most frequently queried associated questions that can represent the associated questions) or xxx (field extracted from the most frequently queried associated questions that can represent the associated questions)? ". It will be appreciated that the template of the first feedback information may be adapted according to the actual implementation.

Alternatively, if the question guidance subsystem 302 fails to accurately match the question posed by the user, i.e., there is no answer in the enterprise knowledge base that matches the initial question, the question guidance subsystem 302 may generate hinted first feedback information using artificial intelligence algorithms, such as by entering the question posed by the user using a text generation model pre-trained in the related art, and then generating a hinted suggestion that directs the user to ask a question again or provide more information. For example, the user may present a fuzzy question: "I want to know information about artificial intelligence," problem guidance subsystem 302 can generate suggestions: "please ask you what artificial intelligence to know what aspect? Such as natural language processing, computer vision, etc., to guide the user to re-ask questions or provide more information so that the system can better understand the user's needs and give accurate answers, to improve the efficiency of query replies, the accuracy of generating replies and the degree of intellectualization of the intelligent question-answering system.

The knowledge base perfecting subsystem 303 can record feedback records of questions in real time in the actual use process, and establish a use record of each matching answer in the knowledge base. When the second feedback information represents the inaccurate condition of the answer (i.e. the enterprise knowledge base or the enterprise knowledge information in the sub-knowledge base corresponding to the target answer), the knowledge base perfecting subsystem 303 will automatically label and send feedback to the maintainer in the intelligent question-answering system 300, and prompt the maintainer of the related data to update the corresponding knowledge base in time so as to ensure the timeliness and accuracy of the enterprise knowledge base.

Specifically, on the basis of the above operation, the intelligent question-answering method according to the embodiment of the present disclosure may further include the following operations: acquiring second feedback information of the user replying to the target; labeling the target reply according to the second feedback information and the attribute information of the target reply to obtain a labeling result; generating an update notification for updating the target reply if the labeling result meets a predetermined condition; pushing an update notification to the target object to cause the target object to update the target reply.

Alternatively, the second feedback information may be an evaluation by the user of whether the target reply is accurate. Such as information that the target reply is inaccurate, that the target reply is invalid, or that the target reply has been invalid. The knowledge base perfecting subsystem 303 can record the questions posed by the user and the answer results of the intelligent question-answering system to the questions in real time. These records will be used to analyze and improve the accuracy and efficiency of the system.

Optionally, the knowledge base perfecting subsystem 303 may also establish a usage record of each question matching result in the knowledge base, including information such as specific content of the question, matching answer, and second feedback of the user. These records help the knowledge base perfecting subsystem 303 analyze and optimize the use of questions and answers.

Optionally, labeling the target reply according to the second feedback information and the attribute information of the target reply, and obtaining a labeling result may include at least one of the following ways: labeling the target reply with an expiration identifier under the condition that the second feedback information comprises a preset field, and obtaining a labeling result; monitoring the pushed frequency of the target answer, and marking the target answer with an expiration identifier under the condition that the pushed frequency is smaller than a first preset threshold value to obtain a marking result; monitoring the updating frequency of a target sub-knowledge base, and marking an expiration identifier for the target reply to obtain a marking result under the condition that the updating frequency is smaller than a second preset threshold value, wherein the target sub-knowledge base is a sub-knowledge base for storing the target reply; and monitoring version information of the target sub-knowledge base, and labeling an expiration identifier for the target reply under the condition that the version information is lower than the preset version information to obtain a labeling result.

Optionally, the predetermined field is, for example, a field that is invalid, expired, invalid, inaccurate, etc.

The knowledge base perfecting subsystem 303 can judge whether the target answer (i.e. the enterprise knowledge base or the enterprise knowledge information in the sub-knowledge base corresponding to the target answer) is outdated or inaccurate according to the feedback condition of the user. If the target reply is frequently fed back by the user as invalid or inaccurate (e.g., the feedback frequency reaches a predetermined feedback frequency threshold), the knowledge base improvement subsystem 303 may mark the target reply as being out of date, in need of update, in-time in need of update, etc.

The knowledge base perfecting subsystem 303 can also monitor the usage of each target answer (i.e., the enterprise knowledge base or the enterprise knowledge information in the sub-knowledge base corresponding to the target answer) in the enterprise knowledge base, including the number of times checked, the frequency searched, the frequency pushed, etc. If the target reply is not used for a long period of time (or pushed less frequently than a first predetermined threshold) the knowledge base improvement subsystem 303 may mark the target reply as being out of date, in need of update, in error of information, etc., and may need to be updated or eliminated.

The knowledge base perfecting subsystem 303 may also set an update frequency standard of the enterprise knowledge base or the sub-knowledge base, for example, at least once every month, so as to ensure timeliness of the enterprise knowledge base or the sub-knowledge base. If the enterprise repository or sub-repository is not updated for a long period of time (or the update frequency is less than a second predetermined threshold), the repository perfecting subsystem 303 may mark the enterprise repository or sub-repository as being out of date, needing to be updated, information errors, etc.

The knowledge base perfecting subsystem 303 may also set an update frequency standard of the enterprise knowledge information in the enterprise knowledge base or the sub-knowledge base, for example, at least once per month, so as to ensure timeliness of the enterprise knowledge information in the enterprise knowledge base or the sub-knowledge base. If the enterprise knowledge information in the enterprise knowledge base or the sub-knowledge base is not updated for a long time within the third predetermined period of time, the knowledge base perfecting subsystem 303 may mark the enterprise knowledge information in the enterprise knowledge base or the sub-knowledge base as an identifier such as expiration, update, information error, etc.

The knowledge base perfecting subsystem 303 can also realize version control of the enterprise knowledge base or the sub knowledge base, and record the content and time of each update. If a version of the enterprise knowledge base or sub-knowledge base is not updated or used for a long period of time (or the version information of the enterprise knowledge base or sub-knowledge base is below the predetermined version information), the knowledge base refinement subsystem 303 may prompt maintenance personnel to check and update.

It is to be understood that the above-described predetermined feedback frequency threshold, the first predetermined period of time, the first predetermined threshold, the second predetermined period of time, the third predetermined period of time, the fourth predetermined period of time, and the predetermined version information may be adaptively adjusted according to actual situations.

By the above manner, the knowledge base perfecting subsystem 303 can more comprehensively judge whether the enterprise knowledge base or the enterprise knowledge information in the sub knowledge base is outdated or inaccurate, so that the maintenance efficiency and accuracy of the enterprise knowledge base are improved, and the enterprise knowledge base is ensured to always keep the latest and accurate state. In addition, by automatically marking corresponding marks and sending feedback notification to maintenance personnel in the intelligent question-answering system, the maintenance personnel can quickly know the nature of the problem and process the problem correspondingly.

Alternatively, in the case where the labeling result satisfies the predetermined condition, for example, when the labeling result characterizes that the target answer (i.e., the enterprise knowledge information in the enterprise knowledge base or the sub-knowledge base corresponding to the target answer) is expired, an update notification for updating the target answer, for example, xx enterprise knowledge information needs to be updated, xx enterprise knowledge information errors, or the like, may be generated.

Alternatively, the target object may include a maintainer of the intelligent question-answering system or a maintainer of the enterprise knowledge information. Specifically, the knowledge base perfecting subsystem 303 may push the update notification to a maintainer of the intelligent question-answering system or a maintainer of the enterprise knowledge information in a notification or reminding manner, so as to prompt the maintainer of the intelligent question-answering system or the maintainer of the enterprise knowledge information to update the enterprise knowledge base or the sub knowledge base in time, and ensure timeliness and accuracy of the knowledge base. This may be accomplished through a workflow or task management system within the system to ensure that feedback of the problem can be handled and updated in time. Through the continuous maintenance and updating of the knowledge base, the intelligent question-answering system can provide accurate and real-time information required by users, help them to solve problems more quickly and improve working efficiency.

According to the embodiment of the disclosure, through the functions of real-time recording of problem feedback, establishment of use records, automatic labeling feedback, supervision and updating of a knowledge base and the like, the accuracy and efficiency of an enterprise dialogue problem management system can be effectively improved, an enterprise is helped to establish a perfect knowledge base, accurate and real-time information required by a user is provided, and further the working efficiency and the service quality are improved.

According to the intelligent question-answering method and the intelligent question-answering system, the community data sub-knowledge base, the expert consultation sub-knowledge base and the platform API sub-knowledge base are integrated, comprehensive knowledge support is provided for users, the users are helped to quickly solve problems encountered in work, and working efficiency and quality are improved. By utilizing the automatic dialogue model established by the artificial intelligence algorithm, the intelligent question-answering system can more intelligently understand the questions posed by the user, give more accurate and personalized answers, and improve the working experience and support of the user. The machine learning algorithm is used for learning and optimizing historical data of user questions and system answers, and the question guiding subsystem can improve accuracy and efficiency of user questions, help users to quickly solve the questions and improve working efficiency. The knowledge base perfecting subsystem can record feedback records of problems in real time, and establishes a use record of each matching result in the enterprise knowledge base, so as to help maintainers update the corresponding knowledge base in time and ensure timeliness and accuracy of the enterprise knowledge base.

It should be noted that, unless there is an execution sequence between different operations or an execution sequence between different operations in technical implementation, the execution sequence between multiple operations may be different, and multiple operations may also be executed simultaneously in the embodiment of the disclosure.

Based on the intelligent question-answering method, the disclosure also provides an intelligent question-answering device. The device will be described in detail below in connection with fig. 4.

Fig. 4 schematically shows a block diagram of the structure of the intelligent question-answering apparatus according to the embodiment of the present disclosure.

As shown in fig. 4, the intelligent question-answering apparatus 400 of this embodiment includes a parsing module 410, a first generating module 420, a determining module 430, a retrieving module 440, and a first pushing module 450.

The parsing module 410 is configured to parse the information to be processed carried in the question-answering task in response to the question-answering task sent by the user, so as to obtain an initial question asked by the user.

The first generation module 420 is configured to generate, based on the initial question, first feedback information having a prompt property when no answer matching the initial question exists in an enterprise knowledge base, where the enterprise knowledge base is obtained by integrating enterprise knowledge information in multiple dimensions.

And a determining module 430, configured to determine a target question asked by the user in response to the user's operation on the first feedback information.

A retrieval module 440 for retrieving a target answer to the target question from the enterprise knowledge base based on the target question;

A first pushing module 450, configured to push the target reply to the user.

According to the embodiment of the disclosure, the intelligent question answering device can further comprise an acquisition module, a labeling module, a second generation module and a second pushing module.

And the acquisition module is used for acquiring second feedback information of the target reply of the user.

And the labeling module is used for labeling the target answer according to the second feedback information and the attribute information of the target answer to obtain a labeling result.

And the second generation module is used for generating an update notification for updating the target answer under the condition that the labeling result meets the preset condition.

And the second pushing module is used for pushing the update notification to the target object so as to enable the target object to update the target reply.

According to an embodiment of the disclosure, the labeling module may include a first labeling sub-module, a second labeling sub-module, a third labeling sub-module, and a fourth labeling sub-module.

And the first labeling sub-module is used for labeling the target reply with the expiration identifier under the condition that the second feedback information comprises the predetermined field, so as to obtain a labeling result.

And the second labeling sub-module is used for monitoring the pushed frequency of the target reply and labeling the target reply with an expiration identifier under the condition that the pushed frequency is smaller than a first preset threshold value to obtain a labeling result.

And the third labeling sub-module is used for monitoring the updating frequency of the target sub-knowledge base and labeling the expiration identifier for the target reply to obtain a labeling result under the condition that the updating frequency is smaller than a second preset threshold value, wherein the target sub-knowledge base is a sub-knowledge base for storing the target reply.

And the fourth labeling sub-module is used for monitoring version information of the target sub-knowledge base and labeling an expiration identifier for the target reply under the condition that the version information is lower than the preset version information to obtain a labeling result.

According to an embodiment of the present disclosure, the first generation module may include an extraction sub-module, a first retrieval sub-module, a sorting sub-module, and a processing sub-module.

An extraction sub-module for extracting, from the initial question, a target field capable of characterizing the initial question.

And the first retrieval sub-module is used for retrieving a plurality of associated questions associated with the target field from the historical question record file according to the target field.

And the sequencing sub-module is used for sequencing the plurality of the associated problems according to the queried frequency of the associated problems to obtain a sequencing result.

And the processing sub-module is used for processing a predetermined number of associated problems selected from the sorting result to obtain first feedback information with prompting property.

According to an embodiment of the present disclosure, the processing sub-module may include a calling unit and a stitching unit.

And the calling unit is used for calling the template of the first feedback information.

And the splicing unit is used for splicing the initial problems, the preset number of associated problems and the templates of the first feedback information to obtain the first feedback information.

According to an embodiment of the present disclosure, the retrieval module may comprise a first determination sub-module and a second retrieval sub-module.

And the first determining sub-module is used for determining the target sub-knowledge base from the plurality of sub-knowledge bases based on the field information of the target problem.

And the second retrieval sub-module is used for retrieving target answers from the target sub-knowledge base based on the target questions.

According to an embodiment of the present disclosure, the second retrieval sub-module may comprise a retrieval unit, a calculation unit and a result unit.

And the searching unit is used for searching the enterprise knowledge information set associated with the target problem from the target sub-knowledge base by using a preset searching method.

And the calculating unit is used for calculating the similarity between the enterprise knowledge information in the enterprise knowledge information set and the target problem.

And a result unit configured to reply, as a target, the enterprise knowledge information corresponding to the similarity satisfying the predetermined condition.

According to an embodiment of the disclosure, the intelligent question answering device may further include a first integration module, a second integration module, and a third integration module.

The first integration module is used for integrating enterprise knowledge information for maintaining the working positions of the enterprises to obtain a community knowledge base.

And the second integration module is used for integrating the expert information table corresponding to the work position responsibilities of the enterprise to obtain an expert consultation sub-knowledge base.

And the third integration module is used for integrating the query application programming interfaces of the enterprise data query system to obtain a platform application programming interface sub-knowledge base.

According to an embodiment of the present disclosure, the parsing module may include a normalization sub-module, a culling sub-module, a word segmentation sub-module, and a fifth labeling sub-module.

And the normalization sub-module is used for normalizing the information to be processed into initial text information.

And the rejecting sub-module is used for rejecting noise information in the initial text information to obtain intermediate text information.

And the word segmentation sub-module is used for segmenting the intermediate text information to obtain target text information.

And the fifth labeling sub-module is used for performing part-of-speech labeling on the target file information to obtain an initial problem.

According to an embodiment of the present disclosure, the standardized sub-module may include an identification unit, a conversion unit, and an integration unit.

The identification unit is used for identifying the object in the image to be processed to obtain an identification result, and taking the identification result as the label information of the image to be processed.

And the conversion unit is used for converting the label information into text information of the image to be processed by using a natural language processing technology.

And the integration unit is used for integrating the text information of the image to be processed with the text to be processed to obtain initial text information.

Any of the parsing module 410, the first generating module 420, the determining module 430, the retrieving module 440, and the first pushing module 450 may be combined in one module to be implemented, or any of the modules may be split into a plurality of modules, according to an embodiment of the present disclosure. Or at least some of the functionality of one or more of the modules may be combined with, and implemented in, at least some of the functionality of other modules. According to embodiments of the present disclosure, at least one of parsing module 410, first generating module 420, determining module 430, retrieving module 440, and first pushing module 450 may be implemented at least in part as hardware circuitry, such as a Field Programmable Gate Array (FPGA), programmable Logic Array (PLA), a system-on-chip, a system-on-substrate, a system-on-package, an Application Specific Integrated Circuit (ASIC), or may be implemented in hardware or firmware in any other reasonable manner of integrating or packaging the circuitry, or in any one of or a suitable combination of three of software, hardware, and firmware. Or at least one of the parsing module 410, the first generating module 420, the determining module 430, the retrieving module 440 and the first pushing module 450 may be at least partially implemented as a computer program module which, when run, may perform the corresponding functions.

It should be noted that, in the embodiment of the present disclosure, the intelligent question answering device portion corresponds to the intelligent question answering method portion in the embodiment of the present disclosure, and the description of the intelligent question answering device portion specifically refers to the intelligent question answering method portion, which is not described herein.

As shown in fig. 5, an electronic device 500 according to an embodiment of the present disclosure includes a processor 501 that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 502 or a program loaded from a storage section 508 into a Random Access Memory (RAM) 503. The processor 501 may include, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or an associated chipset and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), or the like. The processor 501 may also include on-board memory for caching purposes. The processor 501 may comprise a single processing unit or a plurality of processing units for performing different actions of the method flows according to embodiments of the disclosure.

In the RAM 503, various programs and data required for the operation of the electronic apparatus 500 are stored. The processor 501, ROM 502, and RAM 503 are connected to each other by a bus 504. The processor 501 performs various operations of the method flow according to the embodiments of the present disclosure by executing programs in the ROM 502 and/or the RAM 503. Note that the program may be stored in one or more memories other than the ROM 502 and the RAM 503. The processor 501 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in the one or more memories.

According to an embodiment of the present disclosure, the electronic device 500 may also include an input/output (I/O) interface 505, the input/output (I/O) interface 505 also being connected to the bus 504. The electronic device 500 may also include one or more of the following components connected to the I/O interface 505: an input section 506 including a keyboard, a mouse, and the like; an output portion 507 including a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker, and the like; a storage portion 508 including a hard disk and the like; and a communication section 509 including a network interface card such as a LAN card, a modem, or the like. The communication section 509 performs communication processing via a network such as the internet. The drive 510 is also connected to the I/O interface 505 as needed. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 510 as needed so that a computer program read therefrom is mounted into the storage section 508 as needed.

The present disclosure also provides a computer-readable storage medium that may be embodied in the apparatus/device/system described in the above embodiments; or may exist alone without being assembled into the apparatus/device/system. The computer-readable storage medium carries one or more programs which, when executed, implement methods in accordance with embodiments of the present disclosure.

According to embodiments of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, which may include, for example, but is not limited to: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this disclosure, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. For example, according to embodiments of the present disclosure, the computer-readable storage medium may include ROM 502 and/or RAM 503 and/or one or more memories other than ROM 502 and RAM 503 described above.

Embodiments of the present disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowcharts. The program code means for causing a computer system to carry out the intelligent question-answering method provided by the embodiments of the present disclosure when the computer program product is run on the computer system.

The above-described functions defined in the system/apparatus of the embodiments of the present disclosure are performed when the computer program is executed by the processor 501. The systems, apparatus, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the disclosure.

In one embodiment, the computer program may be based on a tangible storage medium such as an optical storage device, a magnetic storage device, or the like. In another embodiment, the computer program may also be transmitted, distributed, and downloaded and installed in the form of a signal on a network medium, and/or installed from a removable medium 511 via the communication portion 509. The computer program may include program code that may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the foregoing.

In such an embodiment, the computer program may be downloaded and installed from a network via the communication portion 509, and/or installed from the removable media 511. The above-described functions defined in the system of the embodiments of the present disclosure are performed when the computer program is executed by the processor 501. The systems, devices, apparatus, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the disclosure.

According to embodiments of the present disclosure, program code for performing computer programs provided by embodiments of the present disclosure may be written in any combination of one or more programming languages, and in particular, such computer programs may be implemented in high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. Programming languages include, but are not limited to, such as Java, c++, python, "C" or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., connected via the Internet using an Internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

Those skilled in the art will appreciate that the features recited in the various embodiments of the disclosure and/or in the claims may be provided in a variety of combinations and/or combinations, even if such combinations or combinations are not explicitly recited in the disclosure. In particular, the features recited in the various embodiments of the present disclosure and/or the claims may be variously combined and/or combined without departing from the spirit and teachings of the present disclosure. All such combinations and/or combinations fall within the scope of the present disclosure.

The embodiments of the present disclosure are described above. These examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the embodiments cannot be used advantageously in combination. The scope of the disclosure is defined by the appended claims and equivalents thereof. Various alternatives and modifications can be made by those skilled in the art without departing from the scope of the disclosure, and such alternatives and modifications are intended to fall within the scope of the disclosure.

Claims

1. An intelligent question-answering method, characterized in that the method comprises the following steps:

Responding to a question-answer task sent by a user, and analyzing information to be processed carried in the question-answer task to obtain an initial question of the user question;

Generating first feedback information with prompt property based on the initial problem under the condition that no answer matched with the initial problem exists in an enterprise knowledge base, wherein the enterprise knowledge base is obtained by integrating enterprise knowledge information in multiple dimensions;

responding to the operation of the user on the first feedback information, and determining a target question asked by the user;

Retrieving a target answer to the target question from the enterprise knowledge base based on the target question;

pushing the target answer to the user.

2. The method according to claim 1, wherein the method further comprises:

acquiring second feedback information of the user replying to the target;

labeling the target reply according to the second feedback information and the attribute information of the target reply to obtain a labeling result;

generating an update notification for updating the target reply if the labeling result meets a predetermined condition;

pushing the update notification to the target object so that the target object updates the target reply.

3. The method according to claim 2, wherein the labeling the target reply according to the second feedback information and the attribute information of the target reply, and the labeling result is obtained by at least one of the following ways:

Labeling the target reply with an expiration identifier under the condition that the second feedback information comprises a preset field, and obtaining the labeling result;

Monitoring the pushed frequency of the target answer, and labeling the target answer with an expiration identifier under the condition that the pushed frequency is smaller than a first preset threshold value to obtain the labeling result;

Monitoring the updating frequency of a target sub-knowledge base, and labeling an expiration identifier for the target reply to obtain the labeling result under the condition that the updating frequency is smaller than a second preset threshold value, wherein the target sub-knowledge base is a sub-knowledge base for storing the target reply;

and monitoring version information of the target sub-knowledge base, and labeling an expiration identifier for the target reply under the condition that the version information is lower than preset version information to obtain the labeling result.

4. The method of claim 1, wherein the generating first feedback information having hint properties based on the initial question comprises:

extracting a target field capable of characterizing the initial question from the initial question;

According to the target field, a plurality of associated questions associated with the target field are retrieved from a history question record file;

Sorting the plurality of associated problems according to the frequency with which the associated problems are queried, so as to obtain a sorting result;

and processing a preset number of associated problems selected from the sorting result to obtain the first feedback information with the prompting property.

5. The method of claim 4, wherein processing the predetermined number of associated questions selected from the ranking results to obtain the first feedback information with prompt properties comprises:

Invoking a template of the first feedback information;

and splicing the initial problems, the preset number of associated problems and the templates of the first feedback information to obtain the first feedback information.

6. The method of claim 1, wherein the enterprise knowledge base comprises a plurality of sub knowledge bases, each sub knowledge base being obtained by integrating enterprise knowledge information of one dimension;

The retrieving a target answer to the target question from the enterprise knowledge base based on the target question, comprising:

determining a target sub-knowledge base from the plurality of sub-knowledge bases based on field information of the target problem;

The target answer is retrieved from the target sub-knowledge base based on the target question.

7. The method of claim 6, wherein retrieving the target answer from the target sub-knowledge base based on the target question comprises:

retrieving a set of enterprise knowledge information associated with the target problem from the target sub-knowledge base using a predetermined retrieval method;

calculating the similarity between the enterprise knowledge information in the enterprise knowledge information set and the target problem;

And taking enterprise knowledge information corresponding to the similarity meeting the preset condition as the target answer.

8. The method of claim 6, wherein the sub-knowledge base comprises at least one of: community data sub-knowledge base, expert consultation sub-knowledge base and platform application programming interface sub-knowledge base;

each sub knowledge base is obtained by integrating enterprise knowledge information in one dimension, and comprises the following steps:

the community knowledge base is obtained by integrating enterprise knowledge information for maintaining enterprise work posts;

the expert consultation sub-knowledge base is obtained by integrating an expert information table corresponding to the responsibility of the enterprise working position;

and integrating query application programming interfaces of the enterprise data query system to obtain the platform application programming interface sub-knowledge base.

9. The method of claim 1, wherein the parsing the information to be processed carried in the question-answering task to obtain the initial question asked by the user includes:

normalizing the information to be processed into initial text information;

Removing noise information in the initial text information to obtain intermediate text information;

Word segmentation is carried out on the intermediate text information, and target text information is obtained;

and marking the part of speech of the target file information to obtain the initial problem.

10. The method of claim 9, wherein the information to be processed comprises an image to be processed and text to be processed;

The normalizing the information to be processed into the initial text information comprises the following steps:

Identifying an object in the image to be processed to obtain an identification result, and taking the identification result as tag information of the image to be processed;

converting the tag information into text information of the image to be processed by using a natural language processing technology;

And integrating the text information of the image to be processed with the text to be processed to obtain the initial text information.

11. An intelligent question-answering device, characterized in that the device comprises:

The analysis module is used for responding to a question-answer task sent by a user and analyzing information to be processed carried in the question-answer task to obtain an initial question of the user question;

the first generation module is used for generating first feedback information with prompt property based on the initial problem under the condition that a reply matched with the initial problem does not exist in an enterprise knowledge base, wherein the enterprise knowledge base is obtained by integrating enterprise knowledge information in multiple dimensions;

the determining module is used for responding to the operation of the user on the first feedback information and determining a target question asked by the user;

A retrieval module for retrieving a target answer to the target question from the enterprise knowledge base based on the target question;

and the first pushing module is used for pushing the target answer to the user.

12. An electronic device, comprising:

one or more processors;

a memory for storing one or more computer programs,

Characterized in that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 10.

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

14. A computer program product comprising a computer program or instructions which, when executed by a processor, implement the steps of the method according to any one of claims 1 to 10.