Human-computer interaction method and device applied to question and answer, equipment and storage medium

By pre-constructing a problem list and reducing large-scale model calls, the problem of inefficient human-computer interaction technology is solved, and a fast response and efficient human-computer interaction experience is achieved.

CN119089012APending Publication Date: 2024-12-06BEIJING BAIDU NETCOM SCI & TECH CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411237138.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-04
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing human-computer interaction technology takes a long time to respond to user requests and consumes a lot of resources, resulting in low efficiency and poor user experience, especially in scenarios where real-time requirements are high.

Method used

Determine the interest information by responding to the user's interaction behavior information, pre-construct the question list, and directly find answers from the database when the user asks questions, reduce the number of calls to the big model and the consumption of computing resources, and use prediction and preparation mechanisms to improve the response speed and efficiency.

Benefits of technology

It significantly improves the efficiency and user experience of human-computer interaction, reduces operating costs, reduces the time for users to wait for answers, and improves the accuracy and accuracy of interaction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089012A_ABST
    Figure CN119089012A_ABST
Patent Text Reader

Abstract

The invention provides a human-computer interaction method, device and equipment applied to questions and answers and a storage medium, and relates to the field of artificial intelligence, in particular to the field of human-computer interaction. According to the specific implementation scheme, in response to interaction behavior information of a user, interest information of the user is determined; wherein the interaction behavior information represents a man-machine interaction behavior made by the user, and the interest information represents a field in which the user is interested; determining a question list corresponding to the interest information; wherein the question list comprises questions belonging to the interest information; responding to question information sent by the user, if the question information exists in the question list, searching reply information corresponding to the question information from a preset database, and feeding back the reply information to the user; wherein the reply information represents answers corresponding to the question information. Before the user puts forward the question, the question list is prepared in advance, and after the question information is obtained, the corresponding question is searched from the pre-prepared question list, so that the man-machine interaction efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of human-computer interaction in the field of artificial intelligence, and in particular to a human-computer interaction method, device, equipment and storage medium for question-answering. Background Art

[0002] Through human-computer interaction, users can ask questions and get answers, realizing intelligent dialogue in the form of question and answer. Currently, users can input questions in the form of text or voice, and the system processes the input questions and outputs the answers to the questions.

[0003] However, current human-computer interaction applications face problems such as high resource consumption and slow response speed, which affects the efficiency of human-computer interaction and thus affects the user's interactive experience. Summary of the invention

[0004] The present invention provides a human-computer interaction method, device, equipment and storage medium for question and answer.

[0005] According to a first aspect of the present disclosure, a human-computer interaction method for question answering is provided, comprising:

[0006] In response to the user's interactive behavior information, determine the user's interest information; wherein the interactive behavior information represents the human-computer interactive behavior performed by the user, and the interest information represents the user's area of ​​interest;

[0007] Determine a list of questions corresponding to the interest information; wherein the list of questions includes questions belonging to the interest information;

[0008] In response to a question sent by a user, if the question exists in the question list, reply information corresponding to the question is searched from a preset database, and the reply information is fed back to the user; wherein the reply information represents the answer corresponding to the question.

[0009] According to a second aspect of the present disclosure, a human-computer interaction device for question-answering is provided, comprising:

[0010] An interest determination unit, configured to determine the user's interest information in response to the user's interaction behavior information; wherein the interaction behavior information represents the human-computer interaction behavior performed by the user, and the interest information represents the user's area of ​​interest;

[0011] A list determining unit, configured to determine a list of questions corresponding to the interest information; wherein the list of questions includes questions belonging to the interest information;

[0012] A reply determination unit is used to respond to a question information sent by a user. If the question information exists in the question list, the reply information corresponding to the question information is searched from a preset database, and the reply information is fed back to the user; wherein the reply information represents the answer corresponding to the question information.

[0013] According to a third aspect of the present disclosure, there is provided an electronic device, including:

[0014] at least one processor; and

[0015] a memory communicatively coupled to the at least one processor;

[0016] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect of the present disclosure.

[0017] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method according to the first aspect of the present disclosure.

[0018] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, which implements the steps of the method described in the first aspect of the present disclosure when executed by a processor.

[0019] According to the technology disclosed in the present invention, the efficiency of human-computer interaction is improved.

[0020] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0022] Figure 1 It is a flowchart of a human-computer interaction method applied to question and answer provided according to an embodiment of the present disclosure;

[0023] Figure 2 It is a flowchart of a human-computer interaction method applied to question and answer provided according to an embodiment of the present disclosure;

[0024] Figure 3 is a schematic diagram of a knowledge graph provided according to an embodiment of the present disclosure;

[0025] Figure 4 is a module structure diagram of a human-computer interaction system provided according to an embodiment of the present disclosure;

[0026] Figure 5 is a schematic diagram of a human-computer interaction process for multiple responses provided according to an embodiment of the present disclosure;

[0027] Figure 6 is a structural block diagram of a human-computer interaction device for question-answering provided according to an embodiment of the present disclosure;

[0028] Figure 7 is a structural block diagram of a human-computer interaction device for question-answering provided according to an embodiment of the present disclosure;

[0029] Figure 8 is a block diagram of an electronic device for implementing the human-computer interaction method for question-answering according to an embodiment of the present disclosure;

[0030] Fig. 9 It is a block diagram of an electronic device used to implement the human-computer interaction method for question-answering according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0031] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0032] When users conduct human-computer interaction, they ask questions and automatically obtain answers. When responding to user requests, current human-computer interaction technology can directly process the input questions and generate answers by LLM (Large Language Model), which is a time-consuming and resource-intensive process.

[0033] In other words, the operation of LLM requires huge computing resources, especially GPU (Graphics Processing Unit) resources are very expensive, and the cost of human-computer interaction is very high. In addition, in scenarios with high real-time requirements, the response speed of LLM is very slow, which is difficult to meet the actual needs of users. The efficiency of human-computer interaction is low, and the user's human-computer interaction experience is poor.

[0034] The present invention provides a human-computer interaction method, device, equipment and storage medium for question answering, which are applied to the human-computer interaction field in the field of artificial intelligence to improve the efficiency of human-computer interaction.

[0035] It should be noted that the model in this embodiment is not a model for a specific user and cannot reflect the personal information of a specific user. It should be noted that the data in this embodiment comes from a public data set.

[0036] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0037] In order to enable readers to more deeply understand the implementation principle of the present disclosure, the following Figure 1-Figure 8 The embodiment is further refined.

[0038] Figure 1 FIG. 1 is a flow chart of a human-computer interaction method for question-answering provided according to an embodiment of the present disclosure, which can be executed by a human-computer interaction device for question-answering. Figure 1 As shown, the method comprises the following steps:

[0039] S101. Determine the user's interest information in response to the user's interactive behavior information; wherein the interactive behavior information represents the human-computer interactive behavior performed by the user, and the interest information represents the user's area of ​​interest.

[0040] Exemplarily, a user can respond to AI (Artificial Intelligence) through interactive behavior. For example, a user can input a question, and AI outputs an answer to the user, that is, the human-computer interaction behavior performed by the user can be the behavior of inputting a question. The user's human-computer interaction behavior can be recorded in real time to obtain the user's interactive behavior information, that is, the interactive behavior information can characterize the human-computer interaction behavior. For example, the interactive behavior information can include the type and content of the human-computer interaction behavior. The human-computer interaction behavior can include multiple types, for example, it can include the user's login behavior, input behavior, confirmation behavior, etc.

[0041] It can respond to the user's interactive behavior information in real time and determine what interactive behavior the user has performed. Based on the user's interactive behavior information, the user's interest information can be determined. The interest information can represent the user's area of ​​interest. For example, if the user is interested in computer knowledge, the interest information can be "computer", and the question the user asks next is likely to be related to the area corresponding to the interest information.

[0042] The interactive behavior information at different times may be different. The interactive behavior information of the user can be determined once before the user asks a question each time. For example, after the user logs in, the interactive behavior information is determined before the first question is asked; after the user asks the first question, the interactive behavior information is determined again before the user asks the second question. Different interest information can be determined for different interactive behavior information of the user. For example, the interactive behavior information represents the human-computer interactive behavior of the user in a historical time period. Based on the interactive behavior information, the content that the user frequently searches for in the past period of time can be determined, and the frequently searched content can be determined as the user's interest information. If the content that the user frequently searches for changes, the interest information may also be different.

[0043] In this embodiment, in response to the user's interactive behavior information, the user's interest information is determined, including: if it is determined based on the interactive behavior information that the user has not asked a question within a preset second time period, the user's account information is obtained as interest information; wherein the account information includes at least one of the user's personal attribute information, location information, and historical behavior information.

[0044] Specifically, the user's interactive behavior information may represent the user's login behavior, or any behavior before the first question is asked during this round of answering. Based on the interactive behavior information, it can be determined whether the user has asked a question in this round of answering. A time period can be preset as the second time period. Based on the interactive behavior information, it can be determined whether the user has asked a question in the second time period. If so, it means that the user has asked the first question in this round of answering; if not, it means that the user has not asked the first question in this round of answering.

[0045] If it is determined based on the interactive behavior information that the user has not sent out any question information within the preset second time period, the user's account information can be directly obtained as interest information. The account information can include at least one of the user's personal attribute information, location information, and historical behavior information. The personal attribute information can be various information related to the user, for example, it can be information such as the tags that the user marked for himself when registering an account, and it can also be information such as demographic attributes, interest characteristics, and social data; the location information can be the user's geographical location; the historical behavior information can be information related to the user's behavior in the historical time period. In other words, before the user asks the first question, the interest information is information related to the user's personal characteristics.

[0046] The beneficial effect of this setting is that the interest information corresponding to the user's first question and non-first question is different. Before the user asks the first question, it is not known what field the user will ask the question. Therefore, the interest information is information related to the user personally, and the knowledge field contained is wide, which makes it easy to find the user's first question from the question list, thereby improving the accuracy and efficiency of human-computer interaction.

[0047] In this embodiment, in response to the user's interactive behavior information, the user's interest information is determined, including: if it is determined based on the interactive behavior information that the user has sent a question information within a preset second time period, then the previous question information at the current moment is obtained as the interest information.

[0048] Specifically, if it is determined based on the interactive behavior information that the user has sent a question within the preset second time period, it is determined that the user has asked the first question in this round of answering, that is, the question asked by the subsequent user is not the first question.

[0049] When predicting and preparing for a non-first question, the last question asked by the user can be obtained as the current interest information. For example, the user asks the first question, and before asking the second question, the first question can be used as the interest information at the current moment; before asking the third question, the second question can be used as the interest information at the current moment. In other words, the interest information determined each time can be different, and the interest information is determined according to the question asked by the user.

[0050] The beneficial effect of this setting is that when the user is not asking a question for the first time, the previous question information is determined as interest information, and the user's question information is accurately predicted. It effectively grasps the knowledge field that the user currently wants to know, improves the accuracy of determining the question list, and then improves the accuracy of human-computer interaction and enhances the user's interactive experience.

[0051] S102: Determine a list of questions corresponding to the interest information; wherein the list of questions includes questions belonging to the interest information.

[0052] Exemplarily, different question lists may be determined according to different interest information. The question list includes multiple questions, and the questions in the question list are questions belonging to the interest information, that is, questions that the user may be interested in. For example, different question lists are preset for different fields, that is, different question lists correspond to different interest information.

[0053] The question list corresponding to the user's interest information can be determined based on the association between the preset interest information and the question list. It is also possible to first find all questions related to the field of interest information based on the interest information, and then select some questions from all questions to build a question list. For example, the number of times these questions are asked by all users within a period of time can be determined, and questions with a large number of questions can be selected to form a question list. The number of questions in the question list can be preset, for example, the number N can be preset, and the question list includes N questions.

[0054] The question list may only contain questions without answers to the questions, thereby reducing the memory occupied by the question list. For example, the number of times a question is asked and the answer to the question may be stored in a preset database, and the number of times a question is asked may be updated in real time with user interaction.

[0055] S103. In response to the question information sent by the user, if the question information exists in the question list, search for reply information corresponding to the question information from a preset database, and feed back the reply information to the user; wherein the reply information represents the answer corresponding to the question information.

[0056] Exemplarily, before the user asks a question, a list of questions is prepared in advance. For example, after the user logs in, the list of questions is determined first, and then the user is waited to send a question message, which means the question raised by the user.

[0057] After the user sends a question, the user's question information is determined, and it is judged whether the question information exists in the question list. If it exists, it means that the answer corresponding to the question information has been stored, and there is no need for a large model to generate the answer now. The answer corresponding to the question information can be found from the preset database, that is, the reply information corresponding to the question information. In other words, for the questions in the question list, the corresponding answers are stored in the preset database. The reply information is fed back to the user to complete a question and answer. For example, the reply information can be displayed on a visual interface for the user to view. If the question information does not exist in the question list, the reply information cannot be directly obtained, and the reply information needs to be calculated based on the question information. For example, semantic recognition is performed on the question information, and the reply information is generated based on the results of the semantic recognition.

[0058] In this embodiment, it also includes: if the question information does not exist in the question list, the question information is input into a preset large language model to obtain reply information corresponding to the question information; wherein the preset large language model is used to determine the reply information based on the question information; the question information and the corresponding reply information are associated and stored in a preset database.

[0059] Specifically, a large language model is pre-set. In this embodiment, the large language model is referred to as a large model. The input data of the large model can be question information, and the output data is reply information. The large model is a pre-trained neural network model. In this embodiment, the model structure of the large model is not specifically limited.

[0060] If it is determined that the question list does not contain the question information sent by the user, it is necessary to calculate the reply information based on the question information. The question information can be input into a preset large model, which extracts and calculates the features of the question information, outputs the reply information corresponding to the question information, and feeds the output reply information back to the user.

[0061] After determining that the question information does not exist in the question list, the question information and the corresponding reply information can also be associated and stored in a preset database, that is, in the subsequent user's questioning process, the question information can be added to the question list. If the user asks the question information, no further calculation is required, and the corresponding reply information can be directly obtained from the preset database. The question probability of the question information can also be stored in the database. Each question in the database corresponds to its own question probability. The question probability can represent the number of times the question is asked. For example, the question probability can be equal to the number of times the question is asked divided by the number of times all questions are asked.

[0062] The beneficial effect of this setting is that for question information that does not exist in the question list, the large model can be called to obtain the answer, thus avoiding the user from not getting a reply and improving the user's interactive experience.

[0063] In the disclosed embodiment, after responding to the user's interactive behavior, the user's field of interest is determined, and a list of questions corresponding to the field is prepared in advance, so as to predict and prepare the user's questions. When the user sends a question, it is directly checked from the question list whether there is a consistent question. If so, the answer to the question can be directly obtained. The number of calls to the large model and the consumption of computing resources are reduced, and the operating cost is reduced. Through the prediction and pre-preparation mechanism, the time users wait for answers can be reduced, the efficiency of human-computer interaction can be improved, and the user's interactive experience can be enhanced.

[0064] Figure 2 A flowchart of a human-computer interaction method for question and answering provided in an embodiment of the present disclosure.

[0065] In this embodiment, determining a list of questions corresponding to the interest information includes: determining candidate questions based on the interest information, and obtaining the question probability of the candidate questions; wherein the candidate questions represent the questions under the interest information, and the question probability represents the number of times the candidate questions are asked within a preset first time period; and sorting the candidate questions according to the question probability of the candidate questions to obtain a question list.

[0066] like Figure 2 As shown, the method comprises the following steps:

[0067] S201. Determine the user's interest information in response to the user's interactive behavior information; wherein the interactive behavior information represents the human-computer interactive behavior performed by the user, and the interest information represents the user's area of ​​interest.

[0068] Exemplarily, this step may refer to the above-mentioned step S101 and will not be described in detail.

[0069] S202. Determine candidate questions based on the interest information, and obtain the questioning probability of the candidate questions; wherein the candidate questions represent questions under the interest information, and the questioning probability represents the number of times the candidate questions are asked within a preset first time period.

[0070] Exemplarily, after obtaining the interest information, multiple questions corresponding to the interest information can be determined as candidate questions. Candidate questions are questions belonging to the field corresponding to the interest information. For example, if the field corresponding to the interest information is a computer-related field, the candidate questions may be who is the inventor of the computer, in which year the computer was invented, etc. Each field corresponds to multiple questions, and the questions corresponding to the interest information can be searched.

[0071] Each question has its own questioning probability. A time period is preset as the first time period. The questioning probability can represent the number of times the candidate question is asked in the preset first time period. For example, the questioning probability of a candidate question can be the number of times the candidate question is asked in the first time period divided by the number of times all candidate questions are asked in the first time period. The questioning probability of each candidate question can be updated in real time.

[0072] In this embodiment, the interest information includes at least one of the user's personal attribute information, location information, and historical behavior information; determining candidate questions based on the interest information includes: determining the user group to which the user belongs based on the interest information; wherein the user group represents the classification result of classifying users based on the interest information; and determining, based on the association relationship between the preset user group and the question, the question corresponding to the user group to which the user belongs as the candidate question.

[0073] Specifically, in each round of answering, before the user asks the first question, the interest information may be the user's personal attribute information, location information, and historical behavior information. Based on this interest information, the user can be categorized and the user group to which the user belongs can be determined. The user group can represent the classification result of the user. Different interest information can correspond to different user groups. The user's field of interest can be determined based on the interest information, thereby determining which user group the user belongs to. For example, if the interest information indicates that the user is a computer engineer, the user can be classified into the user group in the "computer" field.

[0074] Different user groups are pre-associated with multiple questions. According to the association between the preset user groups and the questions, the questions corresponding to the user group to which the user belongs are determined as candidate questions. Multiple candidate questions can be determined. For example, if the user group is "computer", the candidate questions may include who is the inventor of the computer, what is the CPU, how much is the price of the computer, etc.

[0075] The beneficial effect of such a setting is that users can be classified and the user groups to which they belong can be determined, thereby determining the candidate questions corresponding to the user groups, achieving accurate grasp of user interests, and improving the accuracy of human-computer interaction.

[0076] In this embodiment, determining the user group to which the user belongs based on the interest information includes: performing feature extraction processing on the interest information to obtain a user feature vector of the user; wherein the user feature vector represents the interest information of the user; and determining the user group to which the user belongs based on the association relationship between the preset user feature vector and the user group.

[0077] Specifically, before the user asks a question for the first time, the interest information obtained is information related to the user, and the user's personal characteristics can be extracted from the interest information. Users in the same user group have similar characteristics, so the user group to which the user belongs can be determined based on the user characteristics.

[0078] A feature extraction network may be preset, and feature extraction processing may be performed on the interest information through the feature extraction network. In this embodiment, the preset feature extraction network is not specifically limited. The extracted features are determined as user feature vectors. A plurality of user groups are preset, and different user groups correspond to different user feature vectors. According to the extracted user feature vector, a user group corresponding to the user feature vector is determined as the user group to which the user belongs.

[0079] When multiple user groups are pre-set, multiple user feature vectors may be pre-collected, and the user feature vectors may be extracted from demographic attributes, interest features, location features, behavior data, and social data, etc. A preset clustering algorithm is used to perform cluster analysis on the pre-collected user feature vectors, and the multiple user feature vectors are divided into different clusters, each cluster representing a user group. In this embodiment, the preset clustering algorithm is not specifically limited.

[0080] The beneficial effect of such a setting is that user characteristics are extracted from interest information, thereby accurately classifying users, improving the accuracy of determining the question list, and further improving the accuracy of human-computer interaction.

[0081] In this embodiment, the method further includes: if the question information does not exist in the question list, adding the question information to the questions associated with the user group to which the user belongs, and identifying the question probability of the question information.

[0082] Specifically, if the user has not asked a question for the first time, after determining the question list, respond to the question information sent by the user and determine whether the question information exists in the question list. If it does, the reply information can be directly obtained; if it does not, it means that the question information is not in the questions associated with the user group. The question information can be associated with the user group as a new question, that is, the question information is added to the questions associated with the user group to which the user belongs. In the subsequent new round of responses, if it is determined that the user group to which the user belongs is still the user group, the newly added question can be placed in the question list.

[0083] For a newly added question, the probability of the question being asked can be marked. When the question is added, the number of times the question is asked is 1, and the probability of the question being asked is small. As the number of subsequent questions increases, the probability of the question being asked can change in real time. The probability of the question being asked corresponding to each question can be stored in a preset database or in a question list.

[0084] The beneficial effect of such a setting is that if there is no question information in the question list, the question list or the association relationship can be updated, so that as the user continues to use it, all the user's questions are gradually covered, making the performance of human-computer interaction better and better.

[0085] In this embodiment, the interest information is the previous question information at the current moment; based on the interest information, candidate questions are determined, including: if the interest information exists in a preset knowledge graph, then the sub-node of the interest information is searched from the preset knowledge graph; wherein the preset knowledge graph includes multiple nodes, and one node represents a question; the question corresponding to the sub-node of the interest information is determined as the candidate question.

[0086] Specifically, in this round of answering, if the user has already asked the first question, before asking the second, third, and other subsequent questions, the interest information obtained is the previous question information of the question information that has been asked, that is, the previous question information at the current moment. For example, before asking the second question, the interest information is the first question.

[0087] A knowledge graph is pre-set, which includes multiple questions. Each question can be represented as a node. The nodes corresponding to the questions can be connected by edges. The connection relationship can represent the hierarchical relationship between the nodes. For example, a parent node can be connected to multiple child nodes. The knowledge graph can contain questions but not the answers corresponding to the questions.

[0088] Determine whether the interest information exists in the knowledge graph, that is, determine whether the previous question exists in the knowledge graph. If the interest information exists in the knowledge graph, you can find the child node of the interest information from the knowledge graph, that is, find the next-level node connected to the node of the previous question. Determine the question corresponding to the child node of the interest information as a candidate question. For example, if there are ten child nodes corresponding to the interest information, there can be ten candidate questions. Each node in the knowledge graph corresponds to its own question probability, and some questions can be selected as candidate questions from the questions corresponding to these child nodes based on the question probability. For example, if the number of candidate questions is preset to N, the top N questions can be selected as candidate questions based on the question probability.

[0089] That is to say, before asking the first question, the user group is determined based on the user characteristics, and the candidate questions corresponding to the user group are determined, so as to construct the first question list and determine whether the first question exists in the question list. All candidate questions corresponding to the user group exist in the knowledge graph. If the first question exists in the first question list, then before asking the second question, the first question is found from the preset knowledge graph, and the questions of the subordinate nodes connected to the first question are used as new candidate questions to construct the second question list. If the second question exists in the second question list, then before asking the third question, the second question is found from the preset knowledge graph, and the questions of the subordinate nodes connected to the second question are determined and used as new candidate questions again. If a question asked by the user does not exist in the corresponding question list, the subsequent questions can use the large model to generate answers, and the questions that did not appear can be added to the knowledge graph.

[0090] The beneficial effect of this setting is that candidate questions directly related to interest information can be searched level by level in the knowledge graph, so as to achieve accurate prediction of user questions, reduce user waiting time, significantly improve response speed, and enhance user experience.

[0091] In this embodiment, it also includes: if the question information does not exist in the question list, the question information is determined as a new child node of the interest information, and the new child node is added to the preset knowledge graph; and the question probability of the question information is added to the edge connection between the question information and the interest information.

[0092] Specifically, for each question after the first question, if it is determined that the question information does not exist in the question list, the question information can be determined as a new child node of the interest information in the knowledge graph, and the new child node is added to the preset knowledge graph. That is, an edge connection is added under the node of the interest information to connect the question information.

[0093] Each edge connection in the knowledge graph can be marked with a question probability, which indicates the question probability of the child node connected to the edge connection. As users ask questions, the question probabilities can change in real time. When adding a new child node, you can also add the corresponding question probability to the child node. That is, add the question probability of the question information to the edge connection between the question information and the interest information.

[0094] The beneficial effect of this setting is that with user interaction, the knowledge graph can be continuously optimized, the prediction accuracy and response speed can be improved, thereby reducing the number of calls to large models, reducing resource consumption, and improving the efficiency and accuracy of interaction.

[0095] S203. Sort the candidate questions according to their question probabilities to obtain a question list.

[0096] Exemplarily, the probability of each candidate question being asked is determined, and the candidate questions are arranged according to the size of the probability of being asked, for example, they can be arranged in order from large to small. A question list is formed based on the sorted results. That is, the candidate questions can be arranged from top to bottom in the question list, and the candidate question with the highest probability of being asked is arranged at the top.

[0097] In this embodiment, different interest information corresponds to different candidate questions, and the candidate questions are arranged into a question list according to the probability of being asked, so that in subsequent judgments, it is possible to give priority to judging whether the questions asked by the user are predicted questions that may be asked, thereby reducing the user's waiting time and improving the efficiency of human-computer interaction.

[0098] S204. In response to the question information sent by the user, if the question information exists in the question list, search for reply information corresponding to the question information from a preset database, and feed back the reply information to the user; wherein the reply information represents the answer corresponding to the question information.

[0099] For example, this step may refer to the above step S103 and will not be described in detail.

[0100] Figure 3 A schematic diagram of the knowledge graph. Figure 3 In the knowledge graph, "What is a computer" has three sub-nodes. If the first question asked is "What is a computer", the predicted second question may be "What parts does a computer consist of", "What is the price of a computer", and "Who is the inventor of the computer". If the second question asked by the user is "Who is the inventor of the computer", the questions corresponding to the two sub-nodes can be found in the knowledge graph to construct a question list.

[0101] Figure 4 This is a module structure diagram of the human-computer interaction system. Figure 4 In the human-computer interaction system, a prediction module, a user group management module, a knowledge graph management module, a database, and a large model are included. The database stores all questions and answers to the questions. The user group management module is used to determine the user group to which the user belongs when the user asks a question for the first time, and then determine the question list corresponding to the first question, and send the question list to the prediction module. The knowledge graph management module is used to build a question list from the knowledge graph when the user does not ask a question for the first time, and send the question list to the prediction module. The prediction module is used to receive the user's question information and determine whether the question information exists in the question list. If the question information exists in the question list, the user group management module or the knowledge graph management module can be informed to update the question probability and obtain the answer to the question. If the question information does not exist in the question list, the user group management module or the knowledge graph management module is informed to add a new question and obtain the answer to the question based on the large model.

[0102] In the disclosed embodiment, after responding to the user's interactive behavior, the user's field of interest is determined, and a list of questions corresponding to the field is prepared in advance, so as to predict and prepare the user's questions. When the user sends a question, it is directly checked from the question list whether there is a consistent question. If so, the answer to the question can be directly obtained. The number of calls to the large model and the consumption of computing resources are reduced, and the operating cost is reduced. Through the prediction and pre-preparation mechanism, the time users wait for answers can be reduced, the efficiency of human-computer interaction can be improved, and the user's interactive experience can be enhanced.

[0103] Figure 5 Schematic diagram of the human-computer interaction process for multiple responses provided by the embodiment of the present disclosure. Figure 5 As shown, in the scenario of multiple responses, the process of human-computer interaction can be:

[0104] For a whole round of responses, before the user logs in to the human-computer interaction system and sends the first question, the user's interest information is determined based on the user's login behavior or other interactive behaviors. For example, the user's historical search records can be determined based on the user's account information, and the knowledge field corresponding to the historical search records is used as the user's initial interest information. For example, the user's initial interest information is characterized as the computer field.

[0105] Then, based on the initial interest information, the user group to which the user belongs is determined, and a list of questions corresponding to the user group is determined as the first question list. The question list contains multiple questions, each of which corresponds to a question probability. Figure 5 As shown, the initial interest information is the computer field, the user group to which the user belongs is a computer-related user group, and there are four questions in the corresponding question list, and the question probability decreases from top to bottom.

[0106] The user sends a first question message, obtains the first question message, and determines whether the first question message exists in a first question list. Figure 5 In the example, the question information raised by the user is the third question in the question list. Each question in the question list has an answer stored in advance, so the answer to the third question can be directly obtained and replied to the user.

[0107] Before the user asks the second question, the interest information becomes the first question information asked by the user. The second question list is generated based on the new interest information, that is, based on the first question information. It can be determined whether the first question information exists in the preset knowledge graph. If so, the next level question of the first question information is searched from the knowledge graph, and the second question list is generated based on the question probability of the next level question. Figure 5 As shown, there are two questions in the second question list, namely "What is a CPU" and "What is a disk", with question probabilities of 12% and 10% respectively.

[0108] The user sends a second question, obtains the second question, and determines whether the second question exists in the second question list. If so, the answer to the second question can be directly obtained and replied to the user.

[0109] Before the user asks the third question, it is determined whether the second question exists in the preset knowledge graph. If so, the next level question of the second question is searched in the knowledge graph, and the third question list is generated according to the probability of the next level question, until the user stops answering this round.

[0110] Figure 6This is a structural block diagram of a human-computer interaction device for question-answering provided by an embodiment of the present disclosure. For ease of explanation, only the parts related to the embodiment of the present disclosure are shown. Figure 6 The human-computer interaction device 600 applied to question and answer includes: an interest determination unit 601, a list determination unit 602 and a reply determination unit 603.

[0111] The interest determination unit 601 is used to determine the user's interest information in response to the user's interactive behavior information; wherein the interactive behavior information represents the human-computer interactive behavior performed by the user, and the interest information represents the user's area of ​​interest;

[0112] A list determining unit 602 is used to determine a question list corresponding to the interest information; wherein the question list includes questions belonging to the interest information;

[0113] The reply determination unit 603 is used to respond to the question information sent by the user. If the question information exists in the question list, the reply information corresponding to the question information is searched from a preset database, and the reply information is fed back to the user; wherein the reply information represents the answer corresponding to the question information.

[0114] Figure 7 A structural block diagram of a human-computer interaction device for question-answering provided in an embodiment of the present disclosure, such as Figure 7 As shown, the human-computer interaction device 700 applied to question and answer includes an interest determination unit 701, a list determination unit 702 and a reply determination unit 703, wherein the list determination unit 702 includes a candidate determination module 7021 and a list determination module 7022.

[0115] The candidate determination module 7021 is used to determine candidate questions according to the interest information, and obtain the questioning probability of the candidate questions; wherein the candidate questions represent questions under the interest information, and the questioning probability represents the number of times the candidate questions are asked within a preset first time period;

[0116] The list determination module 7022 is used to sort the candidate questions according to the questioning probabilities of the candidate questions to obtain the question list.

[0117] In one example, the interest information includes at least one of the user's personal attribute information, location information, and historical behavior information; the candidate determination module 7021 includes:

[0118] A first determination submodule is used to determine the user group to which the user belongs based on the interest information; wherein the user group represents a classification result of classifying users based on the interest information;

[0119] The second determination submodule is used to determine, based on the association relationship between the preset user group and the question, the question corresponding to the user group to which the user belongs as the candidate question.

[0120] In one example, the first determination submodule is specifically used to:

[0121] Performing feature extraction processing on the interest information to obtain a user feature vector of the user; wherein the user feature vector represents the user's interest information;

[0122] The user group to which the user belongs is determined according to the association relationship between the preset user feature vector and the user group.

[0123] In one example, it also includes:

[0124] The question adding unit is used to add the question information to the questions associated with the user group to which the user belongs if the question information does not exist in the question list, and identify the question probability of the question information.

[0125] In one example, the interest information is the last question information at the current moment; the candidate determination module 7021 includes:

[0126] The third determination submodule is used to search for a subnode of the interest information from the preset knowledge graph if the interest information exists in the preset knowledge graph; wherein the preset knowledge graph includes a plurality of nodes, and one node represents a question;

[0127] The fourth determination submodule is used to determine the question corresponding to the sub-node of the interest information as the candidate question.

[0128] In one example, it also includes:

[0129] A node adding unit, configured to determine the question information as a new child node of the interest information if the question information does not exist in the question list, and add the new child node to the preset knowledge graph;

[0130] The probability adding unit is used to add the question probability of the question information to the edge connection between the question information and the interest information.

[0131] In one example, the interest determination unit includes:

[0132] The first acquisition module is used to obtain the user's account information, which is the interest information, if it is determined based on the interactive behavior information that the user has not sent any question information within a preset second time period; wherein the account information includes at least one of the user's personal attribute information, location information, and historical behavior information.

[0133] In one example, the interest determination unit includes:

[0134] The second acquisition module is used to acquire the last question information at the current moment as the interest information if it is determined, based on the interaction behavior information, that the user has sent question information within a preset second time period.

[0135] In one example, it also includes:

[0136] A model application unit, configured to input the question information into a preset large language model to obtain reply information corresponding to the question information if the question information does not exist in the question list; wherein the preset large language model is used to determine the reply information according to the question information;

[0137] The information storage unit is used to associate and store the question information and the corresponding reply information in a preset database.

[0138] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device.

[0139] Figure 8 A structural block diagram of an electronic device provided in an embodiment of the present disclosure, such as Figure 8 As shown, the electronic device 800 includes: at least one processor 802; and a memory 801 communicatively connected to the at least one processor 802; wherein the memory stores instructions executable by the at least one processor 802, and the instructions are executed by the at least one processor 802 so that the at least one processor 802 can execute the human-computer interaction method applied to question and answer of the present disclosure.

[0140] The electronic device 800 further includes a receiver 803 and a transmitter 804. The receiver 803 is used to receive instructions and data sent by other devices, and the transmitter 804 is used to send instructions and data to external devices.

[0141] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0142] According to an embodiment of the present disclosure, the present disclosure also provides a computer program product, which includes: a computer program, the computer program is stored in a readable storage medium, at least one processor of an electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device executes the solution provided by any of the above embodiments.

[0143] Fig. 9A schematic block diagram of an example electronic device 900 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0144] like Fig. 9 As shown, the device 900 includes a computing unit 901, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The computing unit 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0145] A number of components in the device 900 are connected to the I / O interface 905, including: an input unit 906, such as a keyboard, a mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a disk, an optical disk, etc.; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 909 allows the device 900 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0146] The computing unit 901 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 901 performs the various methods and processes described above, such as a human-computer interaction method applied to question and answer. For example, in some embodiments, the human-computer interaction method applied to question and answer may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 908. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the human-computer interaction method applied to question and answer described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to execute the human-computer interaction method applied to question and answer in any other appropriate manner (eg, by means of firmware).

[0147] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0148] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0149] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0150] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0151] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0152] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0153] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0154] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A human-computer interaction method for question answering, comprising: In response to the user's interactive behavior information, determine the user's interest information; wherein the interactive behavior information represents the human-computer interactive behavior performed by the user, and the interest information represents the user's area of ​​interest; Determine a list of questions corresponding to the interest information; wherein the list of questions includes questions belonging to the interest information; In response to a question sent by a user, if the question exists in the question list, reply information corresponding to the question is searched from a preset database, and the reply information is fed back to the user; wherein the reply information represents the answer corresponding to the question.

2. The method according to claim 1, wherein: The determining a list of questions corresponding to the interest information includes: Determine candidate questions according to the interest information, and obtain the questioning probability of the candidate questions; wherein the candidate questions represent questions under the interest information, and the questioning probability represents the number of times the candidate questions are asked within a preset first time period; The candidate questions are sorted according to the questioning probabilities of the candidate questions to obtain the question list.

3. The method according to claim 2, wherein: The interest information includes at least one of the user's personal attribute information, location information, and historical behavior information; and determining candidate questions based on the interest information includes: Determine the user group to which the user belongs based on the interest information; wherein the user group represents a classification result of classifying users based on the interest information; According to the association relationship between the preset user group and the question, the question corresponding to the user group to which the user belongs is determined as the candidate question.

4. The method according to claim 3, wherein: The determining, based on the interest information, the user group to which the user belongs includes: Performing feature extraction processing on the interest information to obtain a user feature vector of the user; wherein the user feature vector represents the user's interest information; The user group to which the user belongs is determined according to the association relationship between the preset user feature vector and the user group.

5. The method according to claim 3 or 4, further comprising: If the question information does not exist in the question list, the question information is added to the questions associated with the user group to which the user belongs, and the question probability of the question information is identified.

6. The method according to any one of claims 2 to 5, wherein: The interest information is the last question information at the current moment; and determining the candidate questions according to the interest information includes: If the interest information exists in a preset knowledge graph, searching for a subnode of the interest information from the preset knowledge graph; wherein the preset knowledge graph includes a plurality of nodes, and one node represents a question; The question corresponding to the sub-node of the interest information is determined as the candidate question.

7. The method according to claim 6, further comprising: If the question information does not exist in the question list, determining the question information as a new child node of the interest information, and adding the new child node to the preset knowledge graph; The question probability of the question information is added to the edge connection between the question information and the interest information.

8. The method according to any one of claims 1 to 7, wherein: The determining the user's interest information in response to the user's interactive behavior information includes: If it is determined based on the interactive behavior information that the user has not sent any question information within the preset second time period, the user's account information is obtained as the interest information; wherein the account information includes at least one of the user's personal attribute information, location information, and historical behavior information.

9. The method according to claim 8, wherein: The determining the user's interest information in response to the user's interactive behavior information includes: If it is determined according to the interactive behavior information that the user has sent question information within the preset second time period, then the last question information at the current moment is obtained as the interest information.

10. The method according to any one of claims 1 to 9, further comprising: If the question information does not exist in the question list, the question information is input into a preset large language model to obtain reply information corresponding to the question information; wherein the preset large language model is used to determine the reply information according to the question information; The question information and the corresponding reply information are associated and stored in a preset database.

11. A human-computer interaction device for question answering, comprising: An interest determination unit, configured to determine the user's interest information in response to the user's interaction behavior information; wherein the interaction behavior information represents the human-computer interaction behavior performed by the user, and the interest information represents the user's area of ​​interest; A list determining unit, configured to determine a list of questions corresponding to the interest information; wherein the list of questions includes questions belonging to the interest information; A reply determination unit is used to respond to a question information sent by a user. If the question information exists in the question list, the reply information corresponding to the question information is searched from a preset database, and the reply information is fed back to the user; wherein the reply information represents the answer corresponding to the question information.

12. The device according to claim 11, wherein The list determination unit comprises: A candidate determination module, configured to determine candidate questions according to the interest information, and obtain a questioning probability of the candidate questions; wherein the candidate questions represent questions under the interest information, and the questioning probability represents the number of times the candidate questions are asked within a preset first time period; The list determination module is used to sort the candidate questions according to the questioning probabilities of the candidate questions to obtain the question list.

13. The device according to claim 12, wherein: The interest information includes at least one of the user's personal attribute information, location information, and historical behavior information; The candidate determination module comprises: A first determination submodule is used to determine the user group to which the user belongs based on the interest information; wherein the user group represents a classification result of classifying users based on the interest information; The second determination submodule is used to determine, based on the association relationship between the preset user group and the question, the question corresponding to the user group to which the user belongs as the candidate question.

14. The device according to claim 13, wherein: The first determining submodule is specifically used for: Performing feature extraction processing on the interest information to obtain a user feature vector of the user; wherein the user feature vector represents the user's interest information; The user group to which the user belongs is determined according to the association relationship between the preset user feature vector and the user group.

15. The device according to claim 13 or 14, further comprising: The question adding unit is used to add the question information to the questions associated with the user group to which the user belongs if the question information does not exist in the question list, and identify the question probability of the question information.

16. The device according to any one of claims 12 to 15, wherein: The interest information is the last question information at the current moment; The candidate determination module comprises: The third determination submodule is used to search for a subnode of the interest information from the preset knowledge graph if the interest information exists in the preset knowledge graph; wherein the preset knowledge graph includes a plurality of nodes, and one node represents a question; The fourth determination submodule is used to determine the question corresponding to the sub-node of the interest information as the candidate question.

17. The apparatus according to claim 16, further comprising: A node adding unit, configured to determine the question information as a new child node of the interest information if the question information does not exist in the question list, and add the new child node to the preset knowledge graph; The probability adding unit is used to add the question probability of the question information to the edge connection between the question information and the interest information.

18. The device according to any one of claims 11 to 17, wherein: The interest determination unit comprises: The first acquisition module is used to obtain the user's account information, which is the interest information, if it is determined based on the interactive behavior information that the user has not sent any question information within a preset second time period; wherein the account information includes at least one of the user's personal attribute information, location information, and historical behavior information.

19. The device according to claim 18, wherein The interest determination unit comprises: The second acquisition module is used to acquire the last question information at the current moment as the interest information if it is determined, based on the interaction behavior information, that the user has sent question information within a preset second time period.

20. The apparatus according to any one of claims 11 to 19, further comprising: A model application unit, configured to input the question information into a preset large language model to obtain reply information corresponding to the question information if the question information does not exist in the question list; wherein the preset large language model is used to determine the reply information according to the question information; The information storage unit is used to associate and store the question information and the corresponding reply information in a preset database.

21. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 10.

22. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-10.

23. A computer program product, wherein: The invention comprises a computer program, which implements the steps of the method according to any one of claims 1 to 10 when being executed by a processor.

Citation Information

Cited By

  • Intelligent scientific investigation comprehensive management method and system

    CN120743158A