Automatic question answering method, device and equipment based on human-computer interaction and medium

Abstract

The invention provides an automatic question answering method and device based on human-computer interaction, equipment and a medium, and relates to the field of artificial intelligence, in particular to the technical field of intelligent search and natural language processing. According to the specific implementation scheme, in response to a question answering request triggered by a user through a question answering robot, if it is determined that a storage task is received, the storage task is stored in a knowledge base system, and then the question answering request is sent to the knowledge base system; wherein the storage task represents reply information to be stored in the knowledge base system; the question answering request is used for determining reply information corresponding to the question answering request from a knowledge base system; and receiving reply information fed back by the knowledge base system, and feeding back the reply information fed back by the knowledge base system to the user through the question answering robot. Through the method, the accuracy of the reply information corresponding to the question answering request is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of intelligent search and natural language processing technology in the field of artificial intelligence, and in particular to an automatic question-answering method, device, equipment and medium based on human-computer interaction. Background Art

[0002] With the development of intelligent search technology, users can be provided with intelligent question-answering methods to provide users with reply information.

[0003] Furthermore, there is an urgent need for a method that can accurately and effectively provide reply information to users, so as to improve the accuracy of the reply information and enhance the user experience. Summary of the invention

[0004] The present invention provides a method, device, equipment and medium for automatic question answering based on human-computer interaction.

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

[0006] In response to a question-answering request triggered by a user through a question-answering robot, if it is determined that a storage task has been received, after storing the storage task in a knowledge base system, the question-answering request is sent to the knowledge base system; wherein the storage task represents the reply information to be stored in the knowledge base system; the question-answering request is used to determine the reply information corresponding to the question-answering request from the knowledge base system;

[0007] Receive the reply information fed back by the knowledge base system, and feed back the reply information fed back by the knowledge base system to the user through the question answering robot.

[0008] According to a second aspect of the present disclosure, there is provided an automatic question-answering device based on human-computer interaction, comprising:

[0009] A processing unit, configured to respond to a question-answering request triggered by a user through a question-answering robot, and if it is determined that a storage task has been received, after storing the storage task in a knowledge base system, send the question-answering request to the knowledge base system; wherein the storage task represents reply information to be stored in the knowledge base system; and the question-answering request is used to determine reply information corresponding to the question-answering request from the knowledge base system;

[0010] The forwarding unit is used to receive the reply information fed back by the knowledge base system, and feed back the reply information fed back by the knowledge base system to the user through the question-answering robot.

[0011] According to a third aspect of the present disclosure, a computer program product is provided, comprising: a computer program, wherein 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 the at least one processor executes the computer program so that the electronic device executes the method described in the first aspect.

[0012] 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 described in the first aspect.

[0013] According to a fifth aspect of the present disclosure, a question-answering platform is provided, comprising:

[0014] A question-answering center system; and a question-answering robot and a knowledge base system connected to the question-answering center system; wherein the question-answering center system is capable of executing the method described in the first aspect.

[0015] The technology disclosed in the present invention improves the accuracy of reply information corresponding to question-answering requests.

[0016] 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

[0017] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0018] Figure 1 is a scene graph that can implement the embodiments of the present disclosure;

[0019] Figure 2 It is a flowchart of an automatic question-answering method based on human-computer interaction provided by the first embodiment of the present disclosure;

[0020] Figure 3 is a flowchart of an automatic question-answering method based on human-computer interaction provided in the second embodiment of the present disclosure;

[0021] Figure 4 It is a structural block diagram of an automatic question-answering device based on human-computer interaction provided by an embodiment of the present disclosure;

[0022] Figure 5 It is a block diagram of an electronic device used to implement the automatic question-answering method based on human-computer interaction according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0023] 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.

[0024] In the related art, with the development of intelligent search technology, users can ask questions through the search system. However, when using the search system to search for faults or solutions to problems encountered during the use of the search system, the search system usually cannot provide accurate answers, which reduces the user experience.

[0025] In view of this, the present disclosure provides an automatic question-answering method based on human-computer interaction, and provides a human-computer interaction question-answering interface. Users can input question information based on the question-answering interface, and R&D personnel can upload discovered problems and solutions to the problems based on the question-answering interface, and ensure that the solutions uploaded by R&D personnel can be processed with priority, thereby providing a more comprehensive data source for answering user questions, improving the accuracy of answering questions, and enhancing user experience.

[0026] The present invention discloses an automatic question-answering method, device, equipment and medium based on human-computer interaction, which are applied to the fields of intelligent search and natural language processing technology in the field of artificial intelligence to achieve the effect of improving the accuracy of reply information.

[0027] It should be noted that 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.

[0028] Figure 1 is a scene graph that can implement the embodiments of the present disclosure. Figure 1 As shown, the question answering platform 100 may include a question answering robot 101 , a question answering center system 102 and a knowledge base system 103 .

[0029] The question answering robot 101 is used to provide a human-computer interaction interface, receive a question answering request input by a user or supplementary information uploaded by a developer, and store the supplementary information in the knowledge base system as reply information to the question answering request. It should be noted that the present disclosure does not limit the type or interface style of the question answering robot. For example, the question answering robot can be a web version or a mobile application version.

[0030] The question answering center system 102 is used to pre-process the received question answering request or the received supplementary information and forward it to the knowledge base system.

[0031] The knowledge base system 103 is used to reply to the question-answering request based on the stored reply information, or to store the received supplementary information as reply information to the question-answering request.

[0032] Optionally, a large language model may be provided in the knowledge base system to answer question and answer requests. It should be noted that the present disclosure does not limit the type of the large language model, nor does it limit the deployment method of the large language model. For example, the knowledge base system may be integrated with an interface (Application Programming Interface, API) of the large language model, and the knowledge base system may call the interface to process the question and answer request and obtain reply information to the question and answer request.

[0033] Optionally, the knowledge base system may include a database for storing reply information. It should be noted that the present disclosure does not limit the type of database, nor does it limit the deployment method of the database, for example, it may be deployed in the cloud or integrated with the knowledge base system.

[0034] The question-answering robot can be deployed on the terminal device of the user or the R&D personnel. The question-answering center system and the knowledge base system can be integrated and deployed on the same server, or can be distributed and deployed in a server cluster. No limitation is made here.

[0035] Optionally, the question answering robot can be integrated with the search system. For example, an icon of the question answering robot can be set on the front-end display interface of the search system as an entry to the question answering robot. In response to the user's triggering operation on the question answering robot icon, the search system provides the user with a question answering robot's question answering interface so that the user can input a question answering request. It should be noted that the present disclosure does not limit the icon style and display position of the question answering robot.

[0036] In this application scenario, the Q&A platform can provide users with reply information based on the Q&A request input by the user through the Q&A robot, through the Q&A center system and the knowledge base system. The Q&A platform can also store the reply information in the knowledge base system based on the reply information uploaded by the R&D personnel through the Q&A robot, through the Q&A center system.

[0037] Understandably, Figure 1 The scene graph shown is only an exemplary description. The embodiment of the present disclosure does not limit the actual form of the various devices included in the scene graph, nor does it limit the interaction mode between the devices. In the specific application of the solution, it can be set according to actual needs.

[0038] The following specific embodiments are used to describe in detail the technical solution of the present invention and how the technical solution of the present invention solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present invention will be described below in conjunction with the accompanying drawings.

[0039] Figure 2 : is a flowchart of an automatic question answering method based on human-computer interaction provided by the first embodiment of the present disclosure. The execution subject of this embodiment may be the aforementioned question answering center system. The method may include the following steps:

[0040] S201. In response to a question answering request triggered by a user through a question answering robot, determine whether a storage task has been received.

[0041] Exemplarily, the Q&A robot can receive question information input by the user, generate a Q&A request based on the question information, and send the Q&A request to the Q&A center system. The question information may be, for example, "What does it mean when the question code 400 appears during the search process?", or "How to solve the problem code 400?", etc. It should be noted that the present disclosure does not limit the content included in the Q&A request. The Q&A request is used to determine the reply information corresponding to the Q&A request from the knowledge base system. In other words, the Q&A request is used to determine the reply information corresponding to the Q&A request, wherein the reply information is used to represent the answer to the Q&A request, for example, it may be an explanation of the meaning of the problem code, or a solution to the problem. It should be noted that the present disclosure does not limit the content of the reply information corresponding to the Q&A request.

[0042] It should be noted that the present disclosure does not limit how to trigger a question-answering request through a question-answering robot. For example, a user may input question information on the interface of the question-answering robot through an input device to generate a question-answering request. The input device may include, for example, a keyboard, mouse, microphone, stylus, etc.

[0043] Optionally, after receiving the question information input by the user, the question answering robot may adjust the question information to obtain the adjusted question information, and display the adjusted question information on the interface of the question answering robot for the user to confirm. If the user determines that the adjusted question information is correct, the adjusted question information is sent to the question answering center system; if the user denies the adjusted question information, the original question information is sent to the question answering center system.

[0044] Among them, the question-answering robot can use a preset word segmentation model to perform word segmentation on the question information to obtain multiple keywords, wherein the multiple keywords include one or more of the subject, predicate and object, and group the multiple keywords into triples, and use the triples to match the question library, wherein the question library includes the correspondence between multiple triples and the question information, therefore, the matching probability of the triple with each triple in the question library can be obtained, and the question information corresponding to the triple with the largest probability is determined as the adjusted question information.

[0045] Through this step, the Q&A robot can send more accurate Q&A requests to the Q&A center system, thereby improving the accuracy of the reply information corresponding to the Q&A requests.

[0046] As mentioned above, the question-answering robot can receive reply information uploaded by R&D personnel, generate a storage task based on the reply information, and send the storage task to the question-answering center system. The storage task represents the reply information to be stored in the knowledge base system. The reply information in the storage task refers to the supplement to the reply information stored in the knowledge base system, which is used to provide more accurate reply information for question-answering requests. The reply information to be stored in the knowledge base system represented by the storage task can be an explanation of the meaning of the problem code, or a solution to the problem. It should be noted that the present disclosure does not limit the format of the reply information to be stored in the knowledge base system represented by the storage task. In other words, the format of the reply information uploaded by the R&D personnel through the question-answering robot is not limited. For example, it can be a document format, or a piece of text content.

[0047] It should be noted that the present disclosure does not limit how to upload reply information through the question-answering robot. For example, the R&D personnel may input reply information on the interface of the question-answering robot through an input device to generate a storage task. The input device may include, for example, a keyboard, a mouse, a microphone, a stylus, etc.; or, the R&D personnel may upload a document through the interface of the question-answering robot, and the document includes reply information, and a storage task is generated based on the document; or, the question-answering robot may receive reply information sent by other terminal devices, and generate a storage task based on the reply information.

[0048] In some possible implementations, after receiving the question-answering request sent by the question-answering robot, the question-answering center system first determines whether the storage task is received in the question-answering center system. For example, it can determine whether there is a storage task in the cache of the question-answering center system. If it is determined that there is a storage task in the cache, it is determined that the storage task has been received. Specifically, the question-answering center system can determine whether there is a storage task by identifying the task type. For example, the storage task can be set to type 1, and the question-answering request can be set to type 2. Whether the storage task is received is determined by the task type.

[0049] If it is determined that the storage task has been received, step S202 is executed; optionally, if it is determined that the storage task has not been received, the question-answering request is directly sent to the knowledge base system.

[0050] S202: After storing the storage task in the knowledge base system, the question answering request is sent to the knowledge base system.

[0051] Exemplarily, when the question answering center system determines that a storage task has been received, the storage task is first sent to the knowledge base system so that the knowledge base system stores the reply information represented by the storage task, and then the question answering request is sent to the knowledge base system.

[0052] The timing of sending the question answering request may be after a preset time interval after the storage task is sent, or after the question answering center system receives feedback from the knowledge base system that the storage task has been executed.

[0053] It should be noted that the present disclosure does not limit how the knowledge base system stores the storage task. For example, a knowledge graph may be generated based on the reply information represented by the storage task, and the reply information may be stored in the form of a knowledge graph; or a knowledge tree may be generated based on the reply information represented by the storage task, and the reply information may be stored in the form of a knowledge tree.

[0054] In this way, the Q&A center system can ensure that after R&D personnel upload storage tasks, they first store the reply information represented by the storage tasks in the knowledge base system, so as to provide a more comprehensive data source for user-triggered Q&A requests and improve the accuracy of the reply information.

[0055] Optionally, the question answering center system can perform semantic vector conversion processing on the question answering request to obtain a question answering request in a vector format, and send the question answering request in a vector format to the knowledge base system. It should be noted that the present disclosure does not limit the implementation method of the semantic vector conversion processing. For example, the question answering center system can implement the semantic vector conversion processing of the question answering request based on the preset bidirectional encoder from Transformer (Bidirectional Encoder Representations from Transformers, BERT) to obtain a question answering request in a vector format.

[0056] In this way, the question-answering center system can realize vectorized processing of question-answering requests and send the question-answering requests in vector format to the knowledge base system, thereby reducing the processing pressure of the knowledge base system on the question-answering requests and enabling the knowledge base system to directly process the question-answering requests in vector format, thereby improving the processing efficiency of the knowledge base system on the question-answering requests.

[0057] S203: receiving reply information fed back by the knowledge base system, and feeding back the reply information fed back by the knowledge base system to the user through the question answering robot.

[0058] Exemplarily, the reply information fed back by the knowledge base system represents the answer to the question and answer request, and the content of the reply information fed back by the knowledge base system corresponds to the question and answer request, for example, it may be an explanation of the meaning of the question code, and / or the solution to the problem, which is specifically related to the content of the question information indicated in the question and answer request.

[0059] In some possible implementations, after the question-answering center system sends the question-answering request to the knowledge base system, it can receive reply information fed back by the knowledge base system, and then send the reply information fed back by the knowledge base system to the question-answering robot so that the reply information fed back by the knowledge base system can be displayed on the interface of the question-answering robot.

[0060] It should be noted that the present disclosure does not limit how the knowledge base system processes the question and answer request and obtains reply information corresponding to the question and answer request, or in other words, obtains the answer to the question and answer request. For example, the knowledge base system can call the API of the large language model, and process the question information indicated by the question and answer request based on the large language model, so that the large language model can obtain the reply information corresponding to the question and answer request based on the question information and the reply information stored in the knowledge base system, and return the reply information to the question and answer center system.

[0061] Alternatively, the knowledge base system determines the similarity between the question information indicated by the question answering request and the reply information pre-stored in the knowledge base system, and uses the reply information with the highest similarity value as the reply information corresponding to the question answering request, and returns the reply information to the question answering center system. Alternatively, after receiving the question answering request, the knowledge base system may filter the question information indicated by the question answering request for sensitive words, and then process the filtered question information to obtain the reply information. The method of filtering sensitive words is not limited in this disclosure.

[0062] Optionally, after obtaining the reply information corresponding to the question and answer request, the knowledge base system can convert the reply information into a semantic feature vector, and convert the question information indicated by the question and answer request into a semantic feature vector, and determine the similarity between the two semantic feature vectors. If the similarity is determined to be greater than or equal to a second preset threshold, it means that the reply information corresponding to the question and answer request is related to the question and answer request, and it is sent to the question and answer center system.

[0063] The automatic question-answering method based on human-computer interaction provided in this embodiment can, when receiving a question-answering request sent by a question-answering robot, give priority to determining whether there is a received storage task. When there is a received storage task, it will give priority to forwarding the storage task to the knowledge base system for processing, and then forward the question-answering request to the knowledge base system for processing, and receive the reply information returned by the knowledge base system, and feed it back to the user through the question-answering robot. In this way, the question-answering knowledge base system can control the knowledge base system to give priority to the processing of storage tasks, provide a more comprehensive and rich data source for the processing of subsequent question-answering requests, thereby improving the accuracy of the reply information corresponding to the question-answering request and improving the user experience.

[0064] Figure 3 : is a flowchart of an automatic question answering method based on human-computer interaction provided by the second embodiment of the present disclosure. The execution subject of this embodiment may be the aforementioned question answering center system. This embodiment, based on the aforementioned embodiment, describes in detail how to automatically answer questions. The method may include the following steps:

[0065] S301: In response to a question answering request triggered by a user through a question answering robot, determine whether a storage task has been received.

[0066] It should be noted that this step is similar to the aforementioned step S201 and will not be repeated here.

[0067] If it is determined that the question answering center system has received the storage task, step S302 is executed; if it is determined that the storage task has not been received, step S304 is executed.

[0068] S302: Verify the reliability of the storage task.

[0069] Exemplarily, reliability verification refers to verifying the reliability of the reply information represented by the storage task to be stored in the knowledge base system, so as to ensure the security and integrity of the reply information to be stored in the knowledge base system, prevent R&D personnel from uploading erroneous reply information, or prevent the reply information from being maliciously tampered with or lost during transmission.

[0070] In some possible implementations, reliability verification methods may include one or more of the following: format verification of reply information representing the storage task, sensitive word verification of reply information representing the storage task, content relevance verification of reply information representing the storage task, and integrity verification of reply information representing the storage task.

[0071] Among them, the format verification of the reply information represented by the storage task can be to determine whether the format of the reply information meets the preset requirements. The format can refer to the format type of the document. For example, the format of the document of the preset reply information can be in .md or .txt format. The format can also be the text format of the reply information. For example, the text format of the preset reply information is in the form of a key-value pair. It is understandable that the format of the reply information may affect the security of the knowledge base system. For example, if the format of the reply information is an executable script file, it may cause malicious attacks on the knowledge base system and affect the security of the knowledge base system.

[0072] It should be noted that the present disclosure only illustrates the format of the reply information and does not limit the format of the reply information, which can be set according to actual needs.

[0073] The sensitive word verification of the reply information of the storage task representation may be to determine whether there is a sensitive word in the reply information, and the sensitive word represents a word or character that is not allowed to appear in the reply information. The present disclosure does not limit the content of the sensitive word. For example, a sensitive word library may be preset in the question answering center system, and then based on the sensitive word library, it may be determined whether there is a sensitive word in the sensitive word library in the reply information, thereby realizing the sensitive word verification of the reply information of the storage task representation.

[0074] Verifying the content relevance of the reply information representing the stored task may be to determine whether the content of the reply information is correct, in other words, to determine whether the developer has uploaded erroneous reply information, for example, uploading other documents unrelated to the reply information.

[0075] In one possible implementation, when uploading reply information, the R&D personnel need to select the information category of the reply information. The information category can indicate the business type of the question corresponding to the reply information, such as a search-related question, a database-related question, etc. A classification model can be set in the question-answering center system to classify the reply information represented by the storage task. For example, the reply information represented by the storage task is input into a preset classification model, and the category of the reply information represented by the storage task is output; the relevance is determined based on the category of the reply information represented by the storage task and the information category; if it is determined that the relevance is greater than or equal to the first preset threshold, it is determined that the content of the reply information represented by the storage task is related to the information category. Among them, the relevance is used to represent the degree of relevance between the content of the reply information represented by the storage task and the information category.

[0076] Specifically, when determining the category of reply information represented by the storage task and the relevance of the information category, the semantic feature vectors of the two can be determined respectively, and then the cosine similarity of the semantic feature vectors of the two can be calculated to obtain the relevance. It should be noted that the present disclosure does not limit the value of the first preset threshold, which can be set according to actual needs.

[0077] In another possible implementation, based on the aforementioned implementation, detection models of different information categories can be set in the question-answering center system, and the detection model is used to determine the relevance between the reply information represented by the storage task and the information category. For example, the question-answering center system can determine the detection model corresponding to the storage task based on the information category of the reply information represented by the storage task, and then input the reply information represented by the storage task into the detection model, and output the relevance between the reply information represented by the storage task and the information category. If it is determined that the relevance is greater than or equal to the first preset threshold, it is determined that the content of the reply information represented by the storage task is related to the information category.

[0078] It should be noted that the present disclosure does not limit the types of the above-mentioned classification models or detection models. For example, it can be a convolutional neural network model or other machine learning models that can achieve corresponding functions.

[0079] In this way, the question-answering center system can determine the correlation between the reply information represented by the storage task and the information category to be stored in the knowledge base system, ensure the accuracy of the reply information, prevent R&D personnel from uploading erroneous reply information, and ensure the accuracy of the data source in the knowledge base system.

[0080] The integrity verification of the reply information represented by the storage task can be to determine whether the reply information represented by the received storage task is complete, in other words, to determine whether the storage task has been lost or maliciously tampered with during the transmission process. For example, a check frame can be set at the end of the message of the storage task. If the check frame is received, it means that no packet loss has occurred during the transmission process, that is, the reply information is complete; or the preset field value in the storage task represents the hash value of the reply information. When the question-answering system receives the storage task, the hash value of the reply information can be calculated and compared with the preset hash value in the storage task to determine whether the reply information is complete to prevent packet loss or malicious tampering.

[0081] It should be noted that the present disclosure is only an illustration of a reliability verification method. In specific implementation, other reliability verification methods may also be included, and the present disclosure will not exemplify them one by one here.

[0082] In some possible implementations, the question answering center system may perform one or more of the above reliability verification methods on the reply information represented by all storage tasks to verify the reliability of the storage tasks.

[0083] In some possible implementations, as mentioned above, the type of the reply information may be in a document format or in a text content format, and the question answering system may also perform different reliability verification methods according to the type of reply information represented by the storage task.

[0084] For example, the reliability verification methods for storage tasks in document format include one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, and content relevance verification of the reply information represented by the storage task. This is because when the reply information is uploaded in document format, the possibility of incompleteness is relatively small, so the integrity verification process can be reduced.

[0085] The reliability verification methods for non-document format storage tasks include one or more of the following: format verification of the response information represented by the storage task, sensitive word verification of the response information represented by the storage task, and integrity verification of the response information represented by the storage task. This is because when the response information is uploaded in a non-document format, researchers can see the content, so the possibility of irrelevant content is relatively small, so the verification process of content relevance can be reduced.

[0086] In this way, the question-answering center system can determine the reliability verification method based on the type of reply information, and adopt different reliability verification methods for different types of reply information, saving unnecessary verification processes, saving reliability verification time, and improving the execution efficiency of storage tasks.

[0087] If it is determined that the reliability verification of the storage task has passed, step S303 is executed; optionally, if it is determined that the reliability verification has not passed, a prompt message can be returned to the question-answering robot to prompt the user of the problem with the reply message, such as returning a prompt message of an incorrect format to the question-answering robot to prompt the R&D personnel to upload the reply message in the correct format, or returning a prompt message of irrelevant content to the question-answering robot to prompt the R&D personnel to upload the reply message with the correct content.

[0088] S303: If it is determined that the reliability verification of the storage task is passed, the storage task is stored in the knowledge base system.

[0089] It should be noted that this step is similar to the aforementioned step S202 and will not be repeated here.

[0090] In some possible implementations, when the R&D personnel upload the reply information through the Q&A robot, they can set the priority of the reply information, thereby generating a storage task, which has a priority, and set a second message queue in the Q&A center system for caching the storage task. It should be noted that the type of the second message queue is not limited in this disclosure.

[0091] In this implementation, the question-answering center system can determine how to handle the storage task based on the priority of the storage task. For example, priority range 1 can be set as high priority, and priority range 1 can be set as low priority. For example, a priority greater than priority 2 can be set as high priority, and a priority less than priority 2 can be set as low priority. If the storage task is determined to be a high priority, the storage task can be directly sent to the knowledge base system for processing. If the storage task is determined to be a low priority, the storage task can be cached in the second message queue, and the storage tasks can be retrieved from the second message queue in sequence and sent to the knowledge base system for storage.

[0092] In this way, the question-answering center system can control the flow of storage tasks based on the message queue to ensure that the knowledge base system has the ability to process storage tasks. At the same time, based on the priority of the storage tasks, it ensures that high-priority storage tasks can be processed first so that they can be stored in the knowledge base system in a timely manner, ensuring the accuracy of the reply information returned to the user and improving the user experience.

[0093] S304: retrieve the highest priority question-answering request from the first message queue, and send the retrieved question-answering request to the knowledge base system.

[0094] Exemplarily, a first message queue may be set in the question answering center system to cache the question answering request. The type of the first message queue is not limited in this disclosure. After receiving the question answering request, the question answering center system stores the question answering request in the first message queue.

[0095] In some possible implementations, after receiving a question-answering request sent by the question-answering robot, the question-answering center system may set a priority for the question-answering request based on the user who logs into the question-answering robot or the category of the question-answering request.

[0096] For example, if the user who logs in to the Q&A robot is a common user, the Q&A request sent by the Q&A robot can be set to priority 1. If the user who logs in to the Q&A robot is an administrator, the Q&A request sent by the Q&A robot can be set to priority 2. Priority 2 has a higher priority than priority 1. This is because the administrator user's Q&A request is to confirm whether there is a reply information corresponding to the Q&A request in the knowledge base system, or to determine whether to update the reply information corresponding to the Q&A request in order to improve the accuracy of the reply information and enhance the user experience. Therefore, the Q&A request triggered by the administrator user needs to be processed first.

[0097] For another example, when a user inputs question information through a Q&A robot, the user can set the category of the question information, or the Q&A robot can assign categories to the question information based on the question information input by the user, and then the Q&A robot can generate categories of Q&A requests based on the categories of the question information. The category of the Q&A request can be used to describe the category of the question information, such as search-related questions, database-related questions, system control-related questions, etc. After the Q&A center system receives the Q&A request, it can assign a corresponding priority to the Q&A request based on the category of the Q&A request, such as setting Q&A requests for system control-related questions to priority 2 and other types of Q&A requests to priority 1. This is to take into account that if a system-related problem occurs, it can be handled with priority to prevent the search system of the user's terminal device from being unusable, thereby reducing the user experience.

[0098] The Q&A center system can determine the highest priority of all cached Q&A requests from the first message queue, retrieve the highest priority Q&A request and send it to the knowledge base system; if there are multiple highest priority Q&A requests, the highest priority Q&A request at the front will be sent according to the order of the Q&A requests in the first message queue. If the priority of all Q&A requests is the same, they will be sent to the knowledge base system in sequence based on the order in the first message queue.

[0099] Through the above method, the question-answering system can cache question-answering requests based on the first message queue to implement flow control of question-answering requests and prevent excessive requests in the knowledge base system from causing congestion; at the same time, the priority of question-answering requests is used to implement a priority processing mechanism for question-answering requests to improve user experience.

[0100] S305: Receive reply information fed back by the knowledge base system.

[0101] It should be noted that this step is similar to the aforementioned step S203 and will not be repeated here.

[0102] S306: If it is determined that the similarity between the reply information fed back by the knowledge base system and the question-answering request is greater than or equal to a second preset threshold, the reply information fed back by the knowledge base system is fed back to the user via the question-answering robot.

[0103] Exemplarily, after receiving the reply information returned by the knowledge base, the question answering center system can convert the reply information into a semantic feature vector, and convert the question information indicated by the question answering request into a semantic feature vector, determine the similarity between the two semantic feature vectors, and if the similarity is determined to be greater than or equal to the second preset threshold, indicating that the reply information returned by the knowledge base is related to the question answering request, the reply information fed back by the knowledge base system is fed back to the user through the question answering robot. It should be noted that the present disclosure does not limit the value of the second preset threshold, and it can be set according to actual needs.

[0104] Optionally, if it is determined that the similarity is less than the second preset threshold, indicating that the reply information returned by the knowledge base may contain hallucinations, the reply information fed back by the knowledge base system can be fed back to the user through the Q&A robot, and a prompt message can be output to the user to prompt the user that the reply information may contain hallucinations. Alternatively, the Q&A robot can directly return a prompt message to the user that the Q&A request cannot be answered, and output the Q&A request to the R&D personnel's Q&A robot, so that the R&D personnel can promptly supplement the reply information to the knowledge base system based on the Q&A request.

[0105] Optionally, the question answering center system may further perform a sensitive word verification step on the reply information returned by the knowledge base to determine whether sensitive words exist in the reply information.

[0106] Through this step, the Q&A center system can review the response information fed back by the knowledge base to ensure the accuracy of the response information returned to the Q&A robot.

[0107] The automatic question-answering method based on human-computer interaction provided in this embodiment can verify the reliability of the storage task before forwarding it to the knowledge base system, store the storage task that has passed the reliability verification in the knowledge base system, and retrieve the highest priority question-answering request from the first message queue and send it to the knowledge base system for processing. After receiving the reply information returned by the knowledge base system, the reply information and the question-answering request are judged for similarity to determine the accuracy of the reply information. In this way, the question-answering center system can not only ensure the security and integrity of the reply information stored in the knowledge base system, but also use the message queue to implement flow control of the question-answering request, ensuring that the knowledge base system has the processing capacity to process the question-answering request. At the same time, the reply information returned by the knowledge base system can also be confirmed to ensure the accuracy of the reply information sent to the user and improve the user experience.

[0108] Figure 4 400 is a structural block diagram of an automatic question answering device based on human-computer interaction provided by an embodiment of the present disclosure. The device 400 includes the following units:

[0109] Processing unit 401 is used to respond to a question-answering request triggered by a user through a question-answering robot. If it is determined that a storage task has been received, after storing the storage task in the knowledge base system, the question-answering request is sent to the knowledge base system; wherein the storage task represents reply information to be stored in the knowledge base system; the question-answering request is used to determine the reply information corresponding to the question-answering request from the knowledge base system.

[0110] The forwarding unit 402 is used to receive the reply information fed back by the knowledge base system, and feed back the reply information fed back by the knowledge base system to the user through the question answering robot.

[0111] In some possible implementations, the question answering request has a priority, and the question answering request is cached in the first message queue; the processing unit 401 includes:

[0112] The sending module is used to retrieve the highest priority question-answering request from the first message queue, and send the retrieved question-answering request to the knowledge base system.

[0113] In some possible implementations, the priority of the question answering request corresponds to the user, or the priority of the question answering request corresponds to the category of the question answering request.

[0114] In some possible implementations, the processing unit 401 includes:

[0115] The processing module is used to perform semantic vector conversion processing on the question-answering request to obtain the question-answering request in vector format.

[0116] In some possible implementations, the storage task has a priority; the processing unit 401 includes:

[0117] A first storage module, used for sending a high-priority storage task to the knowledge base system for storage;

[0118] The second storage module is used to cache the low-priority storage tasks in the second message queue, retrieve the storage tasks from the second message queue, and store the retrieved storage tasks in the knowledge base system.

[0119] In some possible implementations, the processing unit 401 further includes:

[0120] A verification module is used to verify the reliability of the storage task; wherein the reliability verification method includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, content relevance verification of the reply information represented by the storage task, and integrity verification of the reply information represented by the storage task; if it is determined that the reliability verification of the storage task is passed, the step of storing the storage task in the knowledge base system is executed.

[0121] In some possible implementations, when the storage task is a storage task in a document format, the reliability verification method for the storage task in the document format includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, and content relevance verification of the reply information represented by the storage task;

[0122] When the storage task is a storage task in a non-document format, the reliability verification method for the storage task in a non-document format includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, and integrity verification of the reply information represented by the storage task.

[0123] In some possible implementations, the storage task has an information category; the verification module includes:

[0124] A processing submodule, used for inputting the reply information represented by the stored task into a preset classification model, and outputting the category of the reply information represented by the stored task;

[0125] A first determination submodule is used to determine the relevance according to the category of the reply information represented by the storage task and the information category; wherein the relevance is used to represent the degree of relevance between the content of the reply information represented by the storage task and the information category;

[0126] The second determining submodule is configured to determine that the content of the reply information represented by the stored task is related to the information category when it is determined that the correlation is greater than or equal to a first preset threshold.

[0127] In some possible implementations, the forwarding unit 402 includes:

[0128] The determination module is used to feed back the reply information fed back by the knowledge base system to the user through the question-answering robot when it is determined that the similarity between the reply information fed back by the knowledge base system and the question-answering request is greater than or equal to a second preset threshold.

[0129] 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.

[0130] 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.

[0131] Figure 5 It is a block diagram of an electronic device for implementing the automatic question-answering method based on human-computer interaction of an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, 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.

[0132] like Figure 5 As shown, the device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0133] A number of components in the device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0134] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 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 501 performs the various methods and processes described above, such as the method for automatic question answering based on human-computer interaction. For example, in some embodiments, the method for automatic question answering based on human-computer interaction may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the method for automatic question answering based on human-computer interaction described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to execute the method of automatic question answering based on human-computer interaction in any other appropriate manner (for example, by means of firmware).

[0135] 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.

[0136] 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.

[0137] 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.

[0138] 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).

[0139] 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.

[0140] 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.

[0141] 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.

[0142] 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. An automatic question answering method based on human-computer interaction, the method comprising: In response to a question-answering request triggered by a user through a question-answering robot, if it is determined that a storage task has been received, after storing the storage task in a knowledge base system, the question-answering request is sent to the knowledge base system; wherein the storage task represents the reply information to be stored in the knowledge base system; the question-answering request is used to determine the reply information corresponding to the question-answering request from the knowledge base system; Receive the reply information fed back by the knowledge base system, and feed back the reply information fed back by the knowledge base system to the user through the question answering robot.

2. The method according to claim 1, wherein: The question-answering request has a priority, and the question-answering request is cached in the first message queue; Sending the question-answering request to the knowledge base system includes: The question-answering request with the highest priority is retrieved from the first message queue, and the retrieved question-answering request is sent to the knowledge base system.

3. The method according to claim 2, wherein: The priority of the question answering request corresponds to the user, or the priority of the question answering request corresponds to the category of the question answering request.

4. The method according to any one of claims 1 to 3, wherein: Before sending the question-answering request to the knowledge base system, the method further includes: The question-answering request is subjected to semantic vector conversion processing to obtain a question-answering request in a vector format.

5. The method according to any one of claims 1 to 4, wherein: The storage task has a priority; Storing the storage task in a knowledge base system includes: Sending high-priority storage tasks to the knowledge base system for storage; The low-priority storage tasks are cached in the second message queue, and the storage tasks are retrieved from the second message queue, and the retrieved storage tasks are stored in the knowledge base system.

6. The method according to any one of claims 1 to 5, wherein: Before storing the storage task in the knowledge base system, the method further includes: Performing reliability verification on the storage task; wherein the reliability verification method includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, content relevance verification of the reply information represented by the storage task, and integrity verification of the reply information represented by the storage task; If it is determined that the reliability verification of the storage task is passed, the step of storing the storage task in the knowledge base system is performed.

7. The method according to claim 6, wherein: When the storage task is a storage task in a document format, the reliability verification method for the storage task in the document format includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, and content relevance verification of the reply information represented by the storage task; When the storage task is a storage task in a non-document format, the reliability verification method for the storage task in a non-document format includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, and integrity verification of the reply information represented by the storage task.

8. The method according to claim 6 or 7, wherein: The storage task has an information category; when the reliability verification method is content relevance verification, reliability verification is performed on the storage task, including: Inputting the reply information represented by the stored task into a preset classification model, and outputting the category of the reply information represented by the stored task; Determining the relevance according to the category of the reply information represented by the storage task and the information category; wherein the relevance is used to represent the degree of relevance between the content of the reply information represented by the storage task and the information category; If it is determined that the relevance is greater than or equal to the first preset threshold, it is determined that the content of the reply information represented by the stored task is related to the information category.

9. The method according to any one of claims 1 to 8, wherein: Receiving reply information fed back by the knowledge base system, and feeding back the reply information fed back by the knowledge base system to the user through the question answering robot, including: If it is determined that the similarity between the reply information fed back by the knowledge base system and the question-answering request is greater than or equal to a second preset threshold, the reply information fed back by the knowledge base system is fed back to the user through the question-answering robot.

10. An automatic question-answering device based on human-computer interaction, comprising: A processing unit, configured to respond to a question-answering request triggered by a user through a question-answering robot, and if it is determined that a storage task has been received, after storing the storage task in a knowledge base system, send the question-answering request to the knowledge base system; wherein the storage task represents reply information to be stored in the knowledge base system; and the question-answering request is used to determine reply information corresponding to the question-answering request from the knowledge base system; The forwarding unit is used to receive the reply information fed back by the knowledge base system, and feed back the reply information fed back by the knowledge base system to the user through the question-answering robot.

11. The device according to claim 10, wherein: The question-answering request has a priority, and the question-answering request is cached in the first message queue; The processing unit comprises: The sending module is used to retrieve the highest priority question-answering request from the first message queue, and send the retrieved question-answering request to the knowledge base system.

12. The device according to claim 11, wherein The priority of the question answering request corresponds to the user, or the priority of the question answering request corresponds to the category of the question answering request.

13. The device according to any one of claims 10 to 12, wherein: The processing unit comprises: The processing module is used to perform semantic vector conversion processing on the question-answering request to obtain the question-answering request in vector format.

14. The device according to any one of claims 10 to 13, wherein: The storage task has a priority; the processing unit includes: A first storage module, used for sending a high-priority storage task to the knowledge base system for storage; The second storage module is used to cache the low-priority storage tasks in the second message queue, retrieve the storage tasks from the second message queue, and store the retrieved storage tasks in the knowledge base system.

15. The device according to any one of claims 10 to 14, wherein: The processing unit further includes: A verification module is used to verify the reliability of the storage task; wherein the reliability verification method includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, content relevance verification of the reply information represented by the storage task, and integrity verification of the reply information represented by the storage task; if it is determined that the reliability verification of the storage task is passed, the step of storing the storage task in the knowledge base system is executed.

16. The device according to claim 15, wherein: When the storage task is a storage task in a document format, the reliability verification method for the storage task in the document format includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, and content relevance verification of the reply information represented by the storage task; When the storage task is a storage task in a non-document format, the reliability verification method for the storage task in a non-document format includes one or more of the following: format verification of the reply information represented by the storage task, sensitive word verification of the reply information represented by the storage task, and integrity verification of the reply information represented by the storage task.

17. The device according to claim 15 or 16, wherein: The storage task has an information category; the verification module includes: A processing submodule, used for inputting the reply information represented by the stored task into a preset classification model, and outputting the category of the reply information represented by the stored task; A first determination submodule is used to determine the relevance according to the category of the reply information represented by the storage task and the information category; wherein the relevance is used to represent the degree of relevance between the content of the reply information represented by the storage task and the information category; The second determining submodule is configured to determine that the content of the reply information represented by the stored task is related to the information category when it is determined that the correlation is greater than or equal to a first preset threshold.

18. The device according to any one of claims 10 to 17, wherein: The forwarding unit comprises: The determination module is used to feed back the reply information fed back by the knowledge base system to the user through the question-answering robot when it is determined that the similarity between the reply information fed back by the knowledge base system and the question-answering request is greater than or equal to a second preset threshold.

19. 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 9.

20. 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 1-9.

21. A computer program product, comprising a computer program, which implements the steps of the method according to any one of claims 1 to 9 when executed by a processor.

22. A question-answering platform, comprising: Question answering center system; as well as, A question-answering robot and a knowledge base system connected to the question-answering center system; wherein the question-answering center system is capable of executing the method described in any one of claims 1-9.

Images