Question and answer processing method and device based on large model and knowledge base, equipment and medium

Abstract

The embodiment of the invention discloses a question and answer processing method and device based on a large model and a knowledge base, equipment and a medium, and relates to the technical field of computers, the method comprises the steps of obtaining historical chat records in response to input information of a user; inputting the historical chat record into the large model, and extracting a question text through the large model according to the historical chat record; traversing in a preset knowledge base based on the question text to obtain answer content related to the question text; inputting the answer content into the large model, and generating an answer text according to the answer content through the large model; and generating answer information fed back to the user based on the answer text. According to the method and the device, an administrator can be helped to understand customer problems, a language for replying customers can be organized after knowledge associated with the problems is retrieved, the pressure of replying messages by the administrator can be reduced, customer service personnel can be helped to reply domain-crossing knowledge problems, technical support can be provided for customer service, and the working efficiency of customer service personnel can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a question-answering processing method, device, equipment and medium based on a large model and a knowledge base. Background Art

[0002] At present, with the continuous development of artificial intelligence technology, intelligent customer service question and answer systems are widely used in practical applications. The intelligent customer service question and answer systems can effectively improve the service efficiency for users.

[0003] However, with the expansion and development of business, users' demand for problem consultation is increasing. There are situations where the number of users is large, the types of questions are many, and the questions involve many fields. In these cases, the current intelligent customer service question and answer system can usually only answer questions based on rules, templates or simple statistical models. It is limited by semantic understanding ability and diverse expressions. When faced with complex problems, it is difficult to accurately understand the user's intentions. The generated answer text is easy to deviate from the user's needs. The generated text may be relatively stiff, which leads to the user's inability to accurately get the answers to their needs, and the user experience is low. Summary of the invention

[0004] The embodiment of the present application provides a question-answering processing method, device, equipment and medium based on a large model and a knowledge base, so as to solve the defects of the above-mentioned related technologies. The technical solution is as follows:

[0005] In a first aspect, an embodiment of the present application provides a question-answering processing method based on a large model and a knowledge base, which is applied to a client end, and the method includes:

[0006] Responding to the user's input information, obtaining historical chat records;

[0007] Inputting the historical chat records into a large model, and extracting question texts from the historical chat records through the large model;

[0008] Traversing a preset knowledge base based on the question text to obtain answer content related to the question text;

[0009] Inputting the answer content into the large model, and generating an answer text according to the answer content through the large model;

[0010] Generate answer information to be fed back to the user based on the answer text.

[0011] In an optional solution of the first aspect, the acquiring of historical chat records in response to user input information includes:

[0012] In response to input information from a user, determining a chat time and a user ID of the user based on the input information;

[0013] The input information is stored in a preset database table, and a plurality of historical messages corresponding to the user number before the chat time are traversed from the preset database table;

[0014] The historical chat record is obtained by combining the input information and the multiple historical messages.

[0015] In an optional solution of the first aspect, inputting the historical chat records into a large model, and extracting question texts according to the historical chat records by using the large model, comprises:

[0016] Embedding the historical chat record into a preset first prompt template to construct a first prompt content;

[0017] The first prompt content is input into the big model, so that the big model performs semantic analysis based on the first prompt content to extract the question text.

[0018] In an optional solution of the first aspect, after extracting the question text, the method further includes:

[0019] Processing the question text into a question vector, and updating the question vector into a question field of a preset database table;

[0020] The step of traversing a preset knowledge base based on the question text to obtain answer content related to the question text includes:

[0021] Traversing a preset first knowledge base based on the question vector, and in the case where an answer vector corresponding to the question vector is traversed based on a mapping relationship between the question vector and the answer vector, processing the answer vector to obtain an answer related to the question text, thereby obtaining the answer content;

[0022] Among them, the first knowledge base is constructed based on multiple preset question-answer pairs, and the question-answer pairs are stored in the first knowledge base in the form of question-answer vector pairs, and each question vector in the vector pair has a mapping relationship with an answer vector.

[0023] In an optional solution of the first aspect, after traversing the preset first knowledge base based on the question vector, if no answer vector is obtained through the traversal, the method further includes:

[0024] Traversing a preset second knowledge base based on the question vector, traversing matching knowledge data including each feature word and / or a combination of a preset number of feature words in the question vector from the second knowledge base, to obtain answer content related to the question text;

[0025] The second knowledge base is constructed based on knowledge data in multiple fields and feature words corresponding to the knowledge data.

[0026] In an optional solution of the first aspect, inputting the answer content into the large model, and generating an answer text according to the answer content by using the large model, includes:

[0027] Embedding the answer content into a preset second prompt template to construct a second prompt content;

[0028] Inputting the second prompt content into the large model so that the large model performs semantic analysis on the second prompt content and extracts answer text features for answering the question text;

[0029] The large model is configured to generate the answer text according to the answer format determined by the second prompt template and the answer text features.

[0030] In an optional solution of the first aspect, when no answer vector is obtained through traversal, after generating answer information fed back to the user based on the answer text, the method further includes:

[0031] Based on the correspondence between the question text and the answer information, a new question-answer pair is constructed;

[0032] The question-answer pair is vectorized into a question-answer vector pair, and the question-answer vector pair is updated into the first knowledge base.

[0033] In a second aspect, the embodiment of the present application further provides a question-answering processing device based on a large model and a knowledge base, comprising:

[0034] An information acquisition module, used to acquire historical chat records in response to user input information;

[0035] A question text extraction module, used for inputting the historical chat records into a large model, and extracting question texts from the historical chat records through the large model;

[0036] An answer content extraction module is used to traverse a preset knowledge base based on the question text to obtain answer content related to the question text;

[0037] An answer text generation module, used for inputting the answer content into the large model, and generating an answer text according to the answer content through the large model;

[0038] The answer text feedback module is used to generate answer information fed back to the user based on the answer text.

[0039] In a third aspect, an embodiment of the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method provided in the first aspect of the embodiment of the present application or any one of the implementation methods of the first aspect is implemented.

[0040] In a fourth aspect, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method provided by the first aspect of the embodiment of the present application or any one of the implementations of the first aspect.

[0041] The beneficial effects brought about by the technical solutions provided by some embodiments of the present application include at least:

[0042] The embodiment of the present application provides a question-answering processing method, device, equipment and medium based on a large model and a knowledge base, which can pull chat records based on user input information, thereby retaining a more complete context based on historical chat records, so that the large model can more accurately understand the user's intentions and extract accurate question texts. On the other hand, the use of a knowledge base can narrow the search scope of questions and locate the answer content corresponding to the question text more quickly. Furthermore, by organizing the language through a large language model, the generated answer text is provided to the administrator for reference, which can effectively reduce the administrator's knowledge retrieval and information input pressure. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the present application or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 It is a schematic diagram of the architecture of a question-answering processing system based on a large model and a knowledge base provided in an embodiment of the present application;

[0045] Figure 2 It is a flowchart of a question-answering processing method based on a large model and a knowledge base provided in an embodiment of the present application;

[0046] Figure 3 It is a structural schematic diagram of a question-answering processing device based on a large model and a knowledge base provided in an embodiment of the present application;

[0047] Figure 4 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] The terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or modules is not limited to the listed steps or modules, but may optionally include steps or modules that are not listed, or may optionally include other steps or modules that are inherent to these processes, methods, products or devices.

[0050] It should be noted that the terms "first\second" involved in the present application are only used to distinguish similar objects, and do not represent a specific order for the objects. It is understandable that "first\second" can be interchanged with a specific order or sequence where permitted. It should be understood that the objects distinguished by "first\second" can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those described or illustrated herein.

[0051] Please refer to the following Figure 1 , which is a schematic diagram of the architecture of a question-answering processing system based on a large model and a knowledge base provided by an exemplary embodiment of the present application. Figure 1 As shown, the system includes a client terminal 101 and a user terminal 102.

[0052] Among them, the customer service personnel can receive the user's input information on the customer service terminal 101, and further receive the historical chat records determined by the system based on the input information, and can input the chat records into the big model, extract the question text based on the historical chat records through the big model, traverse the preset knowledge base based on the question text to obtain the answer content related to the question text, input the answer content into the big model, and generate the answer text according to the answer content through the big model; the customer service personnel can adjust, polish and other operations on the answer text generated by the big model on the customer service terminal 101, and then can generate answer information to be fed back to the user terminal 102, or can directly use the answer text generated by the big model as the answer information fed back to the user terminal 102.

[0053] Among them, the user can input the questions to be asked on the user terminal 102, and the user terminal 102 can include but is not limited to a mobile terminal, a personal computer, a tablet, etc. The user can operate the software on the user terminal or through the browser installed on the user terminal. After the client generates the answer information, the user can receive the answer information through the user terminal. This embodiment of the application is not limited to this.

[0054] The present application is described in detail below with reference to specific embodiments.

[0055] Next, combine Figure 2 , taking the question-answering processing method based on a large model and a knowledge base executed by the client as an example, the question-answering processing method based on a large model and a knowledge base provided by the embodiment of the present application is introduced. Figure 2 , Figure 2 FIG. 1 is a flow chart of a question-answering processing method based on a large model and a knowledge base provided in an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:

[0056] S201, in response to user input information, obtaining historical chat records;

[0057] S202, inputting the historical chat records into a large model, and extracting question texts from the historical chat records using the large model;

[0058] S203, traversing a preset knowledge base based on the question text to obtain answer content related to the question text;

[0059] S204, inputting the answer content into the large model, and generating an answer text according to the answer content through the large model;

[0060] S205: Generate answer information fed back to the user based on the answer text.

[0061] In some embodiments, in S201, the user can input questions to be asked in the chat window on the user terminal, where the chat window can be a chat window between the user and the customer service staff, or it can be a related customer service group, where multiple users can input information in the customer service group.

[0062] Specifically, each time a user enters new chat information, the user's chat time and user number can be determined based on the input information, and then the input information can be stored in a preset database table, and multiple historical messages corresponding to the user number before the chat time can be traversed from the preset database table. The historical chat record can be obtained by combining the input information and multiple historical messages.

[0063] Specifically, in a group chat scenario, the group chat number may be determined according to the user number, and then the historical chat records of the corresponding group chat may be extracted from a preset database table.

[0064] In some embodiments, in S202, the natural language understanding capability of the large model can be used to extract the text of the question asked by the user from the historical chat records, specifically including:

[0065] Embedding the historical chat record into a preset first prompt template to construct a first prompt content;

[0066] The first prompt content is input into the big model, so that the big model performs semantic analysis based on the first prompt content to extract the question text.

[0067] It should be noted that the prompt templates of Large Language Models (LLMs) are a structured way for users to interact with the model. The prompt template is essentially a text framework that contains some fixed text and some variable placeholders. These placeholders can be replaced with specific content according to specific needs. In this way, the model can be guided to generate expected answers or creations.

[0068] Among them, the first prompt template is a prompt template used to guide the large model to perform text analysis on historical chat records and extract question text.

[0069] For example, taking a group chat scenario as an example, the group chat includes three users A, B, and C. After the historical chat records are embedded in the placeholder corresponding to the group chat, the first prompt content generated may be:

[0070] [{"role":"system",

[0071] "content":"As a professional chat record intelligent analysis assistant in the xxx field, you need to analyze the group chat records (including group chat history and latest records). Based on the group chat history, determine whether the latest record is a meaningful question related to xxx business and xxx activities. If it is a meaningful question related to xxx business and marketing activities, you need to summarize the questions the speaker actually wants to ask, and present the analysis content and the real question in a specified format string.

[0072] The key takeaways are as follows:

[0073] 1. In the specified format, fill the analysis content into the "content" field and the real question into the "question" field.

[0074] 2. The actual question content in the "question" field is only the question obtained through analysis and summary. This content must be used for subsequent secondary searches in the knowledge base and must not be mixed with other content.

[0075] 3. If the latest record is not a meaningful question in the context, the content of the "question" field should be an empty string.

[0076] The output format example is as follows. Strictly follow the output format below and do not add other content when outputting.

[0077]

[0078] It should be noted that the large model can extract the question text field (question) according to the content field (content), key point field, etc. in the first prompt content.

[0079] In some embodiments, after the question text is extracted in S202, the extracted question text will also be saved in a preset database and saved in the question field of a preset database table in the form of a question vector, so it is easy to determine whether the user's new input information is a new question based on the question vector.

[0080] When the extracted question text is a new question, step S203 is performed, including:

[0081] A question vector corresponding to the question text may be traversed in a preset first knowledge base, and when an answer vector corresponding to the question vector is traversed based on a mapping relationship between the question vector and the answer vector, the answer vector is processed to obtain an answer related to the question text, thereby obtaining the answer content;

[0082] Among them, the first knowledge base is constructed based on multiple preset question-answer pairs, and the question-answer pairs are stored in the first knowledge base in the form of question-answer vector pairs, and each question vector in the vector pair has a mapping relationship with an answer vector.

[0083] It can be understood that the first knowledge base is a QA knowledge base, in which common questions and preset answers are prepared. The frequency of user questions can be collected based on the interaction records between users and customer service personnel in historical data, so as to construct question-answer pairs based on questions and corresponding answers. The embodiment of the present application does not limit the construction method of the first knowledge base.

[0084] In this way, the OA knowledge base can be retrieved to accurately match the responses to questions within the QA question set, thereby quickly generating the answer information required by the user.

[0085] In some embodiments, in S203, after traversing the preset first knowledge base based on the question vector, if no answer vector is obtained through traversal, the following steps are further included:

[0086] Traversing a preset second knowledge base based on the question vector, traversing matching knowledge data including each feature word and / or a combination of a preset number of feature words in the question vector from the second knowledge base, to obtain answer content related to the question text;

[0087] The second knowledge base is constructed based on knowledge data in multiple fields and feature words corresponding to the knowledge data.

[0088] It should be noted that, for some questions raised by users, the same question vector is not traversed in the first knowledge base, and the corresponding answer vector cannot be directly extracted from the first knowledge base.

[0089] Specifically, for questions outside the first knowledge base, the second knowledge base can be searched to obtain corpus information related to the question text, that is, knowledge data. The feature words contained in the question vector can be matched with the feature words in the second knowledge base. For example, a combination of all feature words in the question vector can be selected for matching to find corpus information with the same feature word combination in the second knowledge base. A combination of several feature words or a single feature word in the question vector can also be selected for matching to find corpus information with corresponding feature words. The embodiment of the present application is not limited to this.

[0090] It should be noted that the second knowledge base can select document materials covering multiple fields, can be expanded according to business types, can directly link to existing databases or obtain more data through search engine traversal, and the embodiments of the present application are not limited to this.

[0091] In this way, even if customer service personnel are faced with cross-domain knowledge, they can obtain relevant corpus information through the second knowledge base, and then refine the corpus information through a large model, which will help customer service personnel better respond to questions that cross-domain knowledge.

[0092] In some embodiments, in S204, the answer content may be embedded into a preset second prompt template to construct a second prompt content;

[0093] Inputting the second prompt content into the large model so that the large model performs semantic analysis on the second prompt content and extracts answer text features for answering the question text;

[0094] The large model is configured to generate the answer text according to the answer format determined by the second prompt template and the answer text features.

[0095] Among them, the second prompt template is a prompt template input to the large model for instructing the large model to perform semantic analysis according to the second prompt content, so that the large model can extract the answer text features according to the prompt content. Afterwards, the large model will fill the answer text features into the answer template in the second prompt template according to the format defined by the second prompt content to generate the answer text. For example, the number of words in the output answer text, the form of the answer text (such as striped display features), etc. can be specified by the second prompt template, and this embodiment of the present application does not limit this.

[0096] In some embodiments, in S205, after obtaining the answer text generated by the large model, the customer service staff can review the answer text, or continue to perform semantic analysis on the answer text through the large model, so as to make appropriate adjustments to the answer text, such as deleting some words, correcting typos, adjusting the text paragraph structure, etc., which is not limited to this in the embodiments of the present application.

[0097] In some embodiments, after S205 , the first knowledge base may be updated based on the new question text and the generated answer information.

[0098] Specifically, a new question-answer pair may be constructed based on the correspondence between the question text extracted in S202 and the answer information polished by the customer service staff in S205;

[0099] The question-answer pair is vectorized into a question-answer vector pair, and the question-answer vector pair is updated into the first knowledge base.

[0100] In this way, the first knowledge base can be continuously updated according to the user's questions, which is conducive to expanding the coverage of the first knowledge base, making it possible to generate the answers required by the user more quickly, and is conducive to improving the efficiency of generating the answer text.

[0101] The following are device embodiments of the present application, which can be used to execute the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.

[0102] See next Figure 3, which is a structural diagram of a question-answering processing device based on a large model and a knowledge base provided by an exemplary embodiment of the present application. The device can be implemented as all or part of a terminal through software, hardware, or a combination of both, and can also be integrated on a server as an independent module. The question-answering processing device based on a large model and a knowledge base in the embodiment of the present application can be applied to a terminal or a cloud. The device 30 includes an information acquisition module 301, a question text extraction module 302, an answer content extraction module 303, an answer text generation module 304, and an answer text feedback module 305, wherein:

[0103] The information acquisition module 301 is used to obtain historical chat records in response to user input information;

[0104] The question text extraction module 302 is used to input the historical chat records into the big model, and extract the question text according to the historical chat records through the big model;

[0105] The answer content extraction module 303 is used to traverse the preset knowledge base based on the question text to obtain answer content related to the question text;

[0106] The answer text generation module 304 is used to input the answer content into the large model, and generate an answer text according to the answer content through the large model;

[0107] The answer text feedback module 305 is used to generate answer information fed back to the user based on the answer text.

[0108] It should be noted that when the device 30 provided in the above embodiment executes the question-answering processing method based on a large model and a knowledge base, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device provided in the above embodiment and the embodiment of the question-answering processing method based on a large model and a knowledge base belong to the same concept. The implementation process is detailed in the method embodiment and will not be repeated here.

[0109] An embodiment of the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above-mentioned method embodiments when executing the program.

[0110] See also Figure 4 , which is a structural block diagram of an electronic device provided in an embodiment of the present application.

[0111] like Figure 4 As shown, the electronic device 400 includes: a processor 401 and a memory 402 .

[0112] In the embodiment of the present application, the processor 401 is the control center of the computer system, which can be a processor of a physical machine or a processor of a virtual machine. The processor 401 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 401 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array).

[0113] The processor 401 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also called a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state.

[0114] The memory 402 may include one or more computer-readable storage media, which may be non-transitory. The memory 402 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments of the present application, the non-transitory computer-readable storage medium in the memory 402 is used to store at least one instruction, which is used to be executed by the processor 401 to implement the method in the embodiment of the present application.

[0115] In some embodiments, the electronic device 400 further includes: a peripheral device interface 403 and at least one peripheral device 404. The processor 401, the memory 402 and the peripheral device interface 403 can be connected via a bus or a signal line. Each peripheral device 404 can be connected to the peripheral device interface 403 via a bus, a signal line or a circuit board. Specifically, the peripheral device 404 includes: a display screen, a camera and an audio circuit. The peripheral device interface 403 can be used to connect at least one peripheral device related to I / O (Input / Output) to the processor 401 and the memory 402.

[0116] In some embodiments of the present application, the processor 401, the memory 402, and the peripheral device interface 403 are integrated on the same chip or circuit board; in some other embodiments of the present application, any one or two of the processor 401, the memory 402, and the peripheral device interface 403 can be implemented on a separate chip or circuit board. This embodiment of the present application does not specifically limit this.

[0117] The electronic device structure block diagram shown in the embodiment of the present application does not constitute a limitation on the electronic device 400. The electronic device 400 may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.

[0118] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the method of any of the above embodiments are implemented. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a micro drive, and a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0119] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution can be essentially or in other words, the part that contributes to the relevant technology can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A question-answering processing method based on a large model and a knowledge base, characterized in that: Applied to the client side, the method includes: Responding to the user's input information, obtaining historical chat records; Inputting the historical chat records into a large model, and extracting question texts from the historical chat records through the large model; Traversing a preset knowledge base based on the question text to obtain answer content related to the question text; Inputting the answer content into the large model, and generating an answer text according to the answer content through the large model; Generate answer information to be fed back to the user based on the answer text.

2. A question-answering processing method based on a large model and a knowledge base according to claim 1, characterized in that: The step of obtaining the historical chat records in response to the user's input information includes: In response to input information from a user, determining a chat time and a user ID of the user based on the input information; The input information is stored in a preset database table, and a plurality of historical messages corresponding to the user number before the chat time are traversed from the preset database table; The historical chat record is obtained by combining the input information and the multiple historical messages.

3. A question-answering processing method based on a large model and a knowledge base according to claim 1, characterized in that: The step of inputting the historical chat records into a large model and extracting question texts according to the historical chat records by using the large model includes: Embedding the historical chat record into a preset first prompt template to construct a first prompt content; The first prompt content is input into the big model, so that the big model performs semantic analysis based on the first prompt content to extract the question text.

4. A question-answering processing method based on a large model and a knowledge base according to claim 1, characterized in that: After the question text is extracted, the method further includes: Processing the question text into a question vector, and updating the question vector into a question field of a preset database table; The step of traversing a preset knowledge base based on the question text to obtain answer content related to the question text includes: Traversing a preset first knowledge base based on the question vector, and in the case where an answer vector corresponding to the question vector is traversed based on a mapping relationship between the question vector and the answer vector, processing the answer vector to obtain an answer related to the question text, thereby obtaining the answer content; Among them, the first knowledge base is constructed based on multiple preset question-answer pairs, and the question-answer pairs are stored in the first knowledge base in the form of question-answer vector pairs, and each question vector in the vector pair has a mapping relationship with an answer vector.

5. A question-answering processing method based on a large model and a knowledge base according to claim 4, characterized in that: After traversing the preset first knowledge base based on the question vector, if no answer vector is obtained through traversal, the method further includes: Traversing a preset second knowledge base based on the question vector, traversing matching knowledge data including each feature word and / or a combination of a preset number of feature words in the question vector from the second knowledge base, to obtain answer content related to the question text; The second knowledge base is constructed based on knowledge data in multiple fields and feature words corresponding to the knowledge data.

6. A question-answering processing method based on a large model and a knowledge base according to claim 1, characterized in that: The step of inputting the answer content into the large model and generating an answer text according to the answer content through the large model includes: Embedding the answer content into a preset second prompt template to construct a second prompt content; Inputting the second prompt content into the large model so that the large model performs semantic analysis on the second prompt content and extracts answer text features for answering the question text; The large model is configured to generate the answer text according to the answer format determined by the second prompt template and the answer text features.

7. The method according to claim 5, characterized in that In the case where no answer vector is obtained through traversal, after generating answer information fed back to the user based on the answer text, the method further includes: Based on the correspondence between the question text and the answer information, a new question-answer pair is constructed; the question-answer pair is vectorized into a question-answer vector pair, and the question-answer vector pair is updated to the first knowledge base.

8. A question-answering processing device based on a large model and a knowledge base, characterized in that: include: An information acquisition module, used to acquire historical chat records in response to user input information; A question text extraction module, used for inputting the historical chat records into a large model, and extracting question texts from the historical chat records through the large model; An answer content extraction module is used to traverse a preset knowledge base based on the question text to obtain answer content related to the question text; An answer text generation module, used for inputting the answer content into the large model, and generating an answer text according to the answer content through the large model; The answer text feedback module is used to generate answer information fed back to the user based on the answer text.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images