Response method and device, electronic equipment and storage medium

Abstract

The invention provides a reply method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, target maintenance documents are determined by responding to question information for equipment maintenance, and the target maintenance documents comprise at least two types of maintenance documents with the same equipment characteristics; the source databases of the maintenance documents of each type are different; the target maintenance document is processed on the basis of a preset first large language model, domain information is obtained, the domain information comprises at least one of reply strategy information, document format information and equipment information, and the domain information is used for indicating a preset second large language model to process the target maintenance document; and processing the domain information and the target maintenance document based on the second large language model, and determining reply information corresponding to the question information. The equipment maintenance efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a reply method, device, electronic device, and computer-readable storage medium. Background Art

[0002] When equipment fails, equipment maintenance currently still mainly relies on human experience or equipment manuals.

[0003] Maintenance personnel need to judge the cause of equipment failure based on the equipment failure phenomenon and the equipment manual. The entire maintenance process is difficult to standardize and uniformly manage, and is highly dependent on the experience of maintenance personnel, resulting in low equipment maintenance efficiency and failure to effectively eliminate equipment failures. Summary of the invention

[0004] The embodiments of the present application provide a reply method, apparatus, electronic device, and computer-readable storage medium, aiming to improve equipment maintenance efficiency.

[0005] In a first aspect, an embodiment of the present application provides a reply method, the method comprising:

[0006] In response to the question information regarding equipment maintenance, target maintenance documents are determined, wherein the target maintenance documents include at least two types of maintenance documents having the same equipment characteristics, and each type of the maintenance documents has a different source database;

[0007] Processing the target maintenance document based on a preset first language model to obtain domain information, the domain information including at least one of reply strategy information, document format information and device information, the domain information being used to instruct a preset second language model to process the target maintenance document;

[0008] The domain information and the target maintenance document are processed based on the second largest language model to determine answer information corresponding to the question information.

[0009] Optionally, the responding to the inquiry information regarding equipment maintenance and determining a target maintenance document includes:

[0010] According to the maintenance description information in the question information, query a preset local knowledge base to obtain at least one first maintenance document; wherein the maintenance description information includes at least one of a fault code and / or a fault phenomenon description information;

[0011] Determining whether there is any information missing location in the first maintenance document according to the structural information, logical information and / or semantic information of the first maintenance document;

[0012] If there is an information missing location, querying a preset cloud database according to the information missing location and the context corresponding to the information missing location to obtain a second maintenance document;

[0013] According to the information missing position, the first maintenance document and the second maintenance document are information-fused to obtain a target maintenance document.

[0014] Optionally, the step of constructing the local knowledge base includes:

[0015] Obtain maintenance documentation for multiple devices;

[0016] For each device maintenance document, performing layout analysis on the device maintenance document to identify target elements in the device maintenance document;

[0017] According to the target element, the device maintenance document is sliced ​​to obtain a plurality of knowledge slices;

[0018] For the knowledge slices, vectorization processing is performed on the knowledge slices, and the vectorized knowledge slices are imported into a local knowledge base.

[0019] Optionally, the training step of the second largest language model includes:

[0020] Acquire question training information, and extract maintenance training documents corresponding to the question training information from the local knowledge base;

[0021] Inputting the domain training information input by the user, the question training information and the maintenance training document into the initial large language model to obtain answer prediction information;

[0022] Calculating the recall accuracy and / or question-answering accuracy of the initial large language model according to the answer prediction information and the answer annotation information corresponding to the question training information;

[0023] If at least one of the recall accuracy and the question-answer accuracy is less than a corresponding accuracy threshold, fine-tuning the initial large language model to adjust a weight parameter of the initial large language model;

[0024] Until the recall accuracy and the question-answer accuracy reach corresponding accuracy thresholds, a second largest language model is obtained.

[0025] Optionally, the processing the domain information and the target maintenance document based on the second largest language model to determine the answer information corresponding to the question information includes:

[0026] Inputting the domain information and the associated target maintenance document into the second language model, so that the second language model extracts title information and maintenance content information of each segment in the target maintenance document according to the domain information, and extracts question information on device maintenance in the domain information;

[0027] Calculating a relevance score between the question information and the target maintenance document based on the title information, the maintenance content information and the question information;

[0028] The fragments corresponding to the relevance scores greater than or equal to the preset relevance threshold are marked, and the marked fragments are assembled to obtain reply information, wherein the reply information includes: maintenance content information, source information and reference path information.

[0029] Optionally, before determining the target maintenance document, the method includes:

[0030] Inputting the question information into the second largest language model to obtain initial answer information output by the second largest language model;

[0031] Calculating the recall rate of the answer information of the second largest language model according to the preset answer calibration information corresponding to the initial answer information and the question information;

[0032] If the recall rate of the reply information is lower than or equal to a preset recall rate threshold, determining a target maintenance document according to the question information and a preset target knowledge base is executed;

[0033] If the recall rate of the reply information is higher than the preset recall rate threshold, the initial reply information is set as the reply information.

[0034] Optionally, after responding to the inquiry information regarding equipment maintenance, the method further comprises:

[0035] Performing compliance verification on the question information to verify whether the question information contains sensitive information;

[0036] If the question information does not contain sensitive information, determining the target maintenance document according to the question information and a preset target knowledge base;

[0037] After the processing of the domain information and the target maintenance document based on the second language model to determine the answer information corresponding to the question information, the method further includes:

[0038] Performing compliance verification on the reply information to verify whether the reply information contains sensitive information;

[0039] If the reply information does not contain sensitive information, the reply information is output.

[0040] In a second aspect, an embodiment of the present application provides a reply device, the device further comprising:

[0041] A first determination module, configured to determine target maintenance documents in response to question information regarding equipment maintenance, wherein the target maintenance documents include at least two types of maintenance documents having the same equipment characteristics, and each type of the maintenance documents has a different source database;

[0042] an acquisition module, configured to process the target maintenance document based on a preset first language model to obtain domain information, wherein the domain information includes at least one of reply strategy information, document format information, and device information, and the domain information is used to instruct a preset second language model to process the target maintenance document;

[0043] The second determination module is used to process the domain information and the target maintenance document based on the second largest language model to determine the answer information corresponding to the question information.

[0044] In a third aspect, an embodiment of the present application provides an electronic device, which further includes: a memory and a processor, wherein the memory stores a plurality of instructions; the processor loads instructions from the memory to execute the steps of the reply method of the present application.

[0045] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores a plurality of instructions, and the instructions are suitable for a processor to load to execute the steps in any one of the reply methods provided in the embodiments of the present application.

[0046] Beneficial effects of the embodiments of the present application:

[0047] Compared with the prior art in which maintenance personnel repair equipment failures based on experience, the present application can first extract maintenance documents from different source databases based on the problem information regarding equipment maintenance input by the user, and then use the large language model to generate domain information, and extract reply information from the maintenance documents based on the domain information and return it to the user. It can be seen that the present application can extract maintenance documents from multiple databases, ensuring the integrity and accuracy of the reply information. On this basis, the present application, based on the knowledge aggregation and understanding capabilities of the large language model, can solve various types of questions asked by users, provide users with effective maintenance strategies for different equipment failures, assist users in equipment repair, improve equipment maintenance efficiency, and also reduce equipment maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 This is a first schematic diagram of a reply process provided by an embodiment of the present application;

[0050] Figure 2 It is a schematic diagram of field information provided by an embodiment of the present application;

[0051] Figure 3 is a schematic diagram of a system provided by an embodiment of the present application;

[0052] Figure 4 It is a schematic diagram of knowledge base construction provided by an embodiment of the present application;

[0053] Figure 5 This is a first schematic diagram of large language model training provided by an embodiment of the present application;

[0054] Figure 6 This is a second schematic diagram of large language model training provided by an embodiment of the present application;

[0055] Figure 7 It is a schematic diagram of the interface display provided by the embodiment of the present application;

[0056] Figure 8 It is a first schematic diagram of the compliance verification process provided by an embodiment of the present application;

[0057] Fig. 9 is a schematic diagram of a reply device provided in an embodiment of the present application;

[0058] Fig.10 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0060] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0061] In the embodiment of the present application, "and / or" describes the association relationship of the associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / ", unless otherwise specified, generally indicates that the associated objects before and after are in an "or" relationship.

[0062] In this application, the word "exemplary" is used to mean "serving as an example, illustration, or illustration." Any embodiment described in this application as "exemplary" is not necessarily to be construed as being preferred or advantageous over other embodiments. The following description is given to enable any person skilled in the art to implement and use the invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the invention can be implemented without using these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the invention with unnecessary details. Therefore, the present application is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0063] Combined with the above background technology description of this application, when equipment fails, most industries currently rely mainly on human experience or equipment manuals for equipment maintenance. Maintenance personnel need to judge the cause of the equipment failure based on the equipment failure phenomenon and the equipment manual. The entire maintenance process is difficult to standardize and uniformly manage, and is highly dependent on the experience of maintenance personnel, resulting in low equipment maintenance efficiency and failure to effectively eliminate equipment failures.

[0064] In order to solve the above problems, the present application proposes a reply method, device, electronic device, computer-readable storage medium and computer program product, which aim to return corresponding reply information for the question information of equipment failure, so that maintenance personnel can directly refer to the reply information to repair the equipment, which can not only improve the accuracy of equipment fault diagnosis and equipment maintenance efficiency, but also reduce labor costs, lower the threshold for maintenance personnel, and realize unified management of equipment failure maintenance strategies.

[0065] Specifically, the reply method in the embodiment of the present application is as follows: Figure 1 As shown, the following steps may be included:

[0066] S10, in response to the question information for equipment maintenance, determining a target maintenance document, wherein the target maintenance document includes at least two types of maintenance documents having the same equipment characteristics, and each type of the maintenance documents has a different source database;

[0067] It should be noted that in this embodiment, when a device fails, maintenance personnel can input question information about the device failure to the corresponding terminal device, and the question information input by the user can include device maintenance related information such as fault code and fault description information. For example, maintenance personnel can input "Why is the device indicator light always on red" to the terminal device.

[0068] Among them, the terminal equipment can have a maintenance assistant function, support fault query and analysis functions, and provide maintenance assistance to users.

[0069] The terminal device may be a network device such as a mobile phone, a tablet, a notebook, a server, etc. This embodiment does not specifically limit the type of the terminal device.

[0070] On this basis, the terminal device can respond to the user's question information regarding device maintenance and determine the target maintenance document corresponding to the question information.

[0071] The target maintenance documents in this embodiment may include at least two types of maintenance documents having the same device characteristics, and each type of maintenance document has a different source database.

[0072] For example, in this embodiment, the database may include a local knowledge base and a cloud knowledge base, wherein the local knowledge base stores equipment-related documents, such as product manuals, maintenance records, equipment-related videos, drawings, spare parts library files, and major accident reports, etc. The cloud knowledge base can be understood as Internet information, such as websites, forums, and other equipment-related information. In this embodiment, the cloud knowledge base can be used as a supplement and support for the local knowledge base to provide users with more comprehensive and accurate response information and assist users in equipment maintenance.

[0073] S20, processing the target maintenance document based on a preset first language model to obtain domain information, where the domain information includes at least one of a type, specification, and function of a device, and the domain information is used to instruct a preset second language model to process the target maintenance document;

[0074] It should be noted that, in this embodiment, large language models (LLMs) are a type of natural language processing models with large-scale parameters and computing power. They are trained with a large amount of data and parameters to generate descriptive text similar to humans or answer natural language questions, such as GPT (Generative Pre-trained Transformer), Bard, T5 (Text-to-Text Transfer Transformer), PaLM (PathwaysLanguage Model), etc. This embodiment does not limit the type and specific construction method of the large language model.

[0075] The first language model in this embodiment can be used to process the target maintenance document and generate domain information.

[0076] It is understandable that in the field of artificial intelligence, especially in natural language processing (NLP), "prompt" is a technique that can be used to guide a model to generate a specific output or perform a specific task. In the context of a large language model, a prompt refers to an input provided to the model, which is usually a question, a text, or any type of instruction to guide the model to generate a specific output. In this embodiment, the prompt can specifically be domain information, such as device-related domain knowledge.

[0077] In this embodiment, the preset first language model can be used to process the target maintenance document to obtain domain information. It is understandable that the domain information can also be manually written by the user. However, this embodiment can directly and automatically generate the domain information to inject the domain information into the preset second language model, guiding the second language model to extract more accurate response information.

[0078] For example, Figure 2 The prompt shown includes response strategy information, such as answer generation guidelines, and document format information, such as local knowledge base format instructions.

[0079] It can be understood that the functions of the first language model and the second language model in this embodiment are not the same. The first language model is used to automatically generate domain information, and the second language model is used to automatically generate reply information. Therefore, there are differences in the model hierarchy structure, weights and other parameters of the two large language models, which will not be elaborated.

[0080] S30: Process the domain information and the target maintenance document based on the second largest language model to determine answer information corresponding to the question information.

[0081] In this embodiment, after extracting the target maintenance document and the domain information prompt, the terminal device can input the target maintenance document and the domain information into the second largest language model so that the second largest language model can process the domain information and the target maintenance document to determine the answer information corresponding to the question information.

[0082] Therefore, in this embodiment, the terminal device can respond to the user's question information regarding equipment maintenance, determine the target maintenance document corresponding to the question information, and then use the preset first large language model to process the target maintenance document to obtain domain information, and finally input the target maintenance document and domain information into the second large language model to obtain the reply information corresponding to the above question information. It can be seen that compared with the prior art in which maintenance personnel repair equipment failures based on experience, the present application can first extract maintenance documents from different source databases based on the user's input question information regarding equipment maintenance, and then use the large language model to generate domain information, and extract reply information from the maintenance document based on the domain information and return it to the user. It can be seen that the present application can extract maintenance documents from multiple databases to ensure the integrity and accuracy of the reply information. On this basis, the present application, based on the knowledge aggregation and understanding ability of the large language model, can solve various types of questions asked by users, provide users with effective maintenance strategies for different equipment failures, assist users in equipment maintenance, improve equipment maintenance efficiency, and also reduce equipment maintenance costs.

[0083] In one embodiment, in the above S10, “determining a target maintenance document in response to the inquiry information regarding equipment maintenance” may include:

[0084] S101, querying a preset local knowledge base according to maintenance description information in the question information to obtain at least one first maintenance document; wherein the maintenance description information includes at least one of a fault code and / or a fault phenomenon description information;

[0085] S102, judging whether there is any information missing position in the first maintenance document according to the structural information, logical information and / or semantic information of the first maintenance document;

[0086] S103, if there is an information missing location, query a preset cloud database according to the information missing location and the context corresponding to the information missing location to obtain a second maintenance document;

[0087] S104: According to the information missing position, fuse the first maintenance document and the second maintenance document to obtain a target maintenance document.

[0088] It should be noted that, in this embodiment, Figure 3 As shown, the target database may include a local knowledge base and a cloud database. In combination with the above-mentioned embodiment, the local knowledge base stores equipment-related documents, such as product manuals, maintenance records, equipment-related videos, drawings, spare parts library files, and major accident reports, etc. The cloud knowledge base can be understood as Internet information, such as websites, forums, and other equipment-related information.

[0089] On this basis, after receiving the question information input by the user, the terminal device can query the local knowledge base according to the maintenance description information in the question information to obtain at least one first maintenance document, wherein the maintenance description information in this embodiment may include fault code, fault phenomenon description information, etc. In addition, the maintenance description information in this embodiment may also include fault questions and equipment-related questions, etc. This embodiment does not specifically limit the content of the question information.

[0090] It is understandable that, since the document content in the local knowledge base is limited and is not updated in a timely manner, after querying and obtaining at least one first maintenance document, this embodiment needs to proofread the content integrity of the first maintenance document.

[0091] Specifically, for example, the terminal device may determine whether there is an information missing position in the first maintenance document according to the structural information, logical information and / or semantic information of the first maintenance document.

[0092] If it is determined that there is a missing information location in the first maintenance document, the preset cloud database can be queried according to the missing information location and the context corresponding to the missing information location to obtain the second maintenance document.

[0093] Furthermore, the terminal device may fuse the first maintenance document with the second maintenance document according to the information missing position to obtain the target maintenance document.

[0094] For example, the terminal device may fill the content in the second maintenance document into the information missing position in the first maintenance document to obtain the target maintenance document.

[0095] In one embodiment, if the content of the second maintenance document can cover the content of the first maintenance document, the second maintenance document may also be directly set as the target maintenance document.

[0096] It can be seen that in this embodiment, the cloud knowledge base serves as a backup strategy for the local knowledge base. When the local knowledge base cannot provide users with complete and accurate documents, the cloud knowledge base can supplement the extracted local documents to collect documents that are more comprehensively related to the question information, which will be used for the subsequent generation of response information for equipment maintenance.

[0097] In one embodiment, the steps of constructing a local knowledge base may include:

[0098] Obtain maintenance documentation for multiple devices;

[0099] For each device maintenance document, performing layout analysis on the device maintenance document to identify target elements in the device maintenance document;

[0100] According to the target element, the device maintenance document is sliced ​​to obtain a plurality of knowledge slices;

[0101] For each of the knowledge slices, vectorization is performed on the knowledge slice, and the vectorized knowledge slice is imported into a local knowledge base.

[0102] In this embodiment, the terminal device can obtain multiple equipment maintenance documents. The equipment maintenance documents can be product manuals, historical maintenance records, equipment drawings and other documents. For example, a large number of documents can be collected from the local and cloud, and the documents in the cloud can be downloaded to the local. This embodiment does not limit the type of documents, such as plain text type, graphic type (doc / docx, pdf), table (excel), audio and video, etc.

[0103] On this basis, if Figure 4 As shown, for each device maintenance document, the terminal device can perform layout analysis on the device maintenance document to identify target elements in the device maintenance document, wherein the target elements may include tables, pictures, text, etc. in the device maintenance document, wherein the layout analysis may include but is not limited to optical character recognition (OCR), image extraction, table extraction, layout recovery, etc., which will not be elaborated here.

[0104] Then, the terminal device can slice the device maintenance document according to the target element to obtain multiple knowledge slices, such as knowledge slice 1, knowledge slice 2, ..., knowledge slice n, etc. The knowledge slices are embedded and vectorized, and finally the vectorized knowledge slices are imported into the local knowledge base.

[0105] In this embodiment, the cloud knowledge base can be constructed in a variety of ways according to different technologies and requirements. For example, the following steps may be included:

[0106] Building a cloud database can follow the following steps and principles:

[0107] (1) Architecture design:

[0108] Layered architecture: Divide the database system into multiple layers, such as access layer, computing layer, storage layer, etc., to improve the scalability and maintainability of the system;

[0109] Distributed architecture: Distributed database technologies such as sharding and replication are used to achieve horizontal expansion of data and automatic failure transfer;

[0110] High availability design: Ensure high availability of the database through master-slave replication, multi-copy deployment, etc.

[0111] (2) Technology selection:

[0112] Database system: Choose a suitable database system based on business needs, including relational databases (such as MySQL, PostgreSQL), NoSQL databases (such as MongoDB, Cassandra), or NewSQL databases (such as TiDB, CockroachDB);

[0113] Virtualization technology: Use containerization technologies such as Docker and Kubernetes to achieve rapid deployment, resource isolation, and elastic scaling of databases;

[0114] Cloud service platform: Select a suitable cloud service platform as the underlying infrastructure, and use the computing, storage, network and other resources it provides to build a cloud database environment.

[0115] (3) Deployment and implementation:

[0116] Environment preparation: Create virtual machines, configure networks, mount storage and other resources on the cloud service platform to prepare for database deployment;

[0117] Database installation: Based on the technical selection results, download and install the database software, and configure the necessary parameters and permissions.

[0118] (4) Selection of cloud database service providers, such as Alibaba Cloud, Tencent Cloud, and Baidu Cloud.

[0119] Through the above steps and principles, a stable, reliable, and elastically scalable cloud database service can be built.

[0120] Through the above operations, this embodiment can build a vectorized local knowledge base and a cloud database. On this basis, this embodiment can use the local knowledge base and combine it with the maintenance documents in the cloud knowledge base to train the large language model in this embodiment.

[0121] In one embodiment, the training method of the preset second language model may further include:

[0122] Acquire question training information, and extract maintenance training documents corresponding to the question training information from the local knowledge base;

[0123] Inputting the domain training information input by the user, the question training information and the maintenance training document into the initial large language model to obtain answer prediction information;

[0124] Calculating the recall accuracy and / or question-answering accuracy of the initial large language model according to the answer prediction information and the answer annotation information corresponding to the question training information;

[0125] If at least one of the recall accuracy and the question-answer accuracy is less than a corresponding accuracy threshold, fine-tuning the initial large language model to adjust a weight parameter of the initial large language model;

[0126] Until the recall accuracy and / or the question-answer accuracy reaches a corresponding accuracy threshold, a second largest language model is obtained.

[0127] It should be noted that, in this embodiment, the domain-general model can be used as the initial language model in advance, so that the initial language model can be trained to obtain the preset second largest language model.

[0128] Specifically, for example, in this embodiment, the terminal device may first obtain question training information, wherein the question training information may be historical question information, for example, equipment fault description information and equipment maintenance strategy information raised by users during the operation process may be collected.

[0129] In one embodiment, the terminal device may process the question training information, such as extracting key information from the question training information, including extracting keywords from the question, identifying the user's intention of asking the question based on the keywords, and obtaining optimized question training information.

[0130] The terminal device may query the local knowledge base according to the optimized question training information to extract the maintenance training document corresponding to the question training information from the local knowledge base.

[0131] Then, the domain training information, question training information and maintenance training documents input by the user can be input into the initial large language model, so that the initial large language model processes the domain training information, question training information and maintenance training documents to obtain answer prediction information.

[0132] The terminal device then calculates the recall accuracy and / or question-answering accuracy of the initial large language model based on the answer prediction information and the answer annotation information corresponding to the question training information.

[0133] If it is determined that at least one of the recall accuracy and the question-answer accuracy is less than the corresponding accuracy threshold, it means that the initial large language model cannot accurately predict the answer information corresponding to the question information. In this case, the initial large language model can be fine-tuned to adjust the weight parameters of the initial large language model, for example, Figure 5 As shown, the initial large language model may include a Transformer encoder layer and a Transformer decoder layer. When training the initial large language model, the weight parameters of the initial large language model may be adjusted according to the output recall accuracy and / or question-answering accuracy.

[0134] By repeatedly adjusting the weight parameters of the initial large language model until the recall accuracy and question-answer accuracy reach the corresponding accuracy thresholds, it means that the initial large language model can accurately predict the user's intention based on the question information input by the user and feedback the corresponding reply information. Therefore, the initial large language model at this time can be set as the preset second language large model.

[0135] Specifically, for example, Figure 6 As shown, this embodiment can detect whether the initial large language model is trained through multiple evaluation parameters such as recall accuracy, answer fidelity, and answer completeness output by the initial large language model. If it is detected that the recall accuracy is 99% at this time, which has exceeded the corresponding accuracy threshold (such as 98%), and the question-answer accuracy is 97.34%, which has reached the corresponding accuracy threshold (such as 95%), then the initial large language model at this time can be set as the preset second largest language model.

[0136] Similarly, in this embodiment, the training process of the preset first largest language model is substantially the same as the training process of the second largest language model, for example:

[0137] Obtain question training information, and extract maintenance training documents corresponding to the question training information from the local knowledge base;

[0138] Input the question training information and the maintenance training document into the initial large language model (different from the above large language model, the initial large language model here may be the initial second largest language model) to obtain the domain prediction information;

[0139] Calculate the prediction accuracy of the initial second largest language model based on the domain prediction information and the domain calibration information input by the user;

[0140] If the above prediction accuracy is greater than a preset accuracy threshold, the initial second largest language model may be set as the first largest language model.

[0141] If the above prediction accuracy is less than or equal to the preset accuracy threshold, the initial second largest language model can be fine-tuned to adjust the weight parameters of the large model until the large model can output accurate domain information prompt.

[0142] The first language model can automatically generate domain information prompts based on maintenance documents and question information, and use the domain information prompts to guide the second language model to extract answer information corresponding to the question information from the target maintenance document, so as to accurately predict user intentions and provide support for users in equipment maintenance.

[0143] In one embodiment, in the above S30, “processing the domain information and the target maintenance document based on the second largest language model to determine the answer information corresponding to the question information” may include:

[0144] S301, inputting the domain information and the associated target maintenance document into the second language model, so that the second language model extracts title information and maintenance content information of each segment in the target maintenance document according to the domain information, and extracts question information on device maintenance in the domain information;

[0145] S302, calculating a correlation score between the question information and the target maintenance document according to the title information, the maintenance content information and the question information;

[0146] S303, marking the segments corresponding to the relevance scores greater than or equal to the preset relevance threshold, and assembling the marked segments to obtain reply information, wherein the reply information includes: maintenance content information, source information and reference path information.

[0147] In this embodiment, after obtaining the domain information prompt and the target maintenance document, the terminal device can input the domain information prompt and the target maintenance document into the trained second largest language model, so that the large language model can extract the title information and maintenance content information of each fragment in the target maintenance document, and extract the question information about device maintenance in the domain information.

[0148] It should be noted that, in this embodiment, the domain information prompt may include task requirements:

[0149] (1) Check the title and content of the knowledge base fragments one by one to evaluate their relevance to the user's question; if relevant, add the index value of the document fragment to related_paras;

[0150] (2) If the topic, subject, or object of the user's question matches the title or content of the reference segment, the segment is deemed to be related to the user's question, and its index value is added to related_paras;

[0151] (3) If the snippet contains the answer to the user’s question, it is considered relevant and its index value is added to related_paras.

[0152] (4) If a segment cannot directly answer the user's question but can provide useful auxiliary information, it should also be considered relevant and its index value should be added to related_params;

[0153] (5) If no segment is related to the user question, set relation = false; if there is one or more related segments, set relation = true;

[0154] (6) Output requirements: Use JSON format: {"relation":true / false,"related_paras":[1,2,3]}. No additional explanation is required.

[0155] On this basis, the terminal device can calculate the relevance score between the question information and the target maintenance document according to the title information, the maintenance content information and the question information, and obtain the segment corresponding to the relevance score greater than or equal to the preset relevance threshold.

[0156] Then, the fragments corresponding to the relevance scores greater than or equal to the preset relevance threshold can be marked, and finally all the marked fragments are assembled to obtain the reply information.

[0157] It is worth noting that in this embodiment, Figure 7 As shown, the reply information in this embodiment not only includes the reply content corresponding to the question information, but also may include the source information of the reply content, for example, the reply content is derived from the local knowledge base and / or the cloud knowledge base, and may also include reference path information. For example, if the reply content comes from the cloud knowledge base, the reference path information may be web link information, so that the user can jump to the source document through the web link information.

[0158] In addition, Figure 7As shown, the reply information may also include a knowledge search option, a maintenance record query option, a fault code query option, and the like.

[0159] In one embodiment, before the above S10, “determining the target maintenance document”, the following steps may also be included:

[0160] S40, inputting the question information into the second largest language model to obtain initial answer information output by the second largest language model;

[0161] S50, calculating the recall rate of the reply information of the large language model according to the preset reply calibration information corresponding to the initial reply information and the question information;

[0162] S60, if the recall rate of the reply information is lower than or equal to a preset recall rate threshold, determining a target maintenance document according to the question information and a preset target knowledge base;

[0163] S70: If the recall rate of the reply information is higher than the preset recall rate threshold, the initial reply information is set as the reply information.

[0164] In this embodiment, after obtaining the question information input by the user, the terminal device can directly input the question information into the second largest language model, and the second largest language model processes the question information to obtain the initial answer information output by the second largest language model.

[0165] Then, the recall rate of the reply information of the second largest language model can be calculated by comparing the similarity between the initial reply information and the preset reply calibration information corresponding to the question information, wherein the preset reply calibration information can be understood as the reply information corresponding to each equipment maintenance problem calibrated by expert experience.

[0166] If it is determined that the recall rate of the reply information is lower than or equal to the preset recall rate threshold, it means that the second largest language model cannot accurately predict the user's intention. Then it is still necessary to execute: determine the target maintenance document based on the question information and the preset target knowledge base; process the target maintenance document based on the preset first largest language model; process the domain information and the target maintenance document based on the second largest language model to determine the reply information corresponding to the question information, so as to improve the accuracy of the reply. Please refer to the description of the above embodiment and will not repeat it here.

[0167] If it is determined that the recall rate of the reply information is higher than the preset recall rate threshold, it means that the second largest language model can accurately predict the user's intention. In this case, the initial reply information can be directly set as the reply information, which not only ensures the accuracy of the reply, but also improves the reply efficiency.

[0168] In one embodiment, after the above S10, “responding to the inquiry information regarding equipment maintenance”, the following may also be included:

[0169] S80, performing compliance verification on the question information to verify whether the question information contains sensitive information;

[0170] S90, if the question information does not contain sensitive information, determining a target maintenance document according to the question information and a preset target knowledge base;

[0171] In this embodiment, the terminal device may perform compliance verification on the question information input by the user to verify whether the question information contains sensitive information.

[0172] If it is determined that the question information contains sensitive information, the question information can be ignored and a sensitive information prompt can be output.

[0173] If it is determined that the question information does not contain sensitive information, the following can be executed: determining the target maintenance document, processing the target maintenance document based on the preset first largest language model, and obtaining domain information; processing the domain information and the target maintenance document based on the second largest language model, and determining the reply information corresponding to the question information. Please refer to the description of the above embodiment and will not repeat it here.

[0174] Similarly, after the above S30, “processing the domain information and the target maintenance document based on the second largest language model to determine the answer information corresponding to the question information”, the following may also be included:

[0175] S100, performing compliance verification on the reply information to verify whether the reply information contains sensitive information;

[0176] S110: If the reply information does not contain sensitive information, output the reply information.

[0177] In this embodiment, after the terminal device obtains the reply information corresponding to the question information through the second largest language model, it can perform compliance verification on the reply information to verify whether the reply information contains sensitive information.

[0178] If it is detected that the reply information does not contain sensitive information, the reply information can be output through the display interface of the terminal device, for example, Figure 7 As shown, the reply content, source information, reference path information, etc. in the reply information can be displayed.

[0179] If it is detected that the reply information contains sensitive information, the terminal device can discard the reply information and re-execute: process the domain information and target maintenance document through the second largest language model to determine the reply information corresponding to the question information. The description of the above embodiment of the parameters will not be repeated here.

[0180] In one embodiment, if Figure 8 As shown, this embodiment can perform compliance training on the large language model so that the large language model can identify problem information, domain information, and whether sensitive words exist in the document.

[0181] In general, if Figure 8 As shown, in this embodiment, compliance verification can be performed on both questions and answers, and answers will be output only when the compliance verification passes, ensuring full life cycle application compliance.

[0182] Accordingly, the embodiment of the present application also provides a reply device, such as Fig. 9 As shown, the reply device may include:

[0183] A first determination module 1001 is used to determine target maintenance documents in response to question information about equipment maintenance, wherein the target maintenance documents include at least two types of maintenance documents having the same equipment characteristics, and each type of maintenance document has a different source database;

[0184] An acquisition module 1002 is used to process the target maintenance document based on a preset first language model to obtain domain information, wherein the domain information includes at least one of reply strategy information, document format information and device information, and the domain information is used to instruct a preset second language model to process the target maintenance document;

[0185] The second determination module 1003 is used to process the domain information and the target maintenance document based on the second largest language model to determine the answer information corresponding to the question information.

[0186] Optionally, the first determining module 1001 is further configured to:

[0187] According to the maintenance description information in the question information, query a preset local knowledge base to obtain at least one first maintenance document; wherein the maintenance description information includes at least one of a fault code and / or a fault phenomenon description information;

[0188] Determining whether there is any information missing location in the first maintenance document according to the structural information, logical information and / or semantic information of the first maintenance document;

[0189] If there is an information missing location, querying a preset cloud database according to the information missing location and the context corresponding to the information missing location to obtain a second maintenance document;

[0190] According to the information missing position, the first maintenance document and the second maintenance document are information-fused to obtain a target maintenance document.

[0191] Optionally, the construction of the local knowledge base includes:

[0192] Obtain maintenance documentation for multiple devices;

[0193] For each device maintenance document, performing layout analysis on the device maintenance document to identify target elements in the device maintenance document;

[0194] According to the target element, the device maintenance document is sliced ​​to obtain a plurality of knowledge slices;

[0195] For the knowledge slices, vectorization processing is performed on the knowledge slices, and the vectorized knowledge slices are imported into a local knowledge base.

[0196] Optionally, the training of the second largest language model includes:

[0197] Acquire question training information, and extract maintenance training documents corresponding to the question training information from the local knowledge base;

[0198] Inputting the domain training information input by the user, the question training information and the maintenance training document into the initial large language model to obtain answer prediction information;

[0199] Calculating the recall accuracy and / or question-answering accuracy of the initial large language model according to the answer prediction information and the answer annotation information corresponding to the question training information;

[0200] If at least one of the recall accuracy and the question-answer accuracy is less than a corresponding accuracy threshold, fine-tuning the initial large language model to adjust a weight parameter of the initial large language model;

[0201] Until the recall accuracy and the question-answer accuracy reach corresponding accuracy thresholds, a second largest language model is obtained.

[0202] Optionally, the second determining module 1003 is further configured to:

[0203] Inputting the domain information and the associated target maintenance document into the second language model, so that the second language model extracts title information and maintenance content information of each segment in the target maintenance document according to the domain information, and extracts question information on device maintenance in the domain information;

[0204] Calculating a relevance score between the question information and the target maintenance document based on the title information, the maintenance content information and the question information;

[0205] The fragments corresponding to the relevance scores greater than or equal to the preset relevance threshold are marked, and the marked fragments are assembled to obtain reply information, wherein the reply information includes: maintenance content information, source information and reference path information.

[0206] Optionally, the reply device in the present application further includes:

[0207] An input module, used for inputting the question information into the second largest language model to obtain initial answer information output by the second largest language model;

[0208] A calculation module, used for calculating the recall rate of the answer information of the second largest language model according to the preset answer calibration information corresponding to the initial answer information and the question information;

[0209] A third determination module is used to determine the target maintenance document according to the question information and a preset target knowledge base if the recall rate of the reply information is lower than or equal to a preset recall rate threshold;

[0210] A setting module is used to set the initial reply information as reply information if the recall rate of the reply information is higher than the preset recall rate threshold.

[0211] Optionally, the reply device in the present application further includes:

[0212] A first verification module is used to perform compliance verification on the question information to verify whether the question information contains sensitive information; if the question information does not contain sensitive information, determine the target maintenance document based on the question information and a preset target knowledge base;

[0213] The second verification module is used to perform compliance verification on the reply information to verify whether the reply information contains sensitive information; if the reply information does not contain sensitive information, output the reply information.

[0214] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0215] Accordingly, the present application also provides an electronic device, such as Fig.10 As shown, Fig.10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device 1100 includes a processor 1101 having one or more processing cores, a memory 1102 having one or more computer-readable storage media, and a computer program stored in the memory 1102 and executable on the processor. The processor 1101 is electrically connected to the memory 1102. It will be understood by those skilled in the art that the electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0216] The processor 1101 is the control center of the electronic device 1100, and uses various interfaces and lines to connect various parts of the entire electronic device 1100. By running or loading software programs and / or units stored in the memory 1102, and calling data stored in the memory 1102, the processor 1101 executes various functions of the electronic device 1100 and processes data, thereby monitoring the electronic device 1100 as a whole. The processor 1101 can be a processor CPU, a graphics processor GPU, a network processor (Network Processor, NP), etc., and can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application.

[0217] In the embodiment of the present application, the processor 1101 in the electronic device 1100 will load instructions corresponding to the processes of one or more application programs into the memory 1102 according to the following steps, and the processor 1101 will run the application programs stored in the memory 1102 to implement various functions, such as:

[0218] In response to the question information regarding equipment maintenance, target maintenance documents are determined, wherein the target maintenance documents include at least two types of maintenance documents having the same equipment characteristics, and each type of the maintenance documents has a different source database;

[0219] Processing the target maintenance document based on a preset first language model to obtain domain information, the domain information including at least one of reply strategy information, document format information and device information, the domain information being used to instruct a preset second language model to process the target maintenance document;

[0220] The domain information and the target maintenance document are processed based on the second largest language model to determine answer information corresponding to the question information.

[0221] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0222] Optional, such as Fig.10 As shown, the electronic device 1100 also includes: a touch display screen 1103 , a radio frequency circuit 1104 , an audio circuit 1105 , an input unit 1106 , and a power supply 1107 .

[0223] In addition, the electronic device 1100 further includes: a security module 1108 , a device chip 1109 and a hard disk drive 1110 .

[0224] The processor 1101 is electrically connected to the touch screen 1103, the radio frequency circuit 1104, the audio circuit 1105, the input unit 1106 and the power supply 1107. Those skilled in the art will appreciate that Fig.10The electronic device structure shown in the figure does not constitute a limitation of the electronic device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0225] The touch display screen 1103 can be used to display a graphical user interface and receive operation instructions generated by the user acting on the graphical user interface. The touch display screen 1103 may include a display panel and a touch panel. Among them, the display panel can be used to display information input by the user or information provided to the user and various graphical user interfaces of the electronic device, and these graphical user interfaces can be composed of graphics, text, icons, videos and any combination thereof. Optionally, the display panel can be configured in the form of a liquid crystal display (LCD, Liquid Crystal Display), an organic light-emitting diode (OLED, Organic Light-EmittingDiode) and the like. The touch panel can be used to collect the user's touch operation on or near it (such as the user uses any suitable object or attachment such as a finger, a stylus, etc. on the touch panel or near the touch panel), and generate corresponding operation instructions, and the operation instructions execute corresponding programs. Optionally, the touch panel may include two parts: a touch detection device and a touch controller. Among them, the touch detection device detects the user's touch orientation, detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into the touch point coordinates, and then sends it to the processor 1101, and can receive the command sent by the processor 1101 and execute it. The touch panel can cover the display panel. When the touch panel detects a touch operation on or near it, it is transmitted to the processor 1101 to determine the type of touch event, and then the processor 1101 provides a corresponding visual output on the display panel according to the type of touch event. In an embodiment of the present application, the touch panel and the display panel can be integrated into the touch display screen 1103 to realize the input and output functions. However, in some embodiments, the touch panel and the touch panel can be used as two independent components to realize the input and output functions. That is, the touch display screen 1103 can also be used as a part of the input unit 1106 to realize the input function.

[0226] The radio frequency circuit 1104 may be used to send and receive radio frequency signals, so as to establish wireless communication with a network device or other electronic devices through wireless communication, and to send and receive signals between the network device or other electronic devices.

[0227] The audio circuit 1105 can be used to provide an audio interface between the user and the electronic device through a speaker and a microphone. The audio circuit 1105 can transmit the electrical signal converted from the received audio data to the speaker, which is converted into a sound signal for output; on the other hand, the microphone converts the collected sound signal into an electrical signal, which is received by the audio circuit 1105 and converted into audio data, and then the audio data is output to the processor 1101 for processing, and then sent to another electronic device through the radio frequency circuit 1104, or the audio data is output to the memory 1102 for further processing. The audio circuit 1105 may also include an earplug jack to provide communication between an external headset and an electronic device.

[0228] The input unit 1106 may be used to receive input numbers, character information or user feature information (such as fingerprint, iris, facial information, etc.), and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0229] The power supply 1107 is used to supply power to various components of the electronic device 1100. Optionally, the power supply 1107 can be logically connected to the processor 1101 through a power management system, so that the power management system can manage charging, discharging, power consumption and other functions. The power supply 1107 can also include one or more DC or AC power supplies, recharging systems, power maintenance detection circuits, power converters or inverters, power status indicators and other arbitrary components.

[0230] The security module 1108 can be used to respond to a device startup event on the electronic device 1100 and read encryption information and device information from the non-volatile memory of the device chip 1109; the encryption information and device information stored in the non-volatile memory are obtained by authentication when the electronic device first accesses the cloud; a drive encryption key is generated based on the encryption information and device information, and the hard disk drive of the device chip 1109 is activated; the hard disk drive 1110 can be used to receive data operation instructions triggered by the target application on the activated hard disk drive, and operate the target data based on the drive encryption key.

[0231] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0232] although Fig.10 Not shown, the electronic device 1100 may also include a camera, a sensor, a wireless fidelity module, a Bluetooth module, etc., which will not be described in detail here.

[0233] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0234] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0235] To this end, an embodiment of the present application provides a computer-readable storage medium, in which a plurality of computer programs are stored. The computer program can be loaded by a processor to execute any one of the reply methods provided in the embodiment of the present application. The computer program can execute the steps of the following reply method:

[0236] In response to the question information regarding equipment maintenance, target maintenance documents are determined, wherein the target maintenance documents include at least two types of maintenance documents having the same equipment characteristics, and each type of the maintenance documents has a different source database;

[0237] Processing the target maintenance document based on a preset first language model to obtain domain information, the domain information including at least one of reply strategy information, document format information and device information, the domain information being used to instruct a preset second language model to process the target maintenance document;

[0238] The domain information and the target maintenance document are processed based on the second largest language model to determine answer information corresponding to the question information.

[0239] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0240] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0241] Since the computer program stored in the computer-readable storage medium can execute any one of the reply methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any one of the reply methods provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0242] In the above-mentioned electronic device, computer-readable storage medium, and electronic device embodiments, the description of each embodiment has its own emphasis. For parts not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process and beneficial effects of the electronic device, computer-readable storage medium, electronic device and its corresponding units described above can refer to the description of the reply method in the above embodiment, and will not be repeated here.

[0243] The above is a detailed introduction to a reply method, an electronic device, and a computer-readable storage medium provided in an embodiment of the present application. Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A reply method, characterized in that: The method comprises: In response to the question information regarding equipment maintenance, target maintenance documents are determined, wherein the target maintenance documents include at least two types of maintenance documents having the same equipment characteristics, and each type of the maintenance documents has a different source database; Processing the target maintenance document based on a preset first language model to obtain domain information, the domain information including at least one of reply strategy information, document format information and device information, the domain information being used to instruct a preset second language model to process the target maintenance document; The domain information and the target maintenance document are processed based on the second largest language model to determine answer information corresponding to the question information.

2. The reply method according to claim 1, characterized in that: The response and the inquiry information for equipment maintenance determine the target maintenance document, including: According to the maintenance description information in the question information, query a preset local knowledge base to obtain at least one first maintenance document; wherein the maintenance description information includes at least one of a fault code and / or a fault phenomenon description information; Determining whether there is any information missing location in the first maintenance document according to the structural information, logical information and / or semantic information of the first maintenance document; If there is an information missing location, querying a preset cloud database according to the information missing location and the context corresponding to the information missing location to obtain a second maintenance document; According to the information missing position, the first maintenance document and the second maintenance document are information-fused to obtain a target maintenance document.

3. The reply method according to claim 2, characterized in that: The steps of constructing the local knowledge base include: Obtain maintenance documentation for multiple devices; For each device maintenance document, performing layout analysis on the device maintenance document to identify target elements in the device maintenance document; According to the target element, the device maintenance document is sliced ​​to obtain a plurality of knowledge slices; For the knowledge slices, vectorization processing is performed on the knowledge slices, and the vectorized knowledge slices are imported into a local knowledge base.

4. The reply method according to claim 3, characterized in that: The training steps of the second largest language model include: Acquire question training information, and extract maintenance training documents corresponding to the question training information from the local knowledge base; Inputting the domain training information input by the user, the question training information and the maintenance training document into the initial large language model to obtain answer prediction information; Calculating the recall accuracy and / or question-answering accuracy of the initial large language model according to the answer prediction information and the answer annotation information corresponding to the question training information; If at least one of the recall accuracy and the question-answer accuracy is less than a corresponding accuracy threshold, fine-tuning the initial large language model to adjust a weight parameter of the initial large language model; Until the recall accuracy and the question-answer accuracy reach corresponding accuracy thresholds, a second largest language model is obtained.

5. The reply method according to any one of claims 2 to 4, characterized in that: The processing of the domain information and the target maintenance document based on the second largest language model to determine the answer information corresponding to the question information includes: Inputting the domain information and the associated target maintenance document into the second language model, so that the second language model extracts title information and maintenance content information of each segment in the target maintenance document according to the domain information, and extracts question information on device maintenance in the domain information; Calculating a relevance score between the question information and the target maintenance document based on the title information, the maintenance content information and the question information; The fragments corresponding to the relevance scores greater than or equal to the preset relevance threshold are marked, and the marked fragments are assembled to obtain reply information, wherein the reply information includes: maintenance content information, source information and reference path information.

6. The reply method according to claim 1, characterized in that: Before determining the target maintenance document, include: Inputting the question information into the second largest language model to obtain initial answer information output by the second largest language model; Calculating the recall rate of the answer information of the second largest language model according to the preset answer calibration information corresponding to the initial answer information and the question information; If the recall rate of the reply information is lower than or equal to a preset recall rate threshold, determining a target maintenance document according to the question information and a preset target knowledge base is executed; If the recall rate of the reply information is higher than the preset recall rate threshold, the initial reply information is set as the reply information.

7. The reply method according to claim 1, characterized in that: After responding to the inquiry information about equipment maintenance, the method includes: Performing compliance verification on the question information to verify whether the question information contains sensitive information; If the question information does not contain sensitive information, determining the target maintenance document according to the question information and a preset target knowledge base; After the processing of the domain information and the target maintenance document based on the second language model to determine the answer information corresponding to the question information, the method further includes: Performing compliance verification on the reply information to verify whether the reply information contains sensitive information; If the reply information does not contain sensitive information, the reply information is output.

8. A reply device, characterized in that: The device comprises: A first determination module, configured to determine target maintenance documents in response to question information regarding equipment maintenance, wherein the target maintenance documents include at least two types of maintenance documents having the same equipment characteristics, and each type of the maintenance documents has a different source database; an acquisition module, configured to process the target maintenance document based on a preset first language model to obtain domain information, wherein the domain information includes at least one of reply strategy information, document format information, and device information, and the domain information is used to instruct a preset second language model to process the target maintenance document; The second determination module is used to process the domain information and the target maintenance document based on the second largest language model to determine the answer information corresponding to the question information.

9. An electronic device, characterized in that: The electronic device further comprises: a memory and a processor, wherein the memory stores a plurality of instructions; and the processor loads instructions from the memory to execute the steps of the reply method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the steps of the reply method according to any one of claims 1 to 7.

Images