Intelligent operation and maintenance method based on self-developed GPT model and related equipment thereof
Patent Information
- Application Number
- CN202410918209.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-09
- Publication Date
- 2024-11-08
AI Technical Summary
Traditional operation and maintenance methods rely too much on manual labor, resulting in inefficiency and poor user experience, and are unable to handle online problems efficiently.
An intelligent operation and maintenance method based on self-developed GPT model is adopted. By extracting text content from user questions, knowledge retrieval and quality evaluation are carried out, the question-answer is correct, and manual intervention is carried out when necessary to ensure the accuracy of the answer.
Reduce the dependence on labor during operation and maintenance, improve operation and maintenance efficiency and intelligence, ensure the accuracy of answers and update the knowledge base to improve the accuracy of future answers.
Smart Images

Figure CN118916453A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the fields of intelligent operation and maintenance and financial technology, and is applied to scenarios for operation and maintenance of online problems arising in business, and in particular to an intelligent operation and maintenance method based on a self-developed GPT model and related equipment. Background Art
[0002] With the development of financial technology, online problems for enterprises in related industries will gradually increase as the business scale expands. Generally speaking, operational and consulting issues account for a high proportion of these issues, resulting in greater challenges in terms of average processing time and completion rate within the time limit.
[0003] In response to these problems, in terms of operation and maintenance, the summary of the enterprise's knowledge base is crucial. However, the knowledge base of most enterprises has not been effectively used, and professional operation and maintenance personnel are often required to summarize FAQs based on their own experience to accumulate knowledge or provide solutions to faults. This leads to excessive reliance on manual labor, which requires a lot of manpower to solve problems, is inefficient, and results in poor user experience. Summary of the invention
[0004] The purpose of the embodiments of the present application is to propose an intelligent operation and maintenance method based on a self-developed GPT model and related equipment to solve the technical problem that the traditional operation and maintenance method is too dependent on manual labor and is inefficient when solving problems.
[0005] In order to solve the above technical problems, the embodiment of the present application provides an intelligent operation and maintenance method based on a self-developed GPT model, which adopts the following technical solutions:
[0006] An intelligent operation and maintenance method based on a self-developed GPT model includes the following steps:
[0007] Extract text content from user questions;
[0008] Perform knowledge retrieval according to the text content to obtain knowledge retrieval results;
[0009] Input the text content and the knowledge retrieval results into a preset GPT model to obtain a question-answer pair;
[0010] Performing a quality assessment on the question-answer pair to determine whether the question-answer pair meets a preset condition;
[0011] If the preset condition is met, the question-answer pair is output as the reply content corresponding to the user question;
[0012] If the preset condition is not met, the question-answer pair is displayed on the operation and maintenance interface, an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface is received, and a reply content corresponding to the user question is generated and output according to the operation and maintenance data.
[0013] Furthermore, the step of performing knowledge retrieval according to the text content to obtain knowledge retrieval results specifically includes:
[0014] A search is performed in a preset knowledge base according to the text content to determine whether there is reference information whose relevance to the text content is greater than or equal to a preset association threshold value as the knowledge search result.
[0015] Furthermore, the step of inputting the text content and the knowledge retrieval results into a preset GPT model to obtain a question-answer pair specifically includes:
[0016] identifying the knowledge retrieval result;
[0017] If the knowledge retrieval result indicates that the reference information exists, inputting the text content and the reference information into the GPT model, and calling the GPT model to output the question-answer pair;
[0018] If the knowledge retrieval result is that the reference information does not exist, the text content is input into the GPT model, and the GPT model is called to output the question-answer pair.
[0019] Furthermore, the step of performing quality assessment on the question-answer pair to determine whether the question-answer pair meets a preset condition specifically includes:
[0020] Calculating the similarity between the knowledge points in the knowledge base and the question-answer pair, and determining whether there are similar knowledge points whose similarity is greater than or equal to a preset similarity threshold;
[0021] If the similar knowledge point does not exist, determining that the question-answer pair meets the preset condition;
[0022] If the similar knowledge point exists, it is determined that the question-answer pair does not meet the preset condition.
[0023] Further, if the preset condition is not met, the question-answer pair is displayed on the operation and maintenance interface, an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface is received, and a reply content corresponding to the user question is generated according to the operation and maintenance data and outputted, specifically including:
[0024] If the preset conditions are not met, the question-answer pair and the similar knowledge points are displayed on the operation and maintenance interface; the operation and maintenance instruction is received and the operation and maintenance data is analyzed to obtain a reply suggestion;
[0025] Identify the question type of the question-answer pair, and determine the response template corresponding to the question type based on the pre-stored correspondence between different question types and different response templates;
[0026] The reply template is filled according to the reply suggestion, and the reply content is generated and output.
[0027] Further, after the steps of displaying the question-answer pair on the operation and maintenance interface if the preset condition is not met, receiving an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface, and generating and outputting a reply content corresponding to the user question according to the operation and maintenance data, the method further includes:
[0028] Recording the operation and maintenance data;
[0029] Performing statistical analysis on the operation and maintenance data to obtain quasi-question-answer pairs, and recording the number of occurrences of the quasi-question-answer pairs;
[0030] When the number of occurrences reaches a preset number, the quasi-question-answer pair is stored in the knowledge base.
[0031] Furthermore, the step of extracting text content from the user question specifically includes:
[0032] Identify the document type corresponding to the user question;
[0033] Text extraction is performed according to the document type to obtain the text content.
[0034] In order to solve the above technical problems, the embodiment of the present application also provides an intelligent operation and maintenance system based on a self-developed GPT model, which adopts the following technical solutions:
[0035] An intelligent operation and maintenance system based on a self-developed GPT model, including:
[0036] The extraction module is used to extract text content from user questions;
[0037] A retrieval module, used to perform knowledge retrieval based on the text content to obtain knowledge retrieval results;
[0038] A GPT model processing module, used for inputting the text content and the knowledge retrieval results into a preset GPT model to obtain a question-answer pair;
[0039] An evaluation module, used to perform a quality evaluation on the question-answer pair to determine whether the question-answer pair meets a preset condition;
[0040] An output module, configured to output the question-answer pair as a reply content corresponding to the user question if the preset condition is met;
[0041] The operation and maintenance module is used to display the question-answer pair on the operation and maintenance interface if the preset conditions are not met, receive the operation and maintenance instructions carrying operation and maintenance data triggered by the operation and maintenance interface, generate and output the reply content corresponding to the user question according to the operation and maintenance data.
[0042] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:
[0043] A computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the intelligent operation and maintenance method based on the self-developed GPT model as described above are implemented.
[0044] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:
[0045] A computer-readable storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by a processor, implement the steps of the intelligent operation and maintenance method based on a self-developed GPT model as described above.
[0046] Compared with the prior art, the embodiments of the present application have the following beneficial effects:
[0047] The intelligent operation and maintenance method based on the self-developed GPT model disclosed in this application extracts text content from user questions; and performs knowledge retrieval based on the text content to obtain knowledge retrieval results; then inputs the text content and the knowledge retrieval results into the preset GPT model to obtain a question-answer pair; then, the question-answer pair is quality evaluated to determine whether the question-answer pair meets the preset conditions; if it meets the preset conditions, the question-answer pair is output as the reply content corresponding to the user question; if it does not meet the preset conditions, the question-answer pair is displayed on the operation and maintenance interface, and the operation and maintenance instructions carrying operation and maintenance data triggered by the operation and maintenance interface are received, and the reply content corresponding to the user question is generated and output according to the operation and maintenance data. This application automatically answers user questions in the operation and maintenance process with the help of a self-developed GPT model, and then evaluates the answers generated by the GPT model to determine whether secondary operation and maintenance is required. While ensuring the accuracy of the answers, it reduces the dependence of the operation and maintenance process on manual labor and improves the efficiency and intelligence of operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. 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 any creative work.
[0049] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;
[0050] Figure 2 It is a flowchart of an embodiment of an intelligent operation and maintenance method based on a self-developed GPT model according to the present application;
[0051] Figure 3 It is a structural diagram of an embodiment of an intelligent operation and maintenance system based on a self-developed GPT model according to the present application;
[0052] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION
[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.
[0054] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0055] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.
[0056] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.
[0057] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.
[0058] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 (Moving Picture Experts Group Audio Layer III) players, MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.
[0059] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .
[0060] It should be noted that the intelligent operation and maintenance method based on the self-developed GPT model provided in the embodiment of the present application is generally executed by a server. Accordingly, the intelligent operation and maintenance system based on the self-developed GPT model is generally set in the server.
[0061] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.
[0062] Continue to refer Figure 2 , shows a flow chart of an embodiment of an intelligent operation and maintenance method based on a self-developed GPT model according to the present application. The intelligent operation and maintenance method based on a self-developed GPT model comprises the following steps:
[0063] Step S201, extracting text content from user questions.
[0064] In this embodiment, the intelligent operation and maintenance method based on the self-developed GPT model runs on the electronic device (for example Figure 1 The server shown in the figure) can send or receive data through a wired connection or a wireless connection. It should be noted that the above wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.
[0065] It should be noted that with the development of financial technology, its related industries such as insurance and banking have had large and small innovations in financial technology, including but not limited to channel innovation, product innovation and service innovation based on the Internet. For each company in the insurance and banking industries, the future development direction is to combine financial technology innovations on the basis of business scenarios in order to achieve better results.
[0066] In this embodiment, in order to realize intelligent operation and maintenance based on the self-developed GPT model, it is first necessary to extract text content from user questions. Specifically, user questions are composed of text, which can be a piece of text, or exist in documents, such as HTML documents, PDF documents, word documents, etc., which generally come from "daily consultation" and "operation and maintenance work orders". For each enterprise, especially those involved in financial technology, in the process of business development or iteration, there will be various user problems. The purpose of operation and maintenance is to solve these user problems to ensure the long-term stability of the business system. The NLP algorithm can extract text content from each user question, such as named entity recognition, relationship extraction, event extraction, etc. The extracted text content represents the key information of the user problem, so that intelligent operation and maintenance can be realized according to the text content in the future.
[0067] Step S202: perform knowledge retrieval based on the text content to obtain knowledge retrieval results.
[0068] It should be noted that a knowledge base is a database that centrally stores knowledge. The database can be used to easily create, organize, find and share knowledge. It may include FAQs, manuals, troubleshooting guides, operation manuals, and other information that the team may want or need to know, which helps the team manage knowledge.
[0069] In this embodiment, after extracting the text content from the user's question, knowledge retrieval can be performed based on the text content to obtain knowledge retrieval results. Specifically, in enterprises involved in financial technology, there is generally a knowledge base, and the information in the knowledge base can be uploaded manually or automatically entered by the system. For example, the question-answer pairs generated by the GPT model in the subsequent steps of this application can also be automatically entered into the knowledge base. After extracting the text content from the user's question, you can search in the preset knowledge base based on the text content to determine whether there is reference information as a knowledge retrieval result. These reference information can be knowledge points that have a certain correlation with the text content.
[0070] Step S203: input the text content and the knowledge retrieval results into a preset GPT model to obtain a question-answer pair.
[0071] In this embodiment, after obtaining the knowledge retrieval results, the text content and the knowledge retrieval results can be input into the preset GPT model to obtain a question-answer pair. Specifically, the essence of the GPT model is a generative pre-trained language model based on Transformer, and the preset GPT is the GPT model developed by the enterprise itself. The GPT models developed by various enterprises may be different. This embodiment takes PingAnGPT as an example. According to the knowledge retrieval results, it can be determined whether there is reference information in the knowledge base. If there is reference information, the text content and the reference information can be input into PingAnGPT, and PingAnGPT is called to output the question-answer pair; if there is no reference information, the text content can be input into PingAnGPT, and PingAnGPT is called to directly output the question-answer pair.
[0072] It is understandable that the specific knowledge points stored in the knowledge base can be in the form of a traditional database table or in the form of Key-Value, and the question-answer pair is equivalent to the form of Key-Value, which is helpful for subsequent entry into the knowledge base.
[0073] Step S204: perform a quality assessment on the question-answer pair to determine whether the question-answer pair meets a preset condition.
[0074] In this embodiment, after the text content and knowledge retrieval results are input into the preset GPT and the question-answer pair is obtained, the question-answer pair needs to be quality evaluated to determine whether the question-answer pair meets the preset conditions. Specifically, in order to avoid the repeated generation of some question-answer pairs affecting the intelligence of operation and maintenance, after the question-answer pair is obtained, it is necessary to conduct a quality evaluation on it and screen and remove duplicates by comparing the similarity with the existing knowledge. Therefore, a similarity threshold can be set in advance, and the similarity between the knowledge points in the knowledge base and the question-answer pair is calculated to determine whether there are knowledge points with a similarity greater than or equal to the similarity threshold as similar knowledge points; if there are no similar knowledge points, it is considered that the question-answer pair meets the preset conditions; if there are similar knowledge points, it is considered that the question-answer pair does not meet the preset conditions; subsequently, the corresponding operation and maintenance process can be carried out according to whether the question-answer pair meets the preset conditions.
[0075] Step S205: if the preset condition is met, the question-answer pair is output as the reply content corresponding to the user question.
[0076] In this embodiment, the quality of the question-answer pair is evaluated. If it is determined that the question-answer pair meets the preset conditions, the question-answer pair is output as the reply content corresponding to the user's question. Specifically, the quality evaluation of the question-answer pair can determine whether there are similar knowledge points. If the question-answer pair meets the preset conditions, it means that there are no similar knowledge points, and it is a question-answer pair that is not included in the knowledge base. No corresponding reference information can be found in the knowledge base, so it can be directly used as the reply content of the user's question and output.
[0077] Step S206: If the preset condition is not met, the question-answer pair is displayed on the operation and maintenance interface, an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface is received, and a reply content corresponding to the user question is generated and output according to the operation and maintenance data.
[0078] In this embodiment, the quality of the question-answer pair is evaluated. If it is determined that the question-answer pair does not meet the preset conditions, the question-answer pair is displayed on the operation and maintenance interface, and the operation and maintenance instructions carrying the operation and maintenance data triggered by the operation and maintenance interface are received. The reply content corresponding to the user's question is generated and output according to the operation and maintenance data. Specifically, the quality evaluation of the question-answer pair can determine whether there are similar knowledge points. If the question-answer pair does not meet the preset conditions, it means that there are similar knowledge points, and manual intervention by the operation and maintenance personnel is required. The question-answer pair and similar knowledge points can be displayed on the operation and maintenance interface, and then the operation and maintenance instructions carrying the operation and maintenance data triggered by the operation and maintenance interface are received. By parsing the operation and maintenance data, the reply suggestions entered by the operation and maintenance personnel in the operation and maintenance interface can be obtained, and finally the reply content corresponding to the user's question is generated and output according to the reply suggestions.
[0079] This application uses a self-developed GPT model to automatically answer user questions during the operation and maintenance process, and then evaluates the answers generated by the GPT model to determine whether secondary operation and maintenance is required. While ensuring the accuracy of the answers, it reduces the dependence of the operation and maintenance process on manual labor and improves the efficiency and intelligence of operation and maintenance.
[0080] In some optional implementations of this embodiment, the step of performing knowledge retrieval according to the text content to obtain knowledge retrieval results includes:
[0081] A search is performed in a preset knowledge base according to the text content to determine whether there is reference information whose relevance to the text content is greater than or equal to a preset association threshold value as the knowledge search result.
[0082] In this embodiment, after extracting the text content from the user's question, a search can be performed in the knowledge base based on the text content to determine whether there is reference information whose relevance to the text content is greater than or equal to a preset relevance threshold as a knowledge retrieval result. Specifically, a large amount of information is stored in the knowledge base of the enterprise, which may include FAQs, manuals, troubleshooting guides, operation manuals, and other information that the team may want or need to know. This information may have a certain relevance to the extracted text content. Therefore, a preset relevance threshold can be set in advance, and a search can be performed in the knowledge base based on the text content. It can be determined whether there is information whose relevance is greater than or equal to the threshold by calculating the text relevance. If so, it can be regarded as matching the text content and the information is used as reference information; if not, there is no information matching the text content in the knowledge base, that is, there is no reference information; and whether there is reference information or not, it can be used as a knowledge retrieval result for subsequent generation of question-answer pairs through the self-developed GPT model.
[0083] This application determines whether there is reference information in the text content extracted from the user's question through knowledge retrieval, ensures the accuracy of the answers obtained through the self-developed GPT model, and improves the intelligence of operation and maintenance.
[0084] In some optional implementations, the step of inputting the text content and the knowledge retrieval results into a preset GPT model to obtain a question-answer pair includes:
[0085] If the knowledge retrieval result indicates that the reference information exists, inputting the text content and the reference information into the GPT model, and calling the GPT model to output the question-answer pair;
[0086] If the knowledge retrieval result is that the reference information does not exist, the text content is input into the GPT model, and the GPT model is called to output the question-answer pair.
[0087] In this embodiment, it can be determined whether there is reference information in the text content according to the knowledge retrieval result. After obtaining the knowledge retrieval result, if the knowledge retrieval result is that there is reference information, the text content and the reference information are input into the preset GPT model, and the GPT model is called to output the question-answer pair; if the knowledge retrieval result is that there is no reference information, the text content is input into the preset GPT model, and the GPT model is called to output the question-answer pair. Specifically, PingAnGPT is a self-developed GPT model. Using it as a preset GPT model, when there is reference information corresponding to the text content, PingAnGPT generates and outputs question-answer pairs according to the reference information and the text content; when there is no reference information corresponding to the text content, PingAnGPT directly generates and outputs question-answer pairs according to the text content. It can be understood that for PingAnGPT, the reference information can serve as a reference, and the question-answer pairs generated according to the text content and the reference information should have improved accuracy compared to the directly generated question-answer pairs.
[0088] This application distinguishes whether the text content input to the GPT model contains reference information, and calls the GPT model to generate answers in different ways, thereby ensuring the accuracy of the answers obtained through the self-developed GPT model, reducing the dependence of the operation and maintenance process on manual labor, and improving the efficiency and intelligence of operation and maintenance.
[0089] In some optional implementations, the step of performing quality assessment on the question-answer pair to determine whether the question-answer pair meets a preset condition includes:
[0090] Calculating the similarity between the knowledge points in the knowledge base and the question-answer pair, and determining whether there are similar knowledge points whose similarity is greater than or equal to a preset similarity threshold;
[0091] If the similar knowledge point does not exist, determining that the question-answer pair meets the preset condition;
[0092] If the similar knowledge point exists, it is determined that the question-answer pair does not meet the preset condition.
[0093] It should be noted that the enterprise's knowledge base stores a large amount of information, which may include FAQs, manuals, troubleshooting guides, operation manuals, and other information that the team may want or need to know, and knowledge points are formed by summarizing and summarizing this information, which can be stored in a systematic manner.
[0094] In this embodiment, after obtaining the question-answer pair, it is also necessary to perform a quality assessment on the question-answer pair to determine whether the question-answer pair meets the preset conditions. By calculating the similarity between the knowledge point in the knowledge base and the question-answer pair, it can be determined whether there are similar knowledge points with a similarity greater than or equal to the preset similarity threshold; if there are no similar knowledge points, it is determined that the question-answer pair meets the preset conditions; if there are similar knowledge points, it is determined that the question-answer pair does not meet the preset conditions. Specifically, in order to avoid affecting the intelligence of operation and maintenance, it is necessary to perform a quality assessment on the question-answer pair to screen and remove duplicates. Therefore, a similarity threshold can be set in advance, and an NLP algorithm can be used to determine whether there are knowledge points in the knowledge base with a similarity greater than or equal to the similarity threshold through cosine similarity calculation, TF-IDF calculation, or semantic similarity calculation. If there are, the knowledge point is regarded as a similar knowledge point, and it is considered that the question-answer pair does not meet the preset conditions; if not, it can be considered that the question-answer pair meets the preset conditions; so that the subsequent operation and maintenance process that needs to be performed can be determined based on whether the question-answer pair meets the preset conditions.
[0095] This application evaluates the quality of the answers generated by the GPT model through text similarity, so that duplicates can be screened and removed to determine whether secondary operation and maintenance is required, thereby ensuring the accuracy of the answers and improving the efficiency and intelligence of operation and maintenance.
[0096] In some optional implementations, if the preset condition is not met, the step of displaying the question-answer pair on the operation and maintenance interface, receiving an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface, and generating and outputting a reply content corresponding to the user question according to the operation and maintenance data includes:
[0097] If the preset conditions are not met, the question-answer pair and the similar knowledge points are displayed on the operation and maintenance interface;
[0098] Receiving the operation and maintenance instruction and analyzing the operation and maintenance data to obtain a reply suggestion;
[0099] Identify the question type of the question-answer pair, and determine the response template corresponding to the question type based on the pre-stored correspondence between different question types and different response templates;
[0100] The reply template is filled according to the reply suggestion, and the reply content is generated and output.
[0101] In this embodiment, when performing quality assessment on question-answer pairs, if the question-answer pairs do not meet preset conditions, the question-answer pairs and similar knowledge points can be displayed on the operation and maintenance interface, and operation and maintenance instructions carrying operation and maintenance data triggered by the operation and maintenance interface are received, and the operation and maintenance data is parsed to obtain reply suggestions, identify the question type of the question-answer pair, determine the reply template corresponding to the question type based on the correspondence between different pre-stored question types and different reply templates, fill in the reply template according to the reply suggestion, generate reply content and output it. Specifically, by classifying the questions in advance, various question types can be obtained, and different question types correspond to different reply templates. The operation and maintenance interface is an interface that can be operated by the operation and maintenance personnel and can be displayed on the user end of the operation and maintenance personnel. When the question-answer pair does not meet the preset conditions, that is, there are similar knowledge points, it is determined that the question-answer pair has not passed the quality assessment and needs to be manually intervened by the operation and maintenance personnel. At this time, the question-answer pair and similar knowledge points can be displayed on the operation and maintenance interface. In addition, prompt information, related annotations, etc. can also be displayed, and the operation and maintenance personnel can manually enter the operation and maintenance data on the operation and maintenance interface and trigger the operation and maintenance instructions. When the operation and maintenance instructions carrying the operation and maintenance data triggered by the operation and maintenance interface are received, the operation and maintenance data can be parsed to obtain the reply suggestions entered by the operation and maintenance personnel on the operation and maintenance interface, and then the question type of the question-answer pair is identified, and the reply template corresponding to the question type is determined according to the correspondence between different pre-stored question types and different reply templates, and then the reply suggestions are filled in according to the reply template. Finally, the reply content corresponding to the user's question can be obtained and output.
[0102] When the quality assessment of the answer generated by the GPT model fails, the application can be manually operated for secondary maintenance to generate and output the reply content in a timely manner based on the reply suggestions, thereby ensuring the accuracy of the answer and improving the efficiency of operation and maintenance.
[0103] In some optional implementations, after the steps of displaying the question-answer pair on the operation and maintenance interface if the preset condition is not met, receiving an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface, and generating and outputting a reply content corresponding to the user question according to the operation and maintenance data, the method further includes:
[0104] Recording the operation and maintenance data;
[0105] Performing statistical analysis on the operation and maintenance data to obtain quasi-question-answer pairs, and recording the number of occurrences of the quasi-question-answer pairs;
[0106] When the number of occurrences reaches a preset number, the quasi-question-answer pair is stored in the knowledge base.
[0107] In this embodiment, after generating and outputting the reply content corresponding to the user's question based on the operation and maintenance data carried by the operation and maintenance quality, the operation and maintenance data can also be recorded, and the operation and maintenance data can be statistically analyzed to obtain a quasi-question-answer pair, and the number of occurrences of the quasi-question-answer pair can be recorded at the same time. When the number of occurrences reaches a preset number, the quasi-question-answer pair is stored in the knowledge base. Specifically, since user questions generally come from "daily consultation" and "operation and maintenance work orders", when the reply content corresponding to the user's question is obtained and output, it can be considered that the user's question has been solved and / or the operation and maintenance work order has been completed. At this time, the operation and maintenance data can also be recorded and statistically analyzed based on the operation and maintenance data to obtain a quasi-question-answer pair, and the number of occurrences of the quasi-question-answer pair can be recorded. A number is pre-set, such as 3 times. When the number of occurrences of the quasi-question-answer pair is greater than or equal to 3 times, the quasi-question-answer pair is stored in the knowledge base, thereby updating the knowledge base. If knowledge retrieval is performed later and the quasi-question-answer pair is retrieved, it can be directly output as a question-answer pair.
[0108] Optionally, statistical analysis may be performed regularly based on the recorded operation and maintenance data at preset time intervals to achieve regular updating of the knowledge base. This operation may be performed offline.
[0109] This application realizes the entry and update of the knowledge base through statistical analysis of operation and maintenance data, which can increase the knowledge points in the knowledge base, making the subsequent intelligent operation and maintenance with the help of the self-developed GPT model more accurate, thereby improving the efficiency and intelligence of operation and maintenance.
[0110] In some optional implementations, the step of extracting text content from the user question includes:
[0111] Identify the document type corresponding to the user question;
[0112] Text extraction is performed according to the document type to obtain the text content.
[0113] In this embodiment, in order to realize intelligent operation and maintenance based on the self-developed GPT model, it is first necessary to extract text content from user questions, identify the document type corresponding to the user question, and perform text extraction according to the document type to obtain the text content. Specifically, the user question is composed of text, which can be a piece of text, or exists in a document, such as an html document, a pdf document, a word document, etc. In order to extract the text content, it is first necessary to identify the document type corresponding to the user question, and then perform text extraction according to the document type. Different document types have different ways of text extraction. For example, for documents in html format, after text extraction, the html tags can be removed and the text part can be retained as the text content. Generally speaking, the NLP algorithm is used to extract the text content, such as named entity recognition, relationship extraction, event extraction, etc., and intelligent operation and maintenance can be realized based on the extracted text content later.
[0114] This application extracts text through user questions of different document types to ensure accurate text content, so that the subsequent intelligent operation and maintenance with the help of the self-developed GPT model will be more accurate, thereby improving the efficiency and intelligence of operation and maintenance.
[0115] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.
[0116] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.
[0117] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).
[0118] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.
[0119] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of an intelligent operation and maintenance system based on a self-developed GPT model. Figure 2 Corresponding to the method embodiment shown, the system can be specifically applied to various electronic devices.
[0120] like Figure 3 As shown, the intelligent operation and maintenance system 300 based on the self-developed GPT model described in this embodiment includes: an extraction module 301, a retrieval module 302, a GPT model processing module 303, an evaluation module 304, an output module 305 and an operation and maintenance module 306. Among them:
[0121] Extraction module 301, used to extract text content from user questions;
[0122] A retrieval module 302 is used to perform knowledge retrieval according to the text content to obtain knowledge retrieval results;
[0123] A GPT model processing module 303 is used to input the text content and the knowledge retrieval result into a preset GPT model to obtain a question-answer pair;
[0124] An evaluation module 304 is used to perform a quality evaluation on the question-answer pair to determine whether the question-answer pair meets a preset condition;
[0125] Output module 305, used for outputting the question-answer pair as a reply content corresponding to the user question if the preset condition is met;
[0126] The operation and maintenance module 306 is used to display the question-answer pair on the operation and maintenance interface if the preset condition is not met, receive the operation and maintenance instructions carrying operation and maintenance data triggered by the operation and maintenance interface, generate and output the reply content corresponding to the user question according to the operation and maintenance data.
[0127] The intelligent operation and maintenance system based on the self-developed GPT model provided in this application automatically answers user questions in the operation and maintenance process with the help of the self-developed GPT model, and then evaluates the answers generated by the GPT model to determine whether secondary operation and maintenance is required. While ensuring the accuracy of the answers, it reduces the dependence of the operation and maintenance process on manual labor and improves the efficiency and intelligence of operation and maintenance.
[0128] In some optional implementations of this embodiment, the retrieval module 302 is further used to:
[0129] A search is performed in a preset knowledge base according to the text content to determine whether there is reference information whose relevance to the text content is greater than or equal to a preset association threshold value as the knowledge search result.
[0130] The intelligent operation and maintenance system based on the self-developed GPT model provided in this application determines whether there is reference information in the text content extracted from the user's question through knowledge retrieval, thereby ensuring the accuracy of the answers subsequently obtained through the self-developed GPT model and improving the intelligence of operation and maintenance.
[0131] In some optional implementations of this embodiment, the GPT model processing module 303 is further used to:
[0132] identifying the knowledge retrieval result;
[0133] If the knowledge retrieval result indicates that the reference information exists, inputting the text content and the reference information into the GPT model, and calling the GPT model to output the question-answer pair;
[0134] If the knowledge retrieval result is that the reference information does not exist, the text content is input into the GPT model, and the GPT model is called to output the question-answer pair.
[0135] The intelligent operation and maintenance system based on the self-developed GPT model provided by the present application ensures the accuracy of the answers obtained by the self-developed GPT model by distinguishing whether the text content input into the GPT model contains reference information, and calling the GPT model to generate answers in different ways, thereby reducing the dependence of the operation and maintenance process on manual labor and improving the efficiency and intelligence of the operation and maintenance.
[0136] In some optional implementations of this embodiment, the evaluation module 304 is further configured to:
[0137] Calculating the similarity between the knowledge points in the knowledge base and the question-answer pair, and determining whether there are similar knowledge points whose similarity is greater than or equal to a preset similarity threshold;
[0138] If the similar knowledge point does not exist, determining that the question-answer pair meets the preset condition;
[0139] If the similar knowledge point exists, it is determined that the question-answer pair does not meet the preset condition.
[0140] The intelligent operation and maintenance system based on the self-developed GPT model provided in this application performs quality assessment on the answers generated by the GPT model through text similarity, so that duplicate answers can be screened and removed to determine whether secondary operation and maintenance is required, thereby ensuring the accuracy of the answers and improving the efficiency and intelligence of operation and maintenance.
[0141] In some optional implementations of this embodiment, the operation and maintenance module 306 is further used to:
[0142] If the preset conditions are not met, the question-answer pair and the similar knowledge points are displayed on the operation and maintenance interface;
[0143] Receiving the operation and maintenance instruction and analyzing the operation and maintenance data to obtain a reply suggestion;
[0144] Identify the question type of the question-answer pair, and determine the response template corresponding to the question type based on the pre-stored correspondence between different question types and different response templates;
[0145] The reply template is filled according to the reply suggestion, and the reply content is generated and output.
[0146] The intelligent operation and maintenance system based on the self-developed GPT model provided in this application can be manually operated for secondary operation and maintenance when the quality assessment of the answer generated by the GPT model fails, and the reply content is generated and output in time according to the reply suggestions, thereby ensuring the accuracy of the answer and improving the efficiency of operation and maintenance.
[0147] In some optional implementations of this embodiment, the intelligent operation and maintenance system 300 based on the self-developed GPT model is also used for:
[0148] Recording the operation and maintenance data;
[0149] Performing statistical analysis on the operation and maintenance data to obtain quasi-question-answer pairs, and recording the number of occurrences of the quasi-question-answer pairs;
[0150] When the number of occurrences reaches a preset number, the quasi-question-answer pair is stored in the knowledge base.
[0151] The intelligent operation and maintenance system based on the self-developed GPT model provided in the present application can realize the entry and update of the knowledge base through statistical analysis of the operation and maintenance data, thereby increasing the knowledge points in the knowledge base, making the subsequent intelligent operation and maintenance with the help of the self-developed GPT model more accurate, thereby improving the efficiency and intelligence of the operation and maintenance.
[0152] In some optional implementations of this embodiment, the extraction module 301 is further used to:
[0153] Identify the document type corresponding to the user question;
[0154] Text extraction is performed according to the document type to obtain the text content.
[0155] The intelligent operation and maintenance system based on the self-developed GPT model provided in this application extracts text through user questions of different document types to ensure accurate text content, so that the subsequent intelligent operation and maintenance with the help of the self-developed GPT model is more accurate, thereby improving the efficiency and intelligence of operation and maintenance.
[0156] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.
[0157] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.
[0158] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.
[0159] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 4. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of the intelligent operation and maintenance method based on the self-developed GPT model. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.
[0160] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer-readable instructions or process data stored in the memory 41, such as computer-readable instructions for running the intelligent operation and maintenance method based on the self-developed GPT model.
[0161] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.
[0162] The computer equipment provided in this application automatically answers user questions during the operation and maintenance process with the help of a self-developed GPT model, and then evaluates the answers generated by the GPT model to determine whether secondary operation and maintenance is required. While ensuring the accuracy of the answers, it reduces the dependence of the operation and maintenance process on manual labor and improves the efficiency and intelligence of operation and maintenance.
[0163] The present application also provides another implementation, namely, providing a computer-readable storage medium, which stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the intelligent operation and maintenance method based on the self-developed GPT model as described above.
[0164] The computer-readable storage medium provided in this application automatically answers user questions in the operation and maintenance process by using a self-developed GPT model, and then evaluates the answers generated by the GPT model to determine whether secondary operation and maintenance is required. While ensuring the accuracy of the answers, it reduces the dependence of the operation and maintenance process on manual labor and improves the efficiency and intelligence of operation and maintenance.
[0165] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.
[0166] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.
Claims
1. An intelligent operation and maintenance method based on the self-developed GPT model, It is characterized in that The steps include: Extract text content from user questions; Perform knowledge retrieval according to the text content to obtain knowledge retrieval results; Input the text content and the knowledge retrieval results into a preset GPT model to obtain a question-answer pair; Performing a quality assessment on the question-answer pair to determine whether the question-answer pair meets a preset condition; If the preset condition is met, the question-answer pair is output as the reply content corresponding to the user question; If the preset condition is not met, the question-answer pair is displayed on the operation and maintenance interface, an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface is received, and a reply content corresponding to the user question is generated and output according to the operation and maintenance data.
2. According to the intelligent operation and maintenance method based on the self-developed GPT model according to claim 1, It is characterized in that The step of performing knowledge retrieval according to the text content to obtain knowledge retrieval results specifically includes: A search is performed in a preset knowledge base according to the text content to determine whether there is reference information whose relevance to the text content is greater than or equal to a preset association threshold value as the knowledge search result.
3. According to the intelligent operation and maintenance method based on the self-developed GPT model according to claim 2, It is characterized in that The step of inputting the text content and the knowledge retrieval result into a preset GPT model to obtain a question-answer pair specifically includes: identifying the knowledge retrieval result; If the knowledge retrieval result indicates that the reference information exists, inputting the text content and the reference information into the GPT model, and calling the GPT model to output the question-answer pair; If the knowledge retrieval result is that the reference information does not exist, the text content is input into the GPT model, and the GPT model is called to output the question-answer pair.
4. According to the intelligent operation and maintenance method based on the self-developed GPT model according to claim 1, It is characterized in that The step of performing quality assessment on the question-answer pair to determine whether the question-answer pair meets the preset conditions specifically includes: Calculating the similarity between the knowledge points in the knowledge base and the question-answer pair, and determining whether there are similar knowledge points whose similarity is greater than or equal to a preset similarity threshold; If the similar knowledge point does not exist, determining that the question-answer pair meets the preset condition; If the similar knowledge point exists, it is determined that the question-answer pair does not meet the preset condition.
5. According to the intelligent operation and maintenance method based on the self-developed GPT model according to claim 4, It is characterized in that If the preset condition is not met, the question-answer pair is displayed on the operation and maintenance interface, an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface is received, and a reply content corresponding to the user question is generated according to the operation and maintenance data and outputted, specifically including: If the preset conditions are not met, the question-answer pair and the similar knowledge points are displayed on the operation and maintenance interface; Receiving the operation and maintenance instruction and analyzing the operation and maintenance data to obtain a reply suggestion; Identify the question type of the question-answer pair, and determine the response template corresponding to the question type based on the pre-stored correspondence between different question types and different response templates; The reply template is filled according to the reply suggestion, and the reply content is generated and output.
6. According to the intelligent operation and maintenance method based on the self-developed GPT model according to claim 1, It is characterized in that After the steps of displaying the question-answer pair on the operation and maintenance interface if the preset condition is not met, receiving an operation and maintenance instruction carrying operation and maintenance data triggered by the operation and maintenance interface, and generating and outputting a reply content corresponding to the user question according to the operation and maintenance data, the method further includes: Recording the operation and maintenance data; Performing statistical analysis on the operation and maintenance data to obtain quasi-question-answer pairs, and recording the number of occurrences of the quasi-question-answer pairs; When the number of occurrences reaches a preset number, the quasi-question-answer pair is stored in the knowledge base.
7. According to claim 1, the intelligent operation and maintenance method based on the self-developed GPT model, It is characterized in that The step of extracting text content from the user question specifically includes: Identify the document type corresponding to the user question; Text extraction is performed according to the document type to obtain the text content.
8. An intelligent operation and maintenance system based on the self-developed GPT model, It is characterized in that include: The extraction module is used to extract text content from user questions; A retrieval module, used to perform knowledge retrieval based on the text content to obtain knowledge retrieval results; A GPT model processing module, used for inputting the text content and the knowledge retrieval results into a preset GPT model to obtain a question-answer pair; An evaluation module, used to perform a quality evaluation on the question-answer pair to determine whether the question-answer pair meets a preset condition; An output module, configured to output the question-answer pair as a reply content corresponding to the user question if the preset condition is met; The operation and maintenance module is used to display the question-answer pair on the operation and maintenance interface if the preset conditions are not met, receive the operation and maintenance instructions carrying operation and maintenance data triggered by the operation and maintenance interface, generate and output the reply content corresponding to the user question according to the operation and maintenance data.
9. A computer device, It is characterized in that It includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the intelligent operation and maintenance method based on the self-developed GPT model as described in any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the intelligent operation and maintenance method based on the self-developed GPT model as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Implementation method and implementation equipment for multi-queue help system supporting interclass transfer
CN101588321A
Method and system for mixed customer services of intelligent robots and human beings
CN105591882A
Method and device for customer services, electronic device and computer readable storage medium
CN110069607A
Intelligent dialogue method, system and equipment based on large model and local knowledge base
CN116992005A
Large model question and answer method and device based on domain knowledge
CN117609477A
Cited By
5G private network operation and maintenance method based on large model adaptive capability and related device
CN120264329A
