Large model-based campus intelligent customer service system and processing method

Abstract

The invention provides a campus intelligent customer service system based on a large model and a processing method thereof. The method comprises the following steps: receiving a question input by a campus user, and preprocessing the question to obtain a preprocessed question in a text form; inputting the pre-processed question in the text form into a large language model; the big language model performs semantic recognition according to the received question in the text form, mines a word and sentence relationship, extracts a subject word, and retrieves an answer in the campus knowledge base; if the answer can be retrieved, generating a reply answer by using the answer and a preset Prompt template, and providing the reply answer to the campus user; and if the answer cannot be retrieved, creating a to-be-processed work order according to the identified semantic information, and sending the to-be-processed work order to a corresponding campus administrative department for processing. According to the invention, various campus life problems can be solved more intelligently.

Description

Technical Field

[0001] One or more embodiments of the present invention relate to network communication technology, and more particularly to a campus intelligent customer service system and processing method based on a large model. Background Art

[0002] At present, the operation and maintenance work of campus is becoming increasingly complicated. Whether it is primary and secondary schools or university campuses, there are various operation and maintenance work in campus life. For example, a large number of students, especially freshmen, do not know or will forget the login address / login method of the learning software "Xuexitong"; for another example, students often need to ask for the phone number of the campus network information center; for another example, students often need to report various problems, such as the poor hygiene of the cafeteria and the dishes are often found to be not clean, but students do not know who to report to or how to report.

[0003] With the existing technology, students cannot find an intelligent portal to solve various campus life problems in campus life. If they ask relevant personnel from the school to solve the problems, there are problems such as lack of intelligence, time-consuming, labor-intensive, and inefficient. Summary of the invention

[0004] One or more embodiments of the present invention describe a campus intelligent customer service system and processing method based on a large model, which can solve at least one problem in the prior art.

[0005] According to a first aspect, a campus intelligent customer service system based on a large model is provided, the system comprising: a user input processing module, a large model docking processing module, a large language model, a work order processing module, and a campus knowledge base; wherein,

[0006] The user input processing module is used to receive questions input by campus users, pre-process the questions, and obtain the pre-processed questions in text form;

[0007] The large model docking processing module is used to input the pre-processed text-based questions into the large language model; and provide the reply answers sent by the large language model to the campus user;

[0008] The large language model is used to perform semantic recognition on the received questions, mine word-sentence relationships, extract keywords, and search for answers in the campus knowledge base. If an answer can be retrieved, a reply answer is generated using the answer and a pre-set prompt template, and the reply answer is sent to the large model docking processing module. If an answer cannot be retrieved, the work order processing module is triggered to process according to the recognized semantic information;

[0009] The work order processing module is used to create a work order to be processed after being triggered by the large language model, and send the work order to be processed to the corresponding campus administrative department.

[0010] The user input processing module comprises:

[0011] The text input information processing submodule is used to receive text format questions input by campus users, and perform at least one of corpus cleaning, word segmentation, semantic error correction and part-of-speech tagging on the text format questions, so as to obtain pre-processed text format questions;

[0012] and / or,

[0013] The voice input information processing submodule is used to receive voice-formatted questions input by campus users, convert the voice-formatted questions into text-formatted questions, and then perform at least one of corpus cleaning, word segmentation, semantic error correction, and part-of-speech tagging on the text-formatted questions, so as to obtain pre-processed text-formatted questions.

[0014] The system further includes: a general knowledge base;

[0015] The large language model is further used to search in the general knowledge base when the answer cannot be retrieved in the campus knowledge base and before the identified semantic information is sent to the work order processing module. If the answer can be retrieved in the general knowledge base, a reply answer is generated using the answer and a pre-set Prompt template, and the reply answer is sent to the large model docking processing module. If the answer cannot be retrieved in the general knowledge base, the process of sending the identified semantic information to the work order processing module continues.

[0016] When the large language model executes the triggering of the work order processing module for processing based on the identified semantic information, it is specifically configured to execute: judging whether the user has a complaint intention and / or expresses dissatisfaction or complaint about the campus based on the identified semantic information; if so, using natural language processing capabilities to extract information related to the intention input by the user, and sending the extracted information related to the intention to the work order processing module.

[0017] The system further includes: at least one administrative department processing module within the campus, different administrative department processing modules belonging to different campus administrative departments;

[0018] Each administrative department processing module is used to perform the processing required by the campus user after receiving the work order to be processed sent by the work order processing module, and send the work order processing result to the work order processing module;

[0019] The work order processing module is further used to provide the work order processing result to the campus user after receiving the work order processing result.

[0020] The system further includes: a security and privacy module;

[0021] The large model docking processing module sends the reply answer sent by the large language model to the security and privacy module before providing the reply answer to the campus user; when the received evaluation result is that the answer does not include privacy information, the reply answer is provided to the campus user, otherwise, the answer is refused to be provided to the campus user.

[0022] The security and privacy module is used to check the received reply answers, evaluate whether the answers include privacy information, and notify the large model docking processing module of the evaluation results.

[0023] According to various embodiments of the present invention, a processing method for a campus intelligent customer service system based on a large model is provided, and the method includes:

[0024] receiving questions input by campus users, preprocessing the questions, and obtaining the questions in a preprocessed text form;

[0025] Input the preprocessed text-based questions into the large language model;

[0026] The large language model uses the received questions to perform semantic recognition, mine word-sentence relationships, extract subject terms, and retrieve answers from the campus knowledge base;

[0027] If the answer can be retrieved, a reply answer is generated using the answer and a preset prompt template, and the reply answer is provided to the campus user;

[0028] If the answer cannot be retrieved, a pending work order is created based on the identified semantic information and sent to the corresponding campus administrative department.

[0029] The step of creating a work order to be processed according to the identified semantic information includes:

[0030] Based on the identified semantic information, it is determined whether the user has the intention to complain and / or expresses dissatisfaction or complaints about the campus. If so, the large language model uses natural language processing capabilities to extract information related to the intention input by the user, and sends the extracted information related to the intention to the work order processing module, which creates a work order to be processed.

[0031] Different administrative department processing modules belong to different campus administrative departments;

[0032] The method further includes: after each administrative department processing module receives the work order to be processed sent by the work order processing module, it performs the processing required by the campus user and sends the work order processing result to the work order processing module; after receiving the work order processing result, the work order processing module further provides the work order processing result to the campus user.

[0033] The campus intelligent customer service system and processing method based on a large model provided by various embodiments of the present invention have at least the following beneficial effects:

[0034] 1. In campus life, students can find a unified intelligent entrance that can solve various campus life problems. Students only need to input their own questions about campus life into the user input processing module of the system of the present invention, and they can get the relevant answers output by the system of the present invention. Students do not need to find relevant personnel of the school to solve the problems. The campus intelligent customer service solution can be realized in an intelligent way, saving time, effort and efficiency.

[0035] 2. Based on the system of the present invention, a campus intelligent customer service system solution is implemented, combining the large language model with the campus knowledge base. The large language model can grasp the inherent laws and contextual relationships of the language through deep learning and natural language processing technology, so that it can make more accurate and reasonable responses when faced with complex and changeable language input. The campus knowledge base stores a wealth of campus management knowledge and experience, providing data support for the large language model. Combining the large language model with the campus knowledge base can enable the system to understand user needs more accurately and provide more accurate management services. It effectively improves the answer effect of the model in complex scenarios, ensuring the accuracy of information transmission and user satisfaction. The system of the present invention, with its high efficiency, flexibility and intelligence, has brought new changes and improvements to campus management, and has broad application prospects and social value.

[0036] 3. For issues with complaint intentions, the system of the present invention automatically upgrades to work order services, deeply addresses user demands, and uses the intelligent decision-making ability of the large language model to optimize the work order transfer process, improve problem response speed and problem solving efficiency, reduce waiting time for teachers and students, and enhance campus service satisfaction.

[0037] 4. Traditional customer service mainly provides services through preset questions and rules, with a limited scope of questions and answers, lacks generalization and association capabilities for user questions, and may not be able to answer questions that change slightly. The system and method of the embodiment of the present invention use a large model to train and fine-tune for campus scenarios, can accurately understand user questions, accurately retrieve answers, and organize output in a way that conforms to human language habits through prompt templates.

[0038] 5. Easy to maintain. In actual scenarios, new question-and-answer requirements will continue to arise. Customers can upload information on their own, and there is no need to organize and debug information as time-consuming and labor-intensive as with traditional databases.

[0039] 6. Risk avoidance. By utilizing the capabilities of the large model and security and privacy modules, it is easy to identify users’ uncivilized language and illegal and unlawful words, refuse to provide services to them, and reduce security risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 It is a structural diagram of a campus intelligent customer service system based on a large model in one embodiment of the present invention.

[0042] Figure 2 It is a structural diagram of a campus intelligent customer service system based on a large model in another embodiment of the present invention.

[0043] Figure 3 It is a structural diagram of a campus intelligent customer service system based on a large model in another embodiment of the present invention.

[0044] Figure 4 It is a structural diagram of a campus intelligent customer service system based on a large model in another embodiment of the present invention.

[0045] Figure 5 It is a flowchart of a processing method of a campus intelligent customer service system based on a large model in one embodiment of the present invention.

[0046] Figure 6 It is a flowchart of a processing method of a campus intelligent customer service system based on a large model in another embodiment of the present invention. DETAILED DESCRIPTION

[0047] First of all, it should be noted that the terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "the" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates otherwise.

[0048] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0049] Figure 1 FIG. 1 is a schematic diagram of a campus intelligent customer service system based on a large model in one embodiment of the present invention. Figure 1 The system includes: a user input processing module 101, a large model docking processing module 102, a large language model 103, a work order processing module 104, and a campus knowledge base 105; wherein,

[0050] The user input processing module 101 is used to receive questions input by campus users, pre-process the questions, and obtain the pre-processed questions in text form;

[0051] The large model docking processing module 102 is used to input the pre-processed text-based questions into the large language model 103; and provide the reply answers sent by the large language model 103 to the campus user;

[0052] The large language model 103 is used to perform semantic recognition on the received questions, mine word-sentence relationships, extract keywords, and search for answers in the campus knowledge base 105. If an answer can be retrieved, a reply answer is generated using the answer and a preset prompt template, and the reply answer is sent to the large model docking processing module 102. If an answer cannot be retrieved, the work order processing module 104 is triggered to process according to the recognized semantic information;

[0053] The work order processing module 104 is used to create a work order to be processed after being triggered by the large language model 103, and send the work order to be processed to the corresponding campus administrative department.

[0054] Figure 1 The system of the present invention shown has at least the following beneficial effects:

[0055] 1. In campus life, students can find a unified intelligent entrance that can solve various campus life problems. Students only need to input their own questions about campus life into the user input processing module of the system of the present invention, and they can get the relevant answers output by the system of the present invention. Students do not need to find relevant personnel of the school to solve the problems. The campus intelligent customer service solution can be realized in an intelligent way, saving time, effort and efficiency.

[0056] 2. Based on the system of the present invention, a campus intelligent customer service system solution is implemented, combining the large language model with the campus knowledge base. The large language model can grasp the inherent laws and contextual relationships of the language through deep learning and natural language processing technology, so that it can make more accurate and reasonable responses when faced with complex and changeable language input. The campus knowledge base stores a wealth of campus management knowledge and experience, providing data support for the large language model. Combining the large language model with the campus knowledge base can enable the system to understand user needs more accurately and provide more accurate management services. It effectively improves the answer effect of the model in complex scenarios, ensuring the accuracy of information transmission and user satisfaction. The system of the present invention, with its high efficiency, flexibility and intelligence, has brought new changes and improvements to campus management, and has broad application prospects and social value.

[0057] 3. For issues with complaint intentions, the system of the present invention automatically upgrades to work order services, deeply addresses user demands, and uses the intelligent decision-making ability of the large language model to optimize the work order transfer process, improve problem response speed and problem solving efficiency, reduce waiting time for teachers and students, and enhance campus service satisfaction.

[0058] See also Figure 1 In one embodiment of the system of the present invention, the user input processing module 101 includes:

[0059] The text input information processing submodule 1011 is used to receive text format questions input by campus users, and perform at least one of corpus cleaning, word segmentation, semantic error correction and part-of-speech tagging on the text format questions, so as to obtain pre-processed text format questions. Based on the text input information processing submodule 1011, campus users can directly input text format questions.

[0060] See also Figure 1 In one embodiment of the system of the present invention, the user input processing module 101 includes:

[0061] The voice input information processing submodule 1012 is used to receive the voice format questions input by the campus user, convert the voice format questions into text format questions, and then perform at least one of corpus cleaning, word segmentation, semantic error correction and part-of-speech tagging on the text format questions, so as to obtain the pre-processed text format questions. Based on the voice input information processing submodule 1012, the campus user can directly input questions in voice.

[0062] See also Figure 2 , the system further comprises: a general knowledge base 201;

[0063] The large language model 103 is further used to search in the general knowledge base 201 when the answer cannot be retrieved in the campus knowledge base 105 and before the identified semantic information is sent to the work order processing module 104. If the answer can be retrieved in the general knowledge base 201, a reply answer is generated using the answer and a pre-set Prompt template, and the reply answer is sent to the large model docking processing module 102. If the answer cannot be retrieved in the general knowledge base 201, the process of sending the identified semantic information to the work order processing module 104 continues.

[0064] By using the campus knowledge base 105 specifically for campus life and the general knowledge base 201 for general questions, the search scope can be expanded to ensure that the questions are answered to the greatest extent possible.

[0065] As mentioned above, the system of the present invention automatically upgrades the problem with complaint intention to work order service. At this time, in one embodiment of the present invention, when the large language model 103 triggers the work order processing module 104 to process according to the identified semantic information, it is specifically configured to execute: judging whether the user has the intention to complain and / or expresses dissatisfaction or complaint about the campus according to the identified semantic information, if so, using natural language processing capabilities to extract the information related to the intention input by the user, and sending the extracted information related to the intention to the work order processing module 104.

[0066] As mentioned above, the system of the present invention automatically upgrades the problem with complaint intention to a work order service. At this time, in one embodiment of the present invention, see Figure 3 , the system further comprises: at least one administrative department processing module 301 in the campus, different administrative department processing modules 301 belong to different campus administrative departments;

[0067] Each administrative department processing module 301 is used to perform the processing required by the campus user after receiving the work order to be processed sent by the work order processing module 104, and send the work order processing result to the work order processing module 104;

[0068] The work order processing module 104 is further configured to provide the work order processing result to the campus user after receiving the work order processing result.

[0069] In the embodiment of the present invention, the system further comprises: a security and privacy module 401;

[0070] The large model docking processing module 102 sends the reply answer sent by the large language model 103 to the security and privacy module 401 before providing the reply answer to the campus user; when the received evaluation result is that the answer does not include privacy information, the reply answer is provided to the campus user, otherwise, the answer is refused to be provided to the campus user.

[0071] The security and privacy module 401 is used to check the received reply answer, evaluate whether the reply answer includes privacy information, and notify the large model docking processing module 102 of the evaluation result.

[0072] Figure 5 FIG. 1 is a flowchart of a processing method of a campus intelligent customer service system based on a large model in one embodiment of the present invention. Figure 5 , based on the system in the embodiment of the present invention, the method includes:

[0073] Step 501: receiving a question input by a campus user, preprocessing the question, and obtaining the preprocessed question in text form.

[0074] In various embodiments of the present invention, preprocessing of questions may include: at least one of corpus cleaning, word segmentation, semantic error correction, and part-of-speech tagging, and finally outputting the question in text form without grammatical errors and ambiguity.

[0075] In various embodiments of the present invention, campus users can input questions in text form or directly use voice input. If the questions are in voice form, the voice input information processing submodule 1012 can be used to convert the voice questions into text questions, and then perform preprocessing.

[0076] Step 503: Input the pre-processed text-based question into the large language model.

[0077] Step 505: The large language model performs semantic recognition on the received text-based question, mines word-sentence relationships, and extracts key words.

[0078] Step 507: Determine whether the answer can be retrieved in the campus knowledge base based on the keyword. If so, execute step 509; otherwise, execute step 511.

[0079] Step 509: Generate a reply answer using the answer and a preset prompt template, provide the reply answer to the campus user, and end the current process.

[0080] In various embodiments of the present invention, a prompt can be written in advance as a campus customer service, and a large language model can be injected so that the answer returned by the large language model looks more in line with the customer service role, appears natural and friendly, and enhances the user experience. Therefore, the answer retrieved from the campus knowledge base will not be directly returned as an answer to the question, but will be used as a reference material for the large language model to generate a reply answer, that is, to construct a prompt. In the interaction of the large model, prompt is a key interaction method used to guide the model to generate a specific type of response. For example, in an embodiment of the present invention, the prompt template used by the system is: please learn the following knowledge {………}, now, please deeply understand and answer the following questions, ensure that the answers are accurate and fluent, do not quote any background information, and the question is {…………}. Among them, the content in the first {…………} can be a text answer retrieved from the campus knowledge base, and the content in the second {………} can be a question entered by the user. Through this kind of prompt template, the large voice model can be analyzed and summarized more reasonably.

[0081] Step 511: Create a work order to be processed according to the identified semantic information, and send the work order to be processed to the corresponding campus administrative department.

[0082] In one embodiment of the method of the present invention, in step 511, creating a work order to be processed according to the identified semantic information includes:

[0083] Step 511A: Determine whether the user has a complaint intention and / or expresses dissatisfaction or complaints about the campus based on the identified semantic information. If so, the large language model uses natural language processing capabilities to extract information related to the intention input by the user, and sends the extracted information related to the intention to the work order processing module, which creates a work order to be processed.

[0084] In one embodiment of the method of the present invention, different administrative department processing modules belong to different campus administrative departments;

[0085] The method further includes step 511B: after receiving the work order to be processed sent by the work order processing module, each administrative department processing module performs the processing required by the campus user and sends the work order processing result to the work order processing module; the work order processing module further provides the work order processing result to the campus user after receiving the work order processing result.

[0086] The following example illustrates the processing related to step 511.

[0087] In various embodiments of the present invention, the large language model has the ability to identify user intentions. Specifically, the large language model can determine whether the current user has the intention to complain, whether he or she expresses dissatisfaction or complaints about the campus. If the user has the above intentions, the current conversation may no longer conform to the question-and-answer scenario, but more like a scenario of expressing demands and expecting the problem to be solved (similar to dialing the phone number 12345).

[0088] For example, a campus user, such as a student, may ask the following questions:

[0089] Question 1: How to log in to Xuexitong

[0090] Question 2: What is the phone number of the campus network information center?

[0091] Question 3: The hygiene in the cafeteria is very poor and the dishes are often not washed cleanly. Who should I contact for feedback?

[0092] Question 4: The network in the dormitory has been very unstable recently. What’s going on?

[0093] Among the above four questions, questions one and two are more in line with typical question-and-answer scenarios. After the large language model retrieves the answer in the campus knowledge base / general knowledge base, it generates a reply answer and provides it to the user to complete the task. Questions three and four clearly express the demands. If the system simply replies to such questions, it obviously does not fundamentally solve the problem, and users cannot get a good experience and satisfactory answers in such a customer service system. In particular, users sometimes do not even realize that they have a specific demand to make to the school. If the system can make intelligent judgments and give guidance, this will be a very good user experience. Therefore, through the processing of steps 511A and 511B above, the complaints of campus users can be resolved.

[0094] Figure 6 FIG. 1 is a flowchart of a processing method of a campus intelligent customer service system based on a large model in one embodiment of the present invention. Figure 6 The method is described below based on various systems in the embodiments of the present invention, including:

[0095] Step 601: Campus users access the campus intelligent customer service system based on the big model on a PC or mobile terminal.

[0096] Step 603: The campus user inputs a question in text form; or the campus user inputs a question in voice form, and the system converts the voice question into a text question.

[0097] Step 605: Preprocess the text-based questions, such as corpus cleaning, word segmentation, semantic error correction, part-of-speech tagging, etc., and finally output text information without grammatical errors and ambiguity.

[0098] Step 607: The pre-processed text question is packaged according to the agreed format and then submitted to the large language model.

[0099] Step 609: The large language model performs semantic recognition on the text-based question, mines word-sentence relationships, extracts keywords, and retrieves answers from the campus knowledge base.

[0100] Step 611: Determine whether the answer can be retrieved in the campus knowledge base. If so, execute step 615; otherwise, execute step 613.

[0101] Step 613: Search the answer in the general knowledge base to determine whether the answer can be found in the general knowledge base. If so, execute step 615; otherwise, execute step 619.

[0102] Step 615: Generate a reply answer using the retrieved answer and the preset Prompt template.

[0103] Step 617: Perform a security check on the reply answer to evaluate whether the reply answer is safe and compliant, such as whether it includes privacy information, etc. If it is safe, execute step 621; otherwise, execute step 619.

[0104] Step 619: Refuse to provide answers to the campus user and end the current process.

[0105] Step 621: Provide the reply answer to the campus user.

[0106] Step 623: Determine whether the user has a complaint intention and / or expresses dissatisfaction or complaints about the campus based on the identified semantic information. If so, execute step 625; otherwise, end the current process.

[0107] Step 625: The large language model uses natural language processing capabilities to extract information related to the intention input by the user, and sends the extracted information related to the intention to the work order processing module.

[0108] Step 627: The work order processing module creates a pending work order and sends the pending work order to the administrative department processing module of the corresponding campus administrative department.

[0109] Step 629: After receiving the work order to be processed, the processing module of the campus administration department performs the processing required by the campus user and provides the work order processing result to the campus user.

[0110] It should be noted that the above-mentioned devices are usually implemented on the server side, and can be set on independent servers respectively, or some or all of the devices can be combined and set on the same server. The server can be a single server or a server cluster composed of multiple servers. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system. The above-mentioned devices can also be implemented on a computer terminal with strong computing power.

[0111] An embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed in a computer, the computer is caused to execute a method in any one of the embodiments in the specification.

[0112] An embodiment of the present invention provides a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method in any embodiment of the specification is implemented.

[0113] It is to be understood that the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on the device of the embodiment of the present invention. In other embodiments of the specification, the above-mentioned device may include more or fewer components than those shown in the figure, or combine some components, or split some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0114] The various embodiments of the present invention are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0115] Those skilled in the art should be aware that in one or more of the above examples, the functions described in the present invention can be implemented by hardware, software, widgets, or any combination thereof. When implemented by software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium.

[0116] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. The campus intelligent customer service system based on the big model is characterized by: The system includes: a user input processing module, a large model docking processing module, a large language model, a work order processing module, and a campus knowledge base; wherein, The user input processing module is used to receive questions input by campus users, pre-process the questions, and obtain the pre-processed questions in text form; The large model docking processing module is used to input the pre-processed text-based questions into the large language model; and provide the reply answers sent by the large language model to the campus user; The large language model is used to perform semantic recognition on the received questions, mine word-sentence relationships, extract keywords, and search for answers in the campus knowledge base. If an answer can be retrieved, a reply answer is generated using the answer and a pre-set prompt template, and the reply answer is sent to the large model docking processing module. If an answer cannot be retrieved, the work order processing module is triggered to process according to the recognized semantic information; The work order processing module is used to create a work order to be processed after being triggered by the large language model, and send the work order to be processed to the corresponding campus administrative department.

2. The system according to claim 1, characterized in that The user input processing module comprises: The text input information processing submodule is used to receive text format questions input by campus users, and perform at least one of corpus cleaning, word segmentation, semantic error correction and part-of-speech tagging on the text format questions, so as to obtain pre-processed text format questions; and / or, The voice input information processing submodule is used to receive voice-formatted questions input by campus users, convert the voice-formatted questions into text-formatted questions, and then perform at least one of corpus cleaning, word segmentation, semantic error correction, and part-of-speech tagging on the text-formatted questions, so as to obtain pre-processed text-formatted questions.

3. The system according to claim 1, characterized in that The system further includes: a general knowledge base; The large language model is further used to search in the general knowledge base when the answer cannot be retrieved in the campus knowledge base and before the identified semantic information is sent to the work order processing module. If the answer can be retrieved in the general knowledge base, a reply answer is generated using the answer and a pre-set Prompt template, and the reply answer is sent to the large model docking processing module. If the answer cannot be retrieved in the general knowledge base, the process of sending the identified semantic information to the work order processing module continues.

4. The system according to claim 1, characterized in that When the large language model executes the triggering of the work order processing module for processing based on the identified semantic information, it is specifically configured to execute: judging whether the user has a complaint intention and / or expresses dissatisfaction or complaint about the campus based on the identified semantic information; if so, using natural language processing capabilities to extract information related to the intention input by the user, and sending the extracted information related to the intention to the work order processing module.

5. The system according to claim 1 or 4, characterized in that: The system further includes: at least one administrative department processing module within the campus, different administrative department processing modules belonging to different campus administrative departments; Each administrative department processing module is used to perform the processing required by the campus user after receiving the work order to be processed sent by the work order processing module, and send the work order processing result to the work order processing module; The work order processing module is further used to provide the work order processing result to the campus user after receiving the work order processing result.

6. The system according to claim 1, characterized in that The system further includes: a security and privacy module; The large model docking processing module sends the reply answer sent by the large language model to the security and privacy module before providing the reply answer to the campus user; when the received evaluation result is that the answer does not include privacy information, the reply answer is provided to the campus user, otherwise, the answer is refused to be provided to the campus user. The security and privacy module is used to check the received reply answer, evaluate whether the reply answer includes privacy information, and notify the large model docking processing module of the evaluation result.

7. The processing method of the campus intelligent customer service system based on a large model according to any one of claims 1 to 6, characterized in that: The method includes: Receive questions input by campus users, pre-process the questions, and obtain the questions in a pre-processed text form; Input the preprocessed text-based questions into the large language model; The large language model uses the received text-based questions to perform semantic recognition, mine word-sentence relationships, extract subject terms, and retrieve answers from the campus knowledge base; If the answer can be retrieved, a reply answer is generated using the answer and a preset prompt template, and then the reply answer is provided to the campus user; If the answer cannot be retrieved, a pending work order is created based on the identified semantic information and sent to the corresponding campus administrative department.

8. The method according to claim 7, characterized in that The step of creating a work order to be processed according to the identified semantic information includes: Based on the identified semantic information, it is determined whether the user has the intention to complain and / or expresses dissatisfaction or complaints about the campus. If so, the large language model uses natural language processing capabilities to extract information related to the intention input by the user, and sends the extracted information related to the intention to the work order processing module, which creates a work order to be processed.

9. The method according to claim 7 or 8, characterized in that: Different administrative department processing modules belong to different campus administrative departments; The method further includes: after each administrative department processing module receives the work order to be processed sent by the work order processing module, it performs the processing required by the campus user and sends the work order processing result to the work order processing module; after receiving the work order processing result, the work order processing module further provides the work order processing result to the campus user.

10. A computing device comprising a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method according to any one of claims 7 to 9 is implemented.

Images