Intelligent contradictory dispute resolving method and device

Through data collection, cleaning and model fine-tuning, a knowledge base is built and an intelligent mediation plan is generated, which solves the time-consuming and labor-intensive problems of traditional conflict and dispute resolution methods, and realizes the rapid identification and accurate analysis of conflict and disputes, improves resolution efficiency, and assists Mediators learn.

CN119006229APending Publication Date: 2024-11-22INSPUR SOFTWARE TECH CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411009877.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-26
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

Traditional conflict and dispute resolution methods are time-consuming and labor-intensive. Relying on manual experience, it is difficult to quickly identify, accurately analyze and effectively resolve conflicts and disputes.

Method used

Through data collection, cleaning, model fine-tuning and knowledge base construction, intelligent mediation solutions are generated, ChatGLM3-6B model is fine-tuned using Lora method, knowledge base is built in combination with LangChain-Chatchat system, and answers are generated using RAG to achieve intelligent resolution of conflicts and disputes.

Benefits of technology

The efficiency of conflict resolution has been improved, and the rapid identification and accurate analysis of conflicts and disputes has been achieved. The generated intelligent resolution plan can assist in learning and enhance the legal knowledge and practical training of mediators.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006229A_ABST
    Figure CN119006229A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of artificial intelligence, and particularly provides an intelligent contradictory dispute resolving method and device, and the method comprises the following steps: S1, data collection; s2, data cleaning and conversion; s3, fine adjustment of the model; s4, constructing a knowledge base; and S5, generating an intelligent mediation scheme. Compared with the prior art, an intelligent solution scheme can be generated, learning can be assisted, and learning and training of mediation personnel on legal knowledge theories and practical knowledge can be enhanced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and specifically provides a method and device for intelligently resolving conflicts and disputes. Background Art

[0002] With the rapid development of social economy and the improvement of people's living standards, people's demand for fairness, justice and protection of rights and interests is growing.

[0003] The intricate problems in society are becoming increasingly prominent, and conflicts and disputes are gradually showing new characteristics of diversification, complexity and groupization. How to resolve various conflicts and disputes efficiently and quickly is the focus and difficulty of current governance.

[0004] Although traditional conflict and dispute resolution methods can solve problems to a certain extent, they are often time-consuming and labor-intensive. The analysis, summary and prediction of conflict and dispute mediation almost rely on the accumulated work experience of people, and the mediation methods and mediation plans also rely solely on the accumulated work experience of mediators. How to achieve rapid identification, accurate analysis and effective resolution of disputes through the integration of information and intelligent technologies is an urgent problem to be solved by technical personnel in this field. Summary of the invention

[0005] The present invention aims at the deficiencies of the above-mentioned prior art and provides a highly practical method for intelligently resolving conflicts and disputes.

[0006] A further technical task of the present invention is to provide an intelligent conflict resolution device that is rationally designed, safe and applicable.

[0007] The technical solution adopted by the present invention to solve its technical problem is:

[0008] An intelligent method for resolving conflicts and disputes comprises the following steps:

[0009] S1, data collection;

[0010] S2, data cleaning and conversion;

[0011] S3, model fine-tuning;

[0012] S4, knowledge base construction;

[0013] S5. Generation of intelligent mediation solution.

[0014] Furthermore, in step S1, conflict and dispute data from various sources are collected, conflict and dispute related data are extracted, or conflict and dispute event data are entered.

[0015] Furthermore, in step S2, the data gathered in step S1 include conflict and dispute event data, handling and feedback information, including structural information including major categories of conflicts and disputes, detailed categories of conflicts and disputes, descriptions of conflict and dispute events, mediation records-mediation subject types, mediation records-mediation process records, mediation records-mediation results, mediation results and whether there is escalation or recurrence.

[0016] Furthermore, in step S3, ChatGLM3-6B is fine-tuned using the Lora method. ChatGLM3-6B recommends fine-tuning using a multi-round dialogue format, using descriptions of conflict-related events as input and mediation records as output.

[0017] Furthermore, in step S4, the model fine-tuning steps are as follows:

[0018] (1) Prepare training sets and validation sets;

[0019] (2) Use the efficient fine-tuning framework LLaMA-Factory for fine-tuning.

[0020] Furthermore, in step S4, LangChain-Chatchat is used to construct a knowledge base, including the following steps:

[0021] S4-1. Create a knowledge base. Call the interface / knowledge_base / create_knowledge_base to create a knowledge base.

[0022] S4-2, establishing a knowledge base file, relying on the data cleaned in step S2, establishing a knowledge base file;

[0023] S4-3, upload files, call / knowledge_base / upload_docs interface to upload the files in step S4-1 to the knowledge base, and vectorize them;

[0024] S4-4. Upload to the LangChain-Chatchat system for file processing.

[0025] Furthermore, in step S4-4, the file processing process uploaded to the LangChain-Chatchat system is as follows:

[0026] (1) Load the file;

[0027] (2) Read the text;

[0028] (3) Text segmentation;

[0029] (4) Text vectorization;

[0030] (5) Question vectorization;

[0031] (6) Match the text vectors with the ones that are most similar to the question vector;

[0032] (7) The matched text is added to the prompt as context along with the question;

[0033] (8) Submit to ChatGLM to generate the answer.

[0034] Furthermore, in step S5, it further includes:

[0035] S5-1, based on the model fine-tuning in step S3, an intelligent mediation process is generated for new user input, and the model intention is aligned with humans;

[0036] S5-2. Based on the knowledge base in step S4, generate answers using the RAG method for user input.

[0037] An intelligent conflict resolution device, comprising: at least one memory and at least one processor;

[0038] The at least one memory is used to store a machine-readable program;

[0039] The at least one processor is used to call the machine-readable program to execute an intelligent method for resolving conflicts and disputes.

[0040] Compared with the prior art, the intelligent conflict resolution method and device of the present invention has the following outstanding beneficial effects:

[0041] The present invention improves the efficiency of resolving conflicts and disputes at the grassroots level through data aggregation, data classification and cleaning, model fine-tuning, knowledge base construction, and intelligent solution generation. By integrating information technology and intelligent technology, it realizes the rapid identification, accurate analysis, and effective resolution of conflicts and disputes. At the same time, the generated intelligent solution can also be used to assist learning and strengthen the learning and training of mediators on legal knowledge theory and practical knowledge. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] 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.

[0043] Attached Figure 1 It is a flow chart of an intelligent method for resolving conflicts and disputes. DETAILED DESCRIPTION

[0044] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention is further described in detail below in conjunction with specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] A best embodiment is given below:

[0046] like Figure 1 As shown, a method for intelligently resolving conflicts and disputes in this embodiment has the following steps:

[0047] S1, data collection;

[0048] Gather conflict and dispute data from various sources, extract conflict and dispute related data, or enter conflict and dispute event data.

[0049] S2, data cleaning and conversion;

[0050] The data collected in step S1 includes conflict and dispute event data, disposal and feedback information, and mainly includes the following structural information:

[0051]

[0052] S3, model fine-tuning;

[0053] Use Lora to fine-tune ChatGLM3-6B. ChatGLM3-6B recommends fine-tuning using a multi-round dialogue format. Use the description of conflict-related events as input and the mediation record as output. The model fine-tuning steps are as follows:

[0054] S3-1. Dataset preparation;

[0055] Prepare training sets and validation sets. The data format is as follows:

[0056] "conversations":[

[0057] "role":"system",

[0058] "content":"You are an experienced XXXX mediator, and you are very good at mediating disputes between people."

[0059] "role":"user",

[0060] "content":"The current conflict and dispute information is as follows: XXXXX, how to mediate?",

[0062] "role":"assistant",

[0063] "content":"Mediation Record"

[0064] S3-2 is fine-tuned using the efficient fine-tuning framework LLaMA-Factory.

[0065] Execute the following commands to fine-tune, infer, and merge the model:

[0066] CUDA_VISIBLE_DEVICES=0llamafactory-cli train mdjf / lora_single_gpu / lora_sft.yaml

[0067] CUDA_VISIBLE_DEVICES=0llamafactory-cli chat mdjf / inference / lora_sft.yaml

[0068] CUDA_VISIBLE_DEVICES=0llamafactory-cli export mdjf / merge_lora / lora_sft.yaml

[0069] S4, knowledge base construction;

[0070] Since fine-tuning the model is relatively expensive and time-consuming, the fine-tuning dataset is only fine-tuned on the aggregated historical conflict and dispute data.

[0071] The fine-tuned model is used to simulate the tone of the mediator's answer and basic mediation thinking. For the newly collected conflict and dispute data, the latest question and answer data is plugged into ChatGLM3-6B using an external knowledge base to continuously update the content information of the mediation plan.

[0072] The process of using LangChain-Chatchat to build a knowledge base includes the following steps:

[0073] S4-1. Create a knowledge base. Call the interface / knowledge_base / create_knowledge_base to create a knowledge base.

[0074] S4-2, establishing a knowledge base file, relying on the data cleaned in step S2, establishing a knowledge base file;

[0075] S4-3, upload files, call / knowledge_base / upload_docs interface to upload the files in step S4-1 to the knowledge base, and vectorize them;

[0076] S4-4. Upload to the LangChain-Chatchat system for file processing. The process is as follows:

[0077] (1) Load the file;

[0078] (2) Read the text;

[0079] (3) Text segmentation;

[0080] (4) Text vectorization;

[0081] (5) Question vectorization;

[0082] (6) Match the text vectors with the ones that are most similar to the question vector;

[0083] (7) The matched text is added to the prompt as context along with the question;

[0084] (8) Submit to ChatGLM to generate the answer.

[0085] S5, generation of intelligent mediation scheme;

[0086] In response to the user input, steps S3 and S4 are called respectively to generate answers and ranges, as follows:

[0087] S5-1. Based on the fine-tuning of the model in step S3, an intelligent mediation process is generated for new user input. Since the model is fine-tuned in the field of conflicts and disputes, the model intention is aligned with humans.

[0088] S5-2. Based on the knowledge base in step S4, generate answers using the RAG method for user input. Since the answer results rely on the knowledge base, the answer results have more reference basis.

[0089] The generated intelligent mediation method can not only improve the efficiency of mediators in resolving disputes, but also be used to assist learning and strengthen the learning and training of mediators on legal knowledge theory and practical knowledge.

[0090] Based on the above method, a device for intelligently resolving conflicts and disputes in this embodiment includes: at least one memory and at least one processor;

[0091] The at least one memory is used to store a machine-readable program;

[0092] The at least one processor is used to call the machine-readable program to execute an intelligent method for resolving conflicts and disputes.

[0093] The above-mentioned specific implementations are only specific cases of the present invention. The patent protection scope of the present invention includes but is not limited to the above-mentioned specific implementations. Any technical solutions that comply with the claims of the present invention and any appropriate changes or substitutions made by ordinary technicians in the relevant technical field shall fall within the patent protection scope of the present invention.

[0094] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent method for resolving conflicts and disputes, characterized in that: The steps are as follows: S1, data collection; S2, data cleaning and conversion; S3, model fine-tuning; S4, knowledge base construction; S5. Generation of intelligent mediation solution.

2. The method for intelligently resolving conflicts and disputes according to claim 1, characterized in that: In step S1, conflict and dispute data from various sources are collected, conflict and dispute related data are extracted, or conflict and dispute event data are entered.

3. The method for intelligently resolving conflicts and disputes according to claim 2, characterized in that: In step S2, the data gathered in step S1 include conflict and dispute event data, handling and feedback information, including structural information including major categories of conflicts and disputes, detailed categories of conflicts and disputes, descriptions of conflict and dispute events, mediation records-mediation subject types, mediation records-mediation process records, mediation records-mediation results, mediation results and whether there is escalation or recurrence.

4. The method for intelligently resolving conflicts and disputes according to claim 3 is characterized in that: In step S3, ChatGLM3-6B is fine-tuned using the Lora method. ChatGLM3-6B recommends fine-tuning using a multi-round dialogue format, using descriptions of conflict-related events as input and mediation records as output.

5. The method for intelligently resolving conflicts and disputes according to claim 4 is characterized in that: In step S4, the model fine-tuning steps are as follows: (1) Prepare training sets and validation sets; (2) Use the efficient fine-tuning framework LLaMA-Factory for fine-tuning.

6. The method for intelligently resolving conflicts and disputes according to claim 5, characterized in that: In step S4, LangChain-Chatchat is used to construct a knowledge base, including the following steps: S4-1. Create a knowledge base. Call the interface / knowledge_base / create_knowledge_base to create a knowledge base. S4-2, establishing a knowledge base file, relying on the data cleaned in step S2, establishing a knowledge base file; S4-3, upload files, call / knowledge_base / upload_docs interface to upload the files in step S4-1 to the knowledge base, and vectorize them; S4-4. Upload to the LangChain-Chatchat system for file processing.

7. The method for intelligently resolving conflicts and disputes according to claim 6, characterized in that: In step S4-4, the file processing process uploaded to the LangChain-Chatchat system is as follows: (1) Load the file; (2) Read the text; (3) Text segmentation; (4) Text vectorization; (5) Question vectorization; (6) Match the text vectors with the ones that are most similar to the question vector; (7) The matched text is added to the prompt as context along with the question; (8) Submit to ChatGLM to generate the answer.

8. The method for intelligently resolving conflicts and disputes according to claim 7, characterized in that: In step S5, further comprising: S5-1, based on the model fine-tuning in step S3, an intelligent mediation process is generated for new user input, and the model intention is aligned with humans; S5-2. Based on the knowledge base in step S4, generate answers using the RAG method for user input.

9. An intelligent device for resolving conflicts and disputes, characterized in that: include: at least one memory and at least one processor; The at least one memory is used to store a machine-readable program; The at least one processor is configured to call the machine-readable program to execute the method according to any one of claims 1 to 8.

Citation Information

Cited By

  • Basic-level governance duty intelligent matching method, system and device and storage medium

    CN120410132A