Arbitration trial thought chain inference method and device based on large language model

Abstract

This invention discloses a method and apparatus for inferring the arbitration hearing thought chain based on a large language model, belonging to the field of intelligent labor arbitration hearing technology. The information extraction module uses natural language processing technology to parse the application documents and extract keyword data sets. The information analysis module identifies and extracts the applicant's basic information during their employment based on the keyword data sets from the text parsing. The information analysis module also sorts out the disputed matters of the case and retrieves relevant laws, regulations, and judicial interpretations. A mapping relationship is established between the arbitration request and the disputed matters, analyzing the rationality and feasibility of the arbitration request and matching it with the disputed matters to automatically generate an arbitration hearing thought chain. The arbitration hearing thought chain includes the applicant's basic information, the disputed matters, the arbitration request, and the corresponding facts, evidence, and legal basis. This simulates the arbitrator's thinking process, automating the arbitration hearing process. It helps improve arbitration efficiency and ensures the fairness of the hearing results.

Description

Technical Field

[0001] This invention relates to the field of intelligent labor arbitration adjudication technology, and in particular to an arbitration adjudication thought chain inference method and device based on a large language model. Background Technology

[0002] With the rapid increase in the number of labor dispute cases and the increasing complexity of case types, establishing a new diversified dispute resolution model using data and intelligent systems has become an inevitable trend. This will help improve processing efficiency, reduce resource waste, and achieve more precise governance.

[0003] The construction of intelligent arbitration aims to demonstrate its advantages of convenience, efficiency, and affordability. Supported by network technology and big data, intelligent arbitration systems can better serve arbitrators in handling cases and relevant departments in decision-making, promoting the construction of a diversified dispute resolution mechanism centered on labor and personnel dispute arbitration.

[0004] While intelligent arbitration is feasible, there are technical obstacles to its implementation. For example, labor and personnel disputes often involve highly unstructured evidence, which poses a challenge to the level of intelligent implementation.

[0005] Inter-system collaboration issues: Currently, there is a lack of collaborative case-handling systems between arbitration and litigation, resulting in a digital disconnect between arbitration and judgment. In this situation, information from the arbitration stage needs to be retrieved offline during the trial stage, affecting overall efficiency and accuracy.

[0006] The existing labor dispute resolution mechanisms still rely on a traditional single-threaded model, which is insufficient to cope with the high incidence of labor disputes. Courts, arbitration, and judicial mediation operate under different systems, leading to poor coordination among the three parties and increasing the time and financial burden on the parties involved.

[0007] In summary, although intelligent labor arbitration has made some progress in improving efficiency and convenience, it still faces many challenges and shortcomings in terms of technological implementation, system collaboration, and model innovation. It cannot achieve intelligent arbitration based on the arbitration hearing thought chain. Summary of the Invention

[0008] To address the shortcomings of intelligent labor arbitration hearings mentioned above, this invention provides a method and apparatus for inferring the thought chain in arbitration hearings based on a large language model. The technical solution adopted is as follows:

[0009] The method for inferring the thought chain in arbitration proceedings based on a large language model includes the following steps:

[0010] Step 1: Enter the electronic application documents for labor arbitration into the information entry module;

[0011] Step 2: The information extraction module uses natural language processing technology to parse the application documents and extract keyword data groups;

[0012] Step 3: The information analysis module identifies and extracts the applicant's basic information during employment based on the keyword data group of text parsing. The applicant's basic information during employment includes: date of employment, date of resignation, salary structure, position and job responsibilities.

[0013] Step 4: The information analysis module sorts out the disputed matters of the case, uses a large language model to analyze the application documents to extract disputed keywords, clusters the keywords, classifies the labor dispute matters into categories, and forms a list of disputed matters;

[0014] Step 5: The information analysis module retrieves the relevant laws, regulations, and judicial interpretations for the case, uses knowledge graph technology to associate the disputed matters with legal provisions, compares the factual descriptions in the application documents, and matches appropriate evidentiary materials.

[0015] Step 6: The information analysis module uses NLP technology to perform semantic analysis on the applicant's arbitration request and extract key elements of the arbitration request, such as the matters requested, the amount, and the method of compensation.

[0016] By mapping the arbitration request to the disputed matters, establishing a mapping relationship between the arbitration request and the disputed matters, analyzing the rationality and feasibility of the arbitration request, matching it with the disputed matters, and automatically generating an arbitration hearing thought chain, the arbitration hearing thought chain includes the applicant's basic information, the disputed matters, the arbitration request and the corresponding facts, evidence and legal basis.

[0017] By adopting the above technical solutions, the information entry module can realize the input of electronic labor arbitration application documents. The information entry module can be in the form of an online information entry page, allowing applicants to enter their prepared electronic labor arbitration application documents and other relevant supporting materials. It can also simultaneously upload relevant complaints and supporting materials from enterprises. The information analysis module can perform text parsing of the application documents based on natural language processing technology. The information analysis module can be implemented using devices with autonomous computing capabilities, such as computers or intelligent chips, and can use natural language processing (NLP) technology to parse the application documents. Identifying and extracting the following information—employment date, termination date, salary structure, position, job responsibilities, etc.—to build a database of the applicant's basic information, extracting keyword data groups, and sorting out the disputed matters in the arbitration application documents are key steps. NLP technology can be used to analyze the application documents, extract disputed keywords, and cluster them according to the keywords to classify the disputed matters (such as labor remuneration, termination of labor contract, etc.).

[0018] The disputed matters are mapped to relevant facts, evidence, and legal provisions. Knowledge graph technology is used to associate the disputed matters with relevant legal provisions, enabling automatic retrieval of laws, regulations, and judicial interpretations related to the case. Appropriate evidentiary materials are matched by comparing the factual descriptions in the application documents.

[0019] Analyze the applicant's arbitration claim. Use NLP technology to perform semantic analysis on the applicant's arbitration claim. Extract key elements of the arbitration claim, such as the matters claimed, the amount, and the method of compensation.

[0020] Establish a mapping relationship between the arbitration request and the disputed matters. Analyze the reasonableness and feasibility of the arbitration request and match it with the disputed matters.

[0021] It automatically generates a thought process chain for arbitration proceedings, including the applicant's basic information, the disputed matters, the arbitration request and corresponding facts, and finally the corresponding evidence and legal basis.

[0022] Through the above intelligent design, the thought process of an arbitrator can be simulated, thus automating the arbitration process. This method helps improve arbitration efficiency and ensures the fairness of the outcome.

[0023] Optionally, after collecting the applicant's basic information during their employment in step 3, the data is also verified. The verification of the applicant's basic information during their employment includes the following steps:

[0024] The data verification module uses optical character recognition technology to scan and verify the supporting documents provided by the applicant and the employee files provided by the company, and verifies the accuracy of the identification of basic information during the employment period based on the recognition results.

[0025] By employing the above technical solutions, data verification can extract the accuracy of information. It involves comparing the data with employee records provided by the company. Optical Character Recognition (OCR) technology is used to scan and verify relevant supporting documents. The specific details of Optical Character Recognition (OCR) technology are as follows:

[0026] Adjust the image's brightness, contrast, and resolution to ensure the text is clearly legible. Use image processing techniques to remove noise and artifacts generated during scanning. If the scanned image is tilted, it needs to be corrected using software. Use OCR software to analyze the pre-processed image and recognize the text content. Convert the OCR-recognized text into an editable text format, such as a TXT or Word document. The OCR software should be able to recognize the text layout and distinguish between different parts such as titles, body text, and tables. Manually check whether the OCR output text matches the original scanned document to ensure the accuracy of the information. Correct any OCR-recognized errors, including spelling mistakes and formatting issues. Categorize the extracted text according to a certain structure, such as classifying pay slip information and overtime records separately. Integrate the OCR-processed data into the arbitration hearing system and link it with other case information.

[0027] Optional categories of labor disputes include disputes concerning the confirmation of labor relations, disputes concerning dismissal, resignation, and departure, disputes concerning working hours and rest / leave, disputes concerning social insurance and welfare, disputes concerning labor remuneration and medical expenses for work-related injuries, and disputes concerning economic compensation and damages.

[0028] Optionally, in step 2, the information extraction formula is:

[0029] ;

[0030] in This represents the information extraction function. This indicates the input application document. This represents the nth keyword extracted.

[0031] Optionally, the following information verification formula may be used to verify the applicant's basic information during their employment period:

[0032] ;

[0033] Where V represents the information verification function. This represents the keyword to be verified, and D represents the employee file provided by the company. This represents the verification result of the m-th keyword.

[0034] Optionally, the formula for identifying disputed items in step 4 is:

[0035] ;

[0036] R is the disputed item identification function, and T is the text content. This represents the k-th disputed item.

[0037] Optionally, in step 5, the following formula is used to link the evidence to the legal provisions:

[0038] ;

[0039] Where A represents the correlation function. Let F represent the k-th disputed item, F represent the set of facts, and L represent the set of legal provisions. This represents the p-th fact associated with the disputed term. This represents the p-th legal provision associated with the disputed item.

[0040] Optionally, in step 6, the formula for matching the arbitration request with the disputed matters is as follows:

[0041] ;

[0042] M represents the function that matches the arbitration request with the disputed matters. Indicate the elements of the arbitration request, Indicates the disputed item, This indicates the matching result between the arbitration request and the disputed matters.

[0043] Optionally, in step 4, the keyword extraction formula for the large language model is:

[0044] ;

[0045] Where K represents the keyword extraction function, and T represents the preprocessed text. Indicates model parameters, This represents the nth keyword;

[0046] The keyword clustering formula is:

[0047] ;

[0048] C represents the clustering function, and K represents the set of extracted keywords. Indicates the parameters of the clustering algorithm. This represents the m-th cluster category;

[0049] The formula for generating the list of disputed items is:

[0050] ;

[0051] L represents the function that generates the list of disputed items, and C represents the clustering result. This represents the generated list of the m-th disputed items.

[0052] By employing the above technical solution, labor arbitration application documents are preprocessed to obtain cleaned text. Keyword extraction models (such as BERT) are used to extract keywords from the text. Clustering algorithms (such as K-means) are applied to cluster the keywords to determine the categories of disputed matters. Based on the clustering results, a list of disputed matters is generated.

[0053] A large amount of labeled data is needed to train the keyword extraction and clustering model before model training.

[0054] The model's performance depends on the accuracy of keyword extraction and clustering, so the model parameters need to be continuously optimized.

[0055] Labor dispute cases are complex and varied, and models may need to be updated regularly to adapt to new types of disputes.

[0056] Based on the above architecture and formulas, a large language model can be built that can process labor arbitration application documents, effectively extract dispute keywords, and classify disputed matters into categories.

[0057] The arbitration hearing thought chain inference device based on the big language model includes an information input module, an information extraction module, an information analysis module, and a result display module. The information input module is used to input the electronic application documents for labor arbitration and scan the supporting materials. The information extraction module is connected to the information extraction module and runs the information extraction program designed in step 2 of the arbitration hearing thought chain inference method based on the big language model to analyze the electronic application documents for labor arbitration and obtain the extracted keyword data group.

[0058] The information analysis module is connected to the information input module and the information extraction module. It uses the analysis program designed in steps 2 to 6 of the arbitration hearing thought chain inference method based on a large language model to obtain the arbitration hearing thought chain analysis results. The result display module communicates and interacts with the information analysis module to display the arbitration hearing thought chain analysis results.

[0059] By adopting the above technical solution, the information input module can be a networked computer, the information extraction module and the information analysis module are computer-based extraction or analysis programs, and the result display module is a monitor.

[0060] In summary, the present invention has at least one of the following beneficial technical effects:

[0061] This invention provides a method and apparatus for inferring the thought chain of arbitration proceedings based on a large language model. The information input module allows for the input of electronic labor arbitration application documents, while the information analysis module performs text parsing of the application documents using Natural Language Processing (NLP) technology. It identifies and extracts information such as start date, end date, salary structure, position, and job responsibilities to build a database of the applicant's basic information. Key steps include extracting keyword data groups and organizing the disputed matters in the arbitration application documents. Using NLP technology to analyze the application documents and extract disputed keywords, the disputed matters can be categorized (e.g., labor remuneration, termination of labor contract, etc.) based on keyword clustering. The disputed matters are then correlated with relevant facts, evidence, and legal provisions. Knowledge graph technology is used to associate the disputed matters with relevant legal provisions, enabling automatic retrieval of relevant laws, regulations, and judicial interpretations. The factual descriptions in the application documents are compared to match appropriate evidentiary materials. NLP technology is used to perform semantic analysis of the applicant's arbitration request, extracting key elements such as the requested matters, amount, and method of compensation.

[0062] Establish a mapping relationship between arbitration requests and disputed matters. Analyze the reasonableness and feasibility of the arbitration requests and match them with the disputed matters. Automatically generate an arbitration hearing thought chain, including the applicant's basic information, the disputed matters, the arbitration requests and corresponding facts, and finally the corresponding evidence and legal basis.

[0063] Through the above intelligent design, the thought process of an arbitrator can be simulated, thus automating the arbitration process. This method helps improve arbitration efficiency and ensures the fairness of the outcome. Attached Figure Description

[0064] Figure 1 This is a flowchart illustrating the arbitration hearing thought chain inference method based on a large language model, as described in this invention. Detailed Implementation

[0065] The present invention will be further described in detail below with reference to the accompanying drawings.

[0066] This invention discloses a method and apparatus for inferring the thought chain in arbitration proceedings based on a large language model.

[0067] Reference Figure 1 The arbitration hearing thought chain inference method based on a large language model includes the following steps:

[0068] Step 1: Enter the electronic application documents for labor arbitration into the information entry module;

[0069] Step 2: The information extraction module uses natural language processing technology to parse the application documents and extract keyword data groups;

[0070] Step 3: The information analysis module identifies and extracts the applicant's basic information during employment based on the keyword data group of text parsing. The applicant's basic information during employment includes: date of employment, date of resignation, salary structure, position and job responsibilities.

[0071] Step 4: The information analysis module sorts out the disputed matters of the case, uses a large language model to analyze the application documents to extract disputed keywords, clusters the keywords, classifies the labor dispute matters into categories, and forms a list of disputed matters;

[0072] Step 5: The information analysis module retrieves the relevant laws, regulations, and judicial interpretations for the case, uses knowledge graph technology to associate the disputed matters with legal provisions, compares the factual descriptions in the application documents, and matches appropriate evidentiary materials.

[0073] Step 6: The information analysis module uses NLP technology to perform semantic analysis on the applicant's arbitration request and extract key elements of the arbitration request, such as the matters requested, the amount, and the method of compensation.

[0074] By mapping the arbitration request to the disputed matters, establishing a mapping relationship between the arbitration request and the disputed matters, analyzing the rationality and feasibility of the arbitration request, matching it with the disputed matters, and automatically generating an arbitration hearing thought chain, the arbitration hearing thought chain includes the applicant's basic information, the disputed matters, the arbitration request and the corresponding facts, evidence and legal basis.

[0075] The information entry module allows for the input of electronic labor arbitration application documents. This can be done online through an entry page, allowing applicants to input their prepared electronic application documents and other relevant supporting materials. It can also simultaneously upload relevant complaints and supporting documents from the company. The information analysis module uses natural language processing (NLP) technology to parse the application documents. This module can be implemented using a computer, smart chip, or other device with autonomous computing capabilities. Key steps include identifying and extracting information such as start date, end date, salary structure, position, and job responsibilities to build a database of the applicant's basic information, extracting keyword data groups, and organizing the disputed matters in the arbitration application documents. NLP technology can be used to analyze the application documents, extract disputed keywords, and cluster these keywords to categorize the disputed matters (e.g., labor remuneration, termination of labor contract).

[0076] The disputed matters are mapped to relevant facts, evidence, and legal provisions. Knowledge graph technology is used to associate the disputed matters with relevant legal provisions, enabling automatic retrieval of laws, regulations, and judicial interpretations related to the case. Appropriate evidentiary materials are matched by comparing the factual descriptions in the application documents.

[0077] Analyze the applicant's arbitration claim. Use NLP technology to perform semantic analysis on the applicant's arbitration claim. Extract key elements of the arbitration claim, such as the matters claimed, the amount, and the method of compensation.

[0078] Establish a mapping relationship between the arbitration request and the disputed matters. Analyze the reasonableness and feasibility of the arbitration request and match it with the disputed matters.

[0079] It automatically generates a thought process chain for arbitration proceedings, including the applicant's basic information, the disputed matters, the arbitration request and corresponding facts, and finally the corresponding evidence and legal basis.

[0080] Through the above intelligent design, the thought process of an arbitrator can be simulated, thus automating the arbitration process. This method helps improve arbitration efficiency and ensures the fairness of the outcome.

[0081] Step 3 involves collecting the applicant's basic information during their employment period and then verifying the data. This verification process includes the following steps:

[0082] The data verification module uses optical character recognition technology to scan and verify the supporting documents provided by the applicant and the employee files provided by the company, and verifies the accuracy of the identification of basic information during the employment period based on the recognition results.

[0083] Data verification can determine the accuracy of information. It involves comparing data with employee records provided by the company. Optical Character Recognition (OCR) technology is used to scan and verify relevant supporting documents. The specific details of Optical Character Recognition (OCR) technology are as follows:

[0084] Adjust the image's brightness, contrast, and resolution to ensure the text is clearly legible. Use image processing techniques to remove noise and artifacts generated during scanning. If the scanned image is tilted, it needs to be corrected using software. Use OCR software to analyze the pre-processed image and recognize the text content. Convert the OCR-recognized text into an editable text format, such as a TXT or Word document. The OCR software should be able to recognize the text layout and distinguish between different parts such as titles, body text, and tables. Manually check whether the OCR output text matches the original scanned document to ensure the accuracy of the information. Correct any OCR-recognized errors, including spelling mistakes and formatting issues. Categorize the extracted text according to a certain structure, such as classifying pay slip information and overtime records separately. Integrate the OCR-processed data into the arbitration hearing system and link it with other case information.

[0085] Labor disputes include disputes concerning the confirmation of employment relationships, disputes concerning dismissal, resignation, and termination of employment, disputes concerning working hours and rest and leave, disputes concerning social insurance and welfare, disputes concerning labor remuneration and medical expenses for work-related injuries, and disputes concerning economic compensation and damages.

[0086] In step 2, the information extraction formula is:

[0087] ;

[0088] in This represents the information extraction function. This indicates the input application document. This represents the nth keyword extracted.

[0089] The following information verification formula is used to verify the applicant's basic information during their employment period:

[0090] ;

[0091] Where V represents the information verification function. This represents the keyword to be verified, and D represents the employee file provided by the company. This represents the verification result of the m-th keyword.

[0092] The formula for identifying disputed items in step 4 is:

[0093] ;

[0094] R is the disputed item identification function, and T is the text content. This represents the k-th disputed item.

[0095] In step 5, the following formula is used to link evidence to legal provisions:

[0096] ;

[0097] Where A represents the correlation function. Let F represent the k-th disputed item, F represent the set of facts, and L represent the set of legal provisions. This represents the p-th fact associated with the disputed term. This represents the p-th legal provision associated with the disputed item.

[0098] In step 6, the formula for matching the arbitration request with the disputed matters is as follows:

[0099] ;

[0100] M represents the function that matches the arbitration request with the disputed matters. Indicate the elements of the arbitration request, Indicates the disputed item, This indicates the matching result between the arbitration request and the disputed matters.

[0101] In step 4, the keyword extraction formula for the large language model is:

[0102] ;

[0103] Where K represents the keyword extraction function, and T represents the preprocessed text. Indicates model parameters, This represents the nth keyword;

[0104] The keyword clustering formula is:

[0105] ;

[0106] C represents the clustering function, and K represents the set of extracted keywords. Indicates the parameters of the clustering algorithm. This represents the m-th cluster category;

[0107] The formula for generating the list of disputed items is:

[0108] ;

[0109] L represents the function that generates the list of disputed items, and C represents the clustering result. This represents the generated list of the m-th disputed items.

[0110] The labor arbitration application documents are preprocessed to obtain cleaned text. Keywords are extracted from the text using a keyword extraction model (such as BERT). Clustering algorithms (such as K-means) are applied to cluster the keywords to determine the categories of disputed matters. Based on the clustering results, a list of disputed matters is generated.

[0111] A large amount of labeled data is needed to train the keyword extraction and clustering model before model training.

[0112] The model's performance depends on the accuracy of keyword extraction and clustering, so the model parameters need to be continuously optimized.

[0113] Labor dispute cases are complex and varied, and models may need to be updated regularly to adapt to new types of disputes.

[0114] Based on the above architecture and formulas, a large language model can be built that can process labor arbitration application documents, effectively extract dispute keywords, and classify disputed matters into categories.

[0115] The arbitration hearing thought chain inference device based on the big language model includes an information input module, an information extraction module, an information analysis module, and a result display module. The information input module is used to input the electronic application documents for labor arbitration and scan the supporting materials. The information extraction module is connected to the information extraction module and runs the information extraction program designed in step 2 of the arbitration hearing thought chain inference method based on the big language model to analyze the electronic application documents for labor arbitration and obtain the extracted keyword data group.

[0116] The information analysis module is connected to the information input module and the information extraction module. It uses the analysis program designed in steps 2 to 6 of the arbitration hearing thought chain inference method based on a large language model to obtain the arbitration hearing thought chain analysis results. The result display module communicates and interacts with the information analysis module to display the arbitration hearing thought chain analysis results.

[0117] The information input module can be a networked computer, the information extraction module and the information analysis module are computer-based extraction or analysis programs, and the result display module is a monitor.

[0118] The following specific embodiments illustrate the implementation principle of the arbitration hearing thought chain inference method and device based on a large language model of the present invention:

[0119] The applicant, Zhang San, joined the respondent, a certain technology company, on [Date] as a software engineer. Zhang San left the company on [Date]. Zhang San claims that the company failed to pay overtime as agreed and also failed to pay his year-end bonus upon his departure.

[0120] Summary of Arbitration Application Documents

[0121] Zhang San's start date is [Date] 2018.

[0122] Zhang San's departure date is [Date] 2024.

[0123] Zhang San's salary consists of a basic salary, performance bonus, and overtime pay.

[0124] Zhang San often worked overtime during his tenure, but the company did not pay overtime pay.

[0125] The company's regulation states that the year-end bonus is 10% of the employee's annual salary, but Zhang San did not receive the year-end bonus when he left the company.

[0126] Part 1: Organize the basic situation of the applicant during his tenure;

[0127] ;

[0128] Among them represents the information extraction function, represents the input application document, represents the nth keyword extracted.

[0129] Among them:

[0130] = "May 1, 2018"; b1 = "X / X / 2018" (start date of employment)

[0131] = "X / X / 2024"; = "October 31, 2023" (end date of employment)

[0132] = "basic salary + performance bonus + overtime pay"; = "basic salary + performance bonus + overtime pay" (composition structure of salary);

[0133] Part 2: Sort out the controversial issues of the case

[0134] Application of the core formula

[0135] ;

[0136] Among them, A represents the association function, represents the kth controversial item, F represents the set of facts, L represents the set of legal provisions, represents the pth fact associated with the controversial item, represents the pth legal provision associated with the controversial item.

[0137] Among them:

[0138] = "overtime pay not paid" d1 = "overtime pay not paid"

[0139] = "Zhang San's overtime record" f1 = "Zhang San's overtime record"

[0140] Overtime Application Form and Attendance Records

[0141] Article 41 of the Labor Law: Employers shall pay overtime wages for overtime work.

[0142] =\text{"Year-end bonus not paid"}d2="Year-end bonus not paid"

[0143] =\text{"Company Year-End Bonus Policy"}f2="Company Year-End Bonus Policy"

[0144] Company rules and regulations, employment contract.

[0145] =\text{"Article 47 of the Labor Contract Law: Employers shall pay remuneration as agreed"}l2="Article 47 of the Labor Contract Law: Employers shall pay remuneration as agreed"

[0146] Part Three: Automated Analysis of the Applicant's Arbitration Request

[0147] The labor arbitration application documents are preprocessed to obtain cleaned text. Keywords are extracted from the text using a keyword extraction model (such as BERT). Clustering algorithms (such as K-means) are applied to cluster the keywords to determine the categories of disputed matters. Based on the clustering results, a list of disputed matters is generated.

[0148] The final disputed items are: the corresponding arbitration requests are: "request for payment of overtime pay" and "request for payment of year-end bonus", and the corresponding disputed matters are "failure to pay overtime pay" and "failure to pay year-end bonus";

[0149] Using the above method, we obtained the following thought chain for arbitration proceedings: Zhang San's basic information is as follows: Zhang San's start date is [Date] 2018.

[0150] Zhang San's departure date is [Date] 2024.

[0151] Zhang San's salary consists of a basic salary, performance bonus, and overtime pay.

[0152] Zhang San frequently worked overtime during his employment, but the company did not pay him overtime wages.

[0153] The company stipulates that the year-end bonus is 10% of the employee's annual salary, but Zhang San did not receive the year-end bonus when he left the company.

[0154] The issues in dispute are: "unpaid overtime pay" and "unpaid year-end bonus";

[0155] The arbitration requests are: "to request payment of overtime pay" and "to request payment of year-end bonus";

[0156] The evidence includes: Zhang San's overtime records, overtime application forms, attendance records, and the company's year-end bonus policy;

[0157] The legal basis is Article 41 of the Labor Law and Article 47 of the Labor Contract Law.

[0158] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for inferring the thought chain in arbitration proceedings based on a large language model, characterized in that: Includes the following steps: Step 1: Enter the electronic application documents for labor arbitration into the information entry module; Step 2: The information extraction module uses natural language processing technology to parse the application documents and extract keyword data groups; Step 3: The information analysis module identifies and extracts the applicant's basic information during employment based on the keyword data group of text parsing. The applicant's basic information during employment includes: date of employment, date of resignation, salary structure, position and job responsibilities. Step 4: The information analysis module sorts out the disputed matters of the case, uses a large language model to analyze the application documents to extract disputed keywords, clusters the keywords, classifies the labor dispute matters into categories, and forms a list of disputed matters; Step 5: The information analysis module retrieves the relevant laws, regulations, and judicial interpretations for the case, uses knowledge graph technology to associate the disputed matters with legal provisions, compares the factual descriptions in the application documents, and matches appropriate evidentiary materials. Step 6: The information analysis module uses NLP technology to perform semantic analysis on the applicant's arbitration request and extract the key elements of the arbitration request; By mapping the arbitration request to the disputed matters, establishing a mapping relationship between the arbitration request and the disputed matters, analyzing the rationality and feasibility of the arbitration request, matching it with the disputed matters, and automatically generating an arbitration hearing thought chain, the arbitration hearing thought chain includes the applicant's basic information, the disputed matters, the arbitration request and the corresponding facts, evidence and legal basis.

2. The arbitration hearing thought chain inference method based on a large language model according to claim 1, characterized in that: Step 3 involves collecting the applicant's basic information during their employment period and then verifying the data. This verification process includes the following steps: The data verification module uses optical character recognition technology to scan and verify the supporting documents provided by the applicant and the employee files provided by the company, and verifies the accuracy of the identification of basic information during the employment period based on the recognition results.

3. The arbitration hearing thought chain inference method based on a large language model according to claim 2, characterized in that: Labor disputes include disputes concerning the confirmation of employment relationships, disputes concerning dismissal, resignation, and termination of employment, disputes concerning working hours and rest and leave, disputes concerning social insurance and welfare, disputes concerning labor remuneration and medical expenses for work-related injuries, and disputes concerning economic compensation and damages.

4. The arbitration hearing thought chain inference method based on a large language model according to claim 3, characterized in that: In step 2, the information extraction formula is: ； in This represents the information extraction function. This indicates the input application document. This represents the nth keyword extracted.

5. The arbitration hearing thought chain inference method based on a large language model according to claim 4, characterized in that: The following information verification formula is used to verify the applicant's basic information during their employment period: ； Where V represents the information verification function. This represents the keyword to be verified, and D represents the employee file provided by the company. This represents the verification result of the m-th keyword.

6. The arbitration hearing thought chain inference method based on a large language model according to claim 5, characterized in that: The formula for identifying disputed items in step 4 is: ； R is the disputed item identification function, and T is the text content. This represents the k-th disputed item.

7. The arbitration hearing thought chain inference method based on a large language model according to claim 6, characterized in that: In step 5, the following formula is used to link evidence to legal provisions: ； Where A represents the correlation function. Let F represent the k-th disputed item, F represent the set of facts, and L represent the set of legal provisions. This represents the p-th fact associated with the disputed term. This represents the p-th legal provision associated with the disputed item.

8. The method for inferring the thought chain of arbitration proceedings based on a large language model according to claim 7, characterized in that: In step 6, the formula for matching the arbitration request with the disputed matters is as follows: ； M represents the function that matches the arbitration request with the disputed matters. Indicate the elements of the arbitration request, Indicates the disputed item. This indicates the matching result between the arbitration request and the disputed matters.

9. The arbitration hearing thought chain inference method based on a large language model according to claim 8, characterized in that: In step 4, the keyword extraction formula for the large language model is: ； Where K represents the keyword extraction function, and T represents the preprocessed text. Indicates model parameters, This represents the nth keyword; The keyword clustering formula is: ； C represents the clustering function, and K represents the set of extracted keywords. Indicates the parameters of the clustering algorithm. This represents the m-th cluster category; The formula for generating the list of disputed items is: ； L represents the function that generates the list of disputed items, and C represents the clustering result. This represents the generated list of the m-th disputed items.

10. An arbitration hearing thought chain inference device based on a large language model, characterized in that: It includes an information input module, an information extraction module, an information analysis module, and a result display module. The information input module is used to input electronic application documents for labor arbitration and scan supporting materials. The information extraction module is connected to the information extraction module and runs the information extraction program designed in step 2 of the arbitration hearing thought chain inference method based on the large language model as described in claim 9 to analyze the electronic application documents for labor arbitration and obtain the extracted keyword data group. The information analysis module is communicatively connected to the information input module and the information extraction module, respectively. It adopts the analysis program designed in steps 2 to 6 of the arbitration hearing thought chain inference method based on the large language model as described in claim 9. The analysis program is run to obtain the arbitration hearing thought chain analysis results. The result display module communicates and interacts with the information analysis module to analyze the results and displays the arbitration hearing thought chain analysis results.

Images