A method and device for constructing civil dispute mediation data

Abstract

The present application relates to a kind of civil dispute mediation data construction method and device. By obtaining civil dispute case original data and being structured extraction, obtain case structured information;Create each party corresponding party intelligent agent and mediator corresponding mediator intelligent agent;Case structured information is input to each intelligent agent, so that it carries out the dialogue of multiple preset mediation stages according to information;After each mediation stage ends, the execution data of multiple rounds of dialogue is obtained and stage index is calculated;According to stage index and scheduling rule, subsequent mediation stage is scheduled;Finally, the mediation data of each mediation stage is stored. Through the method of the present application, dynamic stage scheduling is realized, the calculation amount of mediation data generation process can be flexibly adjusted, and reliable data support is provided for subsequent model training, evaluation and empirical research.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of computer technology, and in particular to a method and apparatus for constructing civil dispute mediation data. Background Technology

[0002] Civil dispute mediation, as an important alternative dispute resolution mechanism, plays a vital role in alleviating the conflict between too many parties and too few resources, reducing the cost of rights protection for litigants, and promoting harmony and stability. Compared with traditional litigation, the mediation process is more flexible, lower in cost, and better able to balance legal and social effects.

[0003] With the development of artificial intelligence, large-scale AI models are gradually being applied to civil dispute mediation. Training these models requires a large amount of high-quality mediation process data. However, since many mediation activities take place in closed, offline settings and involve the privacy of the parties involved, mediation dialogues are often not fully recorded or made public. Even when they are, the public content usually only retains the facts of the case and the outcome. The details of the game and interaction between the parties during the mediation process are omitted, making it difficult to analyze the patterns of the mediation process on large-scale real data and thus hindering the improvement of the effectiveness of large-scale AI models in civil dispute mediation. Summary of the Invention

[0004] To address the problems existing in the prior art, this specification provides a method and apparatus for constructing civil dispute mediation data. Starting from real civil dispute documents, it automatically generates structured mediation inputs, and further achieves dynamic stage scheduling through stage execution budget and indicator monitoring. It ensures consistency and legality of the scheme through deviation judgment and error correction and regeneration, and finally forms high-quality mediation process data that can be used for screening, storage and controllable mechanism research.

[0005] The specific technical solutions of the embodiments in this specification are as follows: On the one hand, the embodiments of this specification provide a method for constructing civil dispute mediation data, the method including: Obtain the original data of civil dispute cases, and perform structured extraction on the original data of civil dispute cases to obtain structured case information; Create a party agent corresponding to each party and a mediator agent corresponding to each mediator in the structured information of the case; The structured case information is input to the parties' intelligent agents and the mediator's intelligent agent, so that the parties' intelligent agents and the mediator's intelligent agent can conduct dialogues in multiple preset mediation stages based on the structured case information; After receiving the dialogue end message of each mediation stage, the execution data of the multi-round dialogue in that mediation stage is obtained, and the stage index of that mediation stage is calculated based on the execution data of the multi-round dialogue; when there is a subsequent mediation stage in that mediation stage, the subsequent mediation stage is scheduled according to the stage index and the stage scheduling rules. Store mediation data for each stage of the mediation process.

[0006] Furthermore, scheduling subsequent mediation stages of the mediation stage according to the aforementioned stage indicators and stage scheduling rules further includes: Based on the stage indicators and the stage scheduling rules, determine whether to skip subsequent mediation stages of this mediation stage; or, The execution budget for subsequent mediation phases of the mediation phase is adjusted based on the phase indicators and the phase scheduling rules.

[0007] Furthermore, the execution data includes: the satisfaction level of the party's intelligent agent and the mediator's intelligent agent, the degree of consensus on key issues in the structured information of the case, the litigation risk of the party's intelligent agent, and the degree of first coverage of the legal basis used by the mediator's intelligent agent with the legal basis in the structured information of the case; The phase indicators for this mediation stage, calculated based on data from multiple rounds of dialogue, further include: The satisfaction level, litigation risk, consensus level, and first coverage level mentioned in the multiple rounds of dialogue were normalized to obtain the satisfaction level index, litigation risk index, consensus level index, and first coverage level index for this mediation stage.

[0008] Furthermore, adjusting the execution budget for subsequent mediation stages based on the aforementioned stage indicators and stage scheduling rules further includes: The adjustment amount and direction of the execution budget for the subsequent mediation stage are determined based on the aforementioned satisfaction index, litigation risk index, consensus index, and first coverage index. The execution budget for the subsequent mediation phase is adjusted according to the stated adjustment amount and direction.

[0009] Furthermore, after calculating the phase indicators for this mediation phase based on the execution data of multiple rounds of dialogue, the method further includes: The decision on whether to revert to enforcement at this mediation stage is based on the satisfaction index, litigation risk index, consensus index, and first coverage index mentioned in the multiple rounds of dialogue.

[0010] Furthermore, the method also includes: After receiving the end message of the last mediation stage, the mediation success rate is calculated based on the acceptance of the final solution obtained after the end of the last mediation stage by each party's intelligent agent. The alignment degree of the dispute focus between the first set of dispute focus generated by the mediator intelligent agent after the end of each mediation stage and the second set of dispute focus in the case structure information is calculated. The second coverage degree of the legal basis set used by the mediator intelligent agent after the end of each mediation stage to the legal basis set in the case structure information is also calculated. If the mediation success rate does not meet the predetermined success rate requirement, the alignment of the dispute focus does not meet the predetermined alignment requirement, and / or the second coverage does not meet the predetermined coverage requirement, the parties' agents and the mediator's agent are controlled to re-engage in multiple preset mediation stages based on the case structured information.

[0011] Furthermore, the multiple preset mediation stages include at least a solution generation stage and a bargaining stage; Scheduling subsequent mediation stages based on the aforementioned stage indicators and stage scheduling rules further includes: The bargaining phase is scheduled according to the phase indicators and phase scheduling rules of the scheme generation phase.

[0012] Furthermore, creating the party agent corresponding to each party and the mediator agent corresponding to the mediator in the structured case information further includes: Create multiple party intelligent agents for each party and multiple mediator intelligent agents for the mediator; The method further includes: Configure the behavioral strategies of each party's intelligent agent. The parameters of the behavioral strategies include the degree of self-assertion and the degree of cooperation. The behavioral strategies of different parties' intelligent agents are different. Configure the knowledge reserve strategy for each mediator agent, and the knowledge reserve strategies for different mediator agents are different; Multiple mediation teams are obtained by combining multiple party agents with different behavioral strategies and multiple mediator agents with different knowledge reserve strategies. The input of the structured case information to the parties' intelligent agents and the mediator's intelligent agent further includes: The structured case information is input into each mediation group so that the parties' agents and the mediator's agents in each mediation group can conduct dialogues in multiple preset mediation stages based on the structured case information.

[0013] Furthermore, the method also includes: After each round of dialogue, reflect on and evaluate the state of the current round of dialogue; Based on the results of the aforementioned reflection and evaluation, the degree of self-assertion and cooperation of the party's intelligent agent are adjusted.

[0014] On the other hand, embodiments of this specification also provide an apparatus for constructing civil dispute mediation data, the apparatus comprising: The raw data acquisition and processing unit is used to acquire raw data of civil dispute cases and perform structured extraction on the raw data of civil dispute cases to obtain structured case information. The agent creation unit is used to create the party agent corresponding to each party in the case structured information and the mediator agent corresponding to the mediator. The dialogue simulation unit is used to input the structured information of the case to each party's intelligent agent and the mediator's intelligent agent, so that the party's intelligent agent and the mediator's intelligent agent can conduct dialogues in multiple preset mediation stages based on the structured information of the case; The mediation stage scheduling unit is used to obtain the execution data of the multi-round dialogue in each mediation stage after receiving the dialogue end message of each mediation stage, and calculate the stage index of the mediation stage based on the execution data of the multi-round dialogue; when there is a subsequent mediation stage in the mediation stage, the subsequent mediation stage of the mediation stage is scheduled according to the stage index and the stage scheduling rules. The mediation data storage unit is used to store mediation data at each stage of the mediation process.

[0015] The beneficial effects of the embodiments in this specification are as follows: by constructing a structured dispute representation and adopting a multi-stage mediation framework, the problem of obtaining high-quality mediation process data is solved, enabling the construction of large-scale civil dispute mediation process data under the premise of privacy compliance; by establishing a stage execution budget and monitoring mechanism, the controllability of the mediation simulation process is achieved, avoiding problems such as long-term interaction instability and stage delays; by constructing a dual evaluation index system and establishing a deviation judgment and error correction regeneration mechanism, the factual consistency, dispute coverage, and legal adaptability of the generated data are ensured, significantly improving the data availability; by realizing dynamic stage scheduling, the computational workload of the mediation data generation process can be flexibly adjusted, providing reliable data support for subsequent model training, evaluation, and empirical research. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 The diagram shown is a flowchart illustrating a method for constructing civil dispute mediation data in an embodiment of this specification. Figure 2 The diagram shown is a flowchart illustrating how the execution budget of the subsequent mediation stage is adjusted according to the stage indicators and the stage scheduling rules in an embodiment of this specification. Figure 3 The diagram shown is a flowchart illustrating the quality assessment of the final solution obtained in an embodiment of this specification. Figure 4 The diagram shown is a flowchart illustrating a method for constructing civil dispute mediation data in another embodiment of this specification. Figure 5 The diagram shown is a structural schematic of a device for constructing civil dispute mediation data according to an embodiment of this specification. Figure 6 The diagram shown is a structural schematic of the computer device in an embodiment of this specification.

[0018] [Explanation of Figure Markers]: 501. Raw data acquisition and processing unit; 502. Agent Creation Unit; 503. Dialogue Simulation Unit; 504. Mediation Phase Dispatch Unit; 505. Adjustment data storage unit; 602. Computer equipment; 604, Processor; 606. Memory; 608. Drive mechanism; 610. Input / output module; 612. Input devices; 614. Output devices; 616. Presentation equipment; 618. Graphical User Interface; 620. Network interface; 622. Communication link; 624. Communication bus. Detailed Implementation

[0019] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the embodiments of this specification.

[0020] It should be noted that the terms "first," "second," etc., in the description, claims, and accompanying drawings of the embodiments herein are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, apparatus, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0021] It should be noted that the acquisition, storage, use, and processing of data in the technical solutions of the embodiments of this specification all comply with the relevant provisions of national laws and regulations.

[0022] It should be noted that in the embodiments of this specification, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, they do not mean that the applicant has used or necessarily used the solution.

[0023] To address the problems existing in the prior art, this specification provides a method for constructing civil dispute mediation data. Starting from real civil dispute documents, it automatically generates structured mediation inputs, and further achieves dynamic stage scheduling through stage execution budgets and indicator monitoring. It ensures consistency and legality of the scheme through deviation judgment and error correction and regeneration, and finally forms high-quality mediation process data that can be used for screening, database entry, and controllable mechanism research. Figure 1 The diagram illustrates a flowchart of a method for constructing civil dispute mediation data according to an embodiment of this specification. The process of constructing mediation data is described in this diagram. The order of steps listed in the embodiment is merely one possible execution order among many steps and does not represent the only possible execution order. In actual system or device products, the methods shown in the embodiment or the accompanying drawings can be executed sequentially or in parallel. Specifically, as shown... Figure 1 As shown, the method can be executed by a server, and the method may include: Step 101: Obtain the original data of civil dispute cases, and perform structured extraction on the original data of civil dispute cases to obtain structured case information; Step 102: Create the party agent corresponding to each party in the structured information of the case and the mediator agent corresponding to the mediator; Step 103: Input the structured case information into the intelligent agents of each party and the intelligent agent of the mediator, so that the intelligent agents of the parties and the intelligent agent of the mediator can conduct dialogues in multiple preset mediation stages based on the structured case information; Step 104: After receiving the dialogue end message of each mediation stage, obtain the execution data of the multi-round dialogue in the mediation stage, and calculate the stage index of the mediation stage based on the execution data of the multi-round dialogue; when the mediation stage has a subsequent mediation stage, schedule the subsequent mediation stage of the mediation stage according to the stage index and the stage scheduling rules. Step 105: Store the mediation data for each stage of mediation.

[0024] The beneficial effects of the embodiments in this specification are as follows: by constructing a structured dispute representation and adopting a multi-stage mediation framework, the problem of obtaining high-quality mediation process data is solved, enabling the construction of large-scale civil dispute mediation process data under the premise of privacy compliance; by establishing a stage execution budget and monitoring mechanism, the controllability of the mediation simulation process is achieved, avoiding problems such as long-term interaction instability and stage delays; by constructing a dual evaluation index system and establishing a deviation judgment and error correction regeneration mechanism, the factual consistency, dispute coverage, and legal adaptability of the generated data are ensured, significantly improving the data availability; by realizing dynamic stage scheduling, the computational workload of the mediation data generation process can be flexibly adjusted, providing reliable data support for subsequent model training, evaluation, and empirical research.

[0025] In the embodiments of this specification, raw data is obtained from a publicly available civil dispute case database. The original dispute set is denoted as D = {D1, D2, …, DN}. Each Di is represented in the following form: Di = (Titlei, Keywordsi, Briefi, Methodi, Basesi), where: Titlei represents the case title; Keywordsi is the set of keywords or cause-of-fact tags for the case; Briefi represents a brief description of the case (a brief explanation of the dispute); Methodi represents a description of the case handling method and process (excluding specific mediation dialogue); and Basesi represents a set of legal basis clauses, either manually annotated or attached to documents.

[0026] Then, the Briefi, Methodi, and Basesi in each Di are anonymized: sensitive information such as the party's name, ID number, contact information, and detailed address are identified and replaced with codes; fields such as amount and time are processed into ranges or fuzzy fields as needed.

[0027] Then, using an information extraction model or a large language model, extract from Di: Dispute Type DTi: Extract or summarize the type of case from Keywordssi (such as "labor contract dispute", "house sale contract dispute", etc.). Dispute Summary DBi: Organizes the background information related to mediation in the Brieffi into a concise and coherent case description; Dispute Facts DFi: Objective factual statements extracted from Methodi, denoted as the fact set DFi = {DFi,1,…,DFi,mi}; The set of parties, DPi: Identify the roles of the parties, such as plaintiff, defendant, third party, etc., from the Briefi, and obtain DPi = {Party i,1, …, Party i,ki}; Points of contention (DPointsi): Extract the main points of contention from the Brief and Method, such as "whether liability for breach of contract exists" and "the amount of compensation." Legal Basis DBasesi: Standardizes the legal provisions in Basesi to obtain a structured set of legal clauses.

[0028] The extracted structured representation is: Di = (DTi, DBi, DFi, DPi, DPointsi, DBasesi).

[0029] This specification's embodiments employ a five-stage mediation framework, denoted as: preparatory stage, presentation stage, solution generation stage, bargaining stage, and closing stage. It should be noted that the above stages are a process framework and do not limit each stage to a fixed number of rounds / fixed duration; the system can dynamically schedule the execution of each stage based on dialogue progress and evaluation indicators, including shortening, extending, skipping, or reverting stages.

[0030] Introducing a "five-stage mediation framework" (including a preparatory stage, a presentation stage, a solution generation stage, a bargaining stage, and a closing stage) into simulated mediation between party agents and mediator agents adds a controllable process skeleton to open-ended long dialogues, transforming mediation from "free dialogue" to a procedural collaboration "driven by objectives." The explicit division of stage objectives effectively suppresses objective drift, digressions, and redundant chatter in multi-agent dialogues, guiding the dialogue to gradually converge around clarifying disputes, forming solutions, exchanging conditions, and solidifying outcomes, thereby improving the stability and interpretability of the generated trajectory.

[0031] The five-stage approach provides clear boundaries for stage-level dynamic control. The system can set execution budgets and constraints for each stage (such as maximum number of rounds, maximum number of searches, etc.) and dynamically schedule stages (shortening, skipping, rolling back, etc.) based on stage progress and quality monitoring results. This reduces delays and ineffective interactions, avoids repeated tug-of-wars in low-value stages, and improves overall generation efficiency and data output stability.

[0032] In terms of quality assurance, the phased structure makes the closed loop of "assessment-deviation determination-error correction-rollback" easier to implement. Because the output form and quality requirements of each stage are more clearly defined, the system can calculate mediation effect indicators and solution quality indicators at key nodes; when inconsistencies in facts, missing points of contention, mismatched legal basis, or results deviating from reasonable expectations occur, the system can quickly locate the stage to which the problem belongs and trigger regeneration or stage rollback correction, thereby improving consistency and availability.

[0033] The five-stage approach also significantly enhances the structuring of the data and its value for training and evaluation. Stage labels, stage logs, and inter-stage transition relationships provide stable anchors for annotation, modeling, and evaluation: training can be performed in a decompositional manner by stage (such as dispute extraction, solution generation, and legal citation), and diagnostic evaluation can be conducted by stage (determining whether failure occurred in fact clarification, solution construction, or convergence negotiation), thereby improving the operability of the data in model iteration and capability assessment.

[0034] Finally, it should be emphasized that these five stages can improve the authenticity and credibility of simulation data and provide a unified and scalable technical foundation for controlled comparison and reproducible research.

[0035] In the mediation process simulation, the main roles include: the parties involved and the mediator. The parties involved represent the positions and demands of each party, while the mediator is responsible for presiding over the proceedings, clarifying the facts, and proposing solutions.

[0036] Therefore, the embodiments of this specification create a party agent corresponding to each party and a mediator agent corresponding to each mediator in the structured information of the case.

[0037] The structured case information is input into the intelligent agents of each party and the mediator, enabling them to conduct the five mediation stages sequentially based on this information. Mediation data for each stage can be stored after the dialogue concludes. A final solution is obtained after the concluding dialogue, and this final solution is also stored after its conclusion. The mediation data from each stage and the final solution stored after the concluding dialogue serve as training data for a large-scale AI model for civil dispute mediation.

[0038] In the embodiments of this specification, when a mediation stage reaches a preset termination condition, the system receives a dialogue termination message for that mediation stage. The system acquires execution data from all rounds of dialogue in that mediation stage, including dialogue rounds, content of statements, emotional tone, and coverage of points of contention. Based on the execution data of the multiple rounds of dialogue, the system calculates stage indicators for that mediation stage, including quantitative evaluation data such as stage completion rate, coverage rate of points of contention, party participation rate, and mediation efficiency indicators.

[0039] Then, based on the current stage indicators and preset stage scheduling rules, the subsequent mediation stages of this mediation stage are scheduled. Stage scheduling rules may include, for example, stage advancement conditions, anomaly handling mechanisms, and quality control standards. When the stage indicators meet the advancement conditions, the system automatically starts the next mediation stage; when an anomaly occurs in the stage indicators, a regeneration mechanism is triggered or the agent's policy parameters are adjusted for re-execution.

[0040] The above methods enable the systematic construction of high-quality, controllable, and assessable data on the civil dispute mediation process, providing a reliable data foundation for empirical research, model training, and the development of intelligent mediation systems in the field of mediation.

[0041] According to one embodiment of this specification, scheduling subsequent mediation stages of the mediation stage based on the stated stage indicators and stage scheduling rules further includes: Based on the stage indicators and the stage scheduling rules, determine whether to skip subsequent mediation stages of this mediation stage; or, The execution budget for subsequent mediation phases of the mediation phase is adjusted based on the phase indicators and the phase scheduling rules.

[0042] In the embodiments described in this specification, a phase indicator monitoring system is first established, including indicators such as phase execution efficiency, coverage of disputed issues, party participation, and mediation progress quality. The system also sets up a phase scheduling rule base, including skip rules and budget adjustment rules, to guide the dynamic scheduling of subsequent mediation phases.

[0043] For example, once a mediation phase is completed, various indicator data for that phase can be collected.

[0044] The efficiency index for each stage is calculated by statistically analyzing parameters such as the number of rounds, the proportion of effective speeches, and the percentage of duplicate content in that stage.

[0045] The coverage index of disputed points is determined based on the ratio of the number of disputed points involved in this stage to the total number of disputed points.

[0046] The participant engagement index is comprehensively assessed based on factors such as the frequency of speaking by both parties, their responsiveness, and changes in attitude. The mediation progress quality index is quantified by analyzing dimensions such as the degree of consensus reached at this stage, the extent to which differences have been narrowed, and the improvement in the feasibility of the proposed solutions.

[0047] Based on the collected stage indicators, the system uses preset stage scheduling rules to determine which stage to skip. For example, when the stage execution efficiency indicator is below a set threshold and the dispute focus coverage indicator has reached the corresponding threshold, the system determines that the mediation stage has fully fulfilled its expected function and can skip the planned subsequent detailed discussion stage. When the party participation indicator continues to decline below the corresponding threshold and the mediation progress quality indicator stagnates, the system determines that continuing the current type of mediation stage will have limited effect and decides to skip subsequent mediation stages of the same type and directly enter a mediation stage of a different nature.

[0048] For subsequent mediation stages that are not skipped, the execution budget is dynamically adjusted based on the current stage's performance indicators. When the coverage indicator of disputed points is low, the system will increase the execution budget for the subsequent dispute resolution stage, for example, by increasing the originally planned number of dialogue rounds to ensure full discussion of all disputed points. When the mediation progress quality indicator is high, the execution budget for the subsequent solution refinement stage will be reduced accordingly, and the originally planned number of dialogue rounds will be shortened to avoid inefficiency caused by excessive discussion.

[0049] In addition, a feedback mechanism for budget adjustments can be established. When the execution efficiency index is high at a certain stage, it indicates that the current budget allocation may be too generous, and the budget for subsequent stages of similar nature will be appropriately reduced. Conversely, when the execution efficiency index is too low, the system will increase the budget reserve for subsequent stages, ensuring sufficient time and space for discussion.

[0050] Through this dynamic adjustment mechanism based on phase indicators and scheduling rules, the mediation process can flexibly adjust its execution strategy according to the actual progress, avoiding wasted time in ineffective phases while ensuring that key issues are fully discussed, significantly improving the overall efficiency and quality of mediation. This method enables the mediation process to be adaptive, allowing for personalized phase arrangements based on the characteristics of different cases and the behavioral strategies of the parties involved, thus achieving optimal allocation of mediation resources.

[0051] Preferably, in this embodiment of the specification, the bargaining stage is scheduled according to the stage indicators and stage scheduling rules of the scheme generation stage.

[0052] The solution generation phase is a crucial part of the mediation process. In this phase, the system, based on structured dispute case input, analyzes the points of contention and relevant legal basis using a large language model agent to generate a targeted mediation solution. This phase has specific performance indicators, such as the completeness of the solution generation, legal suitability, and coverage of the points of contention. Simultaneously, the solution generation phase is configured with corresponding phase scheduling rules, which define the execution budget, quality monitoring standards, and trigger conditions for transitioning to the next phase.

[0053] The bargaining phase is the core interactive stage in the mediation process, where the parties' intelligent agents engage in specific negotiations and discussions. The initiation and execution of this phase are directly influenced by the output of the solution generation phase. The system dynamically adjusts the parameter settings and execution strategies of the bargaining phase based on the performance of the phase indicators from the solution generation phase.

[0054] When the completeness index of the solution generation stage reaches the preset threshold, the system automatically triggers the start of the negotiation stage; based on the evaluation results of the coverage of the disputed issues in the solution generation stage, the system adjusts the strategic focus and issue priority of the parties' intelligent agents in the negotiation stage; based on the legal adaptability score of the solution generation stage, the system sets corresponding compliance constraints for the negotiation stage.

[0055] Key information passed from the proposal generation phase to the negotiation phase can include the content of the generated mediation proposal, the analysis results of the points of contention, and the relevant legal basis. The scheduling rules can also define the execution budget allocation strategy for the negotiation phase, dynamically adjusting the upper limit of rounds and the time budget for the negotiation phase based on the complexity assessment results from the proposal generation phase.

[0056] During actual execution, the system continuously monitors changes in metrics during the scheme generation phase. When metrics become abnormal or deviate, the scheduling rules trigger a corresponding error correction mechanism to adjust the execution strategy for the negotiation phase in real time. This control mechanism based on phase metrics and scheduling rules ensures the stability and controllability of the mediation process, avoiding the phase delays and target drift problems common in traditional methods.

[0057] Preferably, the satisfaction levels of the party intelligence agent and the mediator intelligence agent in the embodiments of this specification, the degree of consensus on key issues in the structured information of the case, the litigation risk of the party intelligence agent, and the degree of first coverage of the legal basis used by the mediator intelligence agent with the legal basis in the structured information of the case are considered.

[0058] The phase indicators for this mediation stage, calculated based on data from multiple rounds of dialogue, further include: The satisfaction level, litigation risk, consensus level, and first coverage level mentioned in the multiple rounds of dialogue were normalized to obtain the satisfaction level index, litigation risk index, consensus level index, and first coverage level index for this mediation stage.

[0059] In the embodiments described in this specification, satisfaction level and consensus level can be directly output during the dialogue between the party's AI agent and the mediator's AI agent. Litigation risk is assessed from a third-party perspective as the risk of the dispute escalating into formal litigation or serious confrontation. For example, satisfaction level is quantified by analyzing the emotional tendencies, acceptance, and cooperation levels expressed by the parties in the dialogue. The system identifies the parties' reactions to the mediator's proposed solutions or suggestions and calculates the ratio of positive to negative responses. Litigation risk is assessed based on the frequency of the parties' threats of litigation, their level of concern about the consequences of mediation failure, and the number of times legal proceedings are mentioned. For example, when the parties frequently mention keywords such as "lawsuit," this indicator value will increase accordingly. Consensus level is calculated by detecting the degree of closeness in the positions of both parties on the points of contention. The system analyzes the trajectory of changes in the statements of both parties on core issues and identifies trends of softening positions, increased willingness to compromise, or widening differences. The first coverage indicator measures the completeness of the discussion of the main points of contention in the case at the current mediation stage, ensuring that important issues are not omitted or avoided.

[0060] Satisfaction level, consensus level, and litigation risk can all be numerical data. Therefore, the embodiments of this specification can separately set scales for satisfaction level, consensus level, and litigation risk, and set mapping values ​​corresponding to each level. For example, five levels can be set as Very Low, Low, Medium, High, and Very High, which are mapped to 0, 1, 2, 3, and 4 respectively. The levels corresponding to satisfaction level, consensus level, and litigation risk in their respective scales are determined, thereby determining the mapping values ​​corresponding to each of satisfaction level, consensus level, and litigation risk, and determining the number of times each level corresponding to satisfaction level, consensus level, and litigation risk is selected. Then, normalization is performed using the following formula:

[0061] Where s is the index value obtained after normalization, and c i This indicates the number of times level i was selected.

[0062] By applying the above formulas to the satisfaction level, consensus level, and litigation risk respectively, we obtain the satisfaction level index, litigation risk index, and consensus level index.

[0063] To determine the first level of coverage, we can compare the legal provisions manually annotated in the structured information of the case with those proposed by the mediator's AI agent, and use the Recall metric (which can be calculated after merging synonymous provisions / equivalent clauses if necessary) to obtain the first level of coverage, i.e., the coverage rate of legal basis.

[0064] Furthermore, such as Figure 2 As shown, adjusting the execution budget for subsequent mediation stages based on the stage indicators and stage scheduling rules further includes: Step 201: Determine the adjustment amount and direction of the execution budget for the subsequent mediation stage based on the satisfaction level index, litigation risk index, consensus level index, and first coverage level index; Step 202: Adjust the execution budget for the subsequent mediation phase according to the adjustment amount and direction.

[0065] In the embodiments of this specification, an execution budget adjustment model, such as a neural network model or a logistic regression model, can be pre-trained. The model takes the satisfaction index, litigation risk index, consensus index, first coverage index, and the initial value of the execution budget as inputs and outputs the adjustment amount and direction of the execution budget.

[0066] Then, the execution budget for the subsequent mediation phase is adjusted according to the stated adjustment amount and direction.

[0067] In the embodiments described in this specification, the budget adjustment process has dynamic feedback characteristics. Therefore, after each mediation stage, various indicators can be re-evaluated, and the budget for the next stage can be adjusted accordingly, forming an adaptive resource optimization cycle that ensures that the mediation process is both efficient and sufficient.

[0068] According to one embodiment of this specification, it can also be determined whether the mediation stage should be reverted to re-enforcement based on the satisfaction index, litigation risk index, consensus index, and first coverage index of the multi-round dialogue.

[0069] In the embodiments of this specification, when the satisfaction index is lower than the preset lower limit, it indicates that the parties are dissatisfied with the current progress of the mediation; when the litigation risk index exceeds the warning line, it indicates that the possibility of the mediation breaking down is increasing; when the consensus index shows that the differences between the two parties are widening rather than narrowing, it indicates that there may be a deviation in the direction of the mediation; when the first coverage index is insufficient, it means that important points of contention have not been fully discussed.

[0070] Therefore, the embodiments of this specification can employ a multi-dimensional comprehensive evaluation mechanism to calculate the weighted scores of each indicator and analyze the correlation between the indicators. For example, even if the satisfaction index is slightly low, if the consensus index shows an upward trend and the first coverage index meets the standard, the system may determine to continue the current stage. Conversely, if multiple indicators simultaneously show a negative trend, especially if the litigation risk index rises sharply while the consensus index continues to decline, the system will trigger a rollback and re-execution mechanism.

[0071] When the system determines that a rollback and re-execution are necessary, it analyzes the specific reasons for the abnormal indicators, identifying issues such as inappropriate mediation strategies, misjudgment of the key points of contention, or failure of emotional management by the parties involved. Based on the problem diagnosis results, the system adjusts the mediator's guidance strategy, re-organizes the discussion order of the key points of contention, or changes the parties' communication methods, and then restarts the mediation phase.

[0072] In some other embodiments of this specification, a detailed decision log can also be generated to record the specific values ​​of each indicator, the basis for judgment, and the reasons for rollback, providing data support for subsequent mediation quality analysis and strategy optimization.

[0073] According to one embodiment of this specification, a deviation judgment mechanism based on mediation effectiveness and solution quality indicators is established. When inconsistencies in facts, insufficient coverage of disputed points, or legal applicability issues are detected, an error correction and regeneration process is automatically triggered to ensure the quality and usability of the generated data. Specifically, after the dialogue in the last mediation stage ends, it is determined whether the final solution obtained meets the requirements. If it does not meet the requirements, the parties' intelligent agents and the mediator's intelligent agents can be controlled to conduct multiple rounds of dialogue in each mediation stage until a final solution that meets the requirements is obtained. Specifically, as shown below... Figure 3 As shown, the steps for quality assessment of the final solution include: Step 301: In response to the dialogue end message of the last mediation stage, calculate the mediation success rate based on the acceptance of the final solution obtained after the dialogue of the last mediation stage by each party's intelligent agent, calculate the alignment degree of the dispute focus between the first dispute focus set generated by the mediator intelligent agent after the dialogue of each mediation stage and the second dispute focus set in the case structure information, and calculate the second coverage degree of the legal basis set used by the mediator intelligent agent after the dialogue of each mediation stage to the legal basis set in the case structure information. Step 302: In response to the mediation success rate not meeting the predetermined success rate requirement, the alignment of the dispute focus not meeting the predetermined alignment requirement, and / or the second coverage not meeting the predetermined coverage requirement, control each party's intelligent agent and the mediator's intelligent agent to re-engage in multiple preset mediation stages of dialogue based on the case structured information.

[0074] In the embodiments described in this specification, the acceptance of the final solution obtained after the dialogue in the last mediation stage by each party's intelligent agent can be analyzed. The ratio of the number of parties accepting the solution to the total number of parties can be calculated to determine the mediation success rate. This success rate reflects the overall effectiveness of the mediation process and provides a key indicator for subsequent quality assessment.

[0075] The system can extract the first set of disputed issues generated by the mediator's agent after each stage of mediation, and simultaneously obtain the second set of disputed issues manually labeled from the case's structured information. Using semantic matching algorithms and keyword comparison technology, the overlap and similarity between the two sets of disputed issues are calculated to derive a numerical value for disputed issue alignment. This metric is used to evaluate whether the mediator's agent accurately identified and addressed the core disputed issues in the case, ensuring that the mediation process focuses on key points of contention.

[0076] This system can collect the set of legal bases cited and used by the mediator agent during the mediation process after each stage of dialogue, including relevant legal provisions, judicial interpretations, and precedents. This set of legal bases is then compared and analyzed with the set of legal bases manually labeled in the case's structured information to calculate the second degree of coverage of the legal bases used by the mediator agent with the standard set of legal bases. This coverage reflects the completeness and accuracy of the application of law during the mediation process.

[0077] This specification allows for the setting of predetermined success rate requirements, predetermined alignment requirements, and predetermined coverage requirements as quality assessment standards. When the mediation success rate fails to meet the predetermined success rate requirement, it indicates poor mediation effectiveness; when the alignment of disputed issues fails to meet the predetermined alignment requirement, it indicates improper handling of key disputed issues; and when the second coverage requirement fails to meet the predetermined coverage requirement, it indicates insufficient application of legal basis. In response to any one or more of the above conditions occurring simultaneously, the system automatically triggers a re-mediation mechanism, controlling the parties' intelligent agents and the mediator's intelligent agent to re-engage in multiple preset mediation stages based on the case's structured information, ensuring that the mediation quality meets the expected standards.

[0078] This can be understood as follows: by establishing a multi-dimensional quality assessment system, dynamic monitoring and automatic error correction of the mediation process are achieved, effectively improving the reliability and practicality of the intelligent mediation system. When the mediation quality is detected to be substandard, the system can promptly initiate a re-mediation process, avoiding the generation of low-quality mediation results and ensuring the high-quality generation of mediation data.

[0079] In one specific embodiment of this specification, in order to improve the applicability of the subsequent civil dispute mediation model, this embodiment of the specification can construct more mediation data for a single case. Therefore, this embodiment of the specification creates multiple party agents for each party and multiple mediator agents for the mediator.

[0080] To broaden the scope of mediation data, such as Figure 4 As shown, the method further includes: Step 401: Configure the behavioral strategies of each party's intelligent agent. The parameters of the behavioral strategies include the degree of self-assertion and the degree of cooperation. The behavioral strategies of different parties' intelligent agents are different. Step 402: Configure the knowledge reserve strategy for each mediator agent; different mediator agents will have different knowledge reserve strategies. Step 403: Combine multiple party agents with different behavioral strategies and multiple mediator agents with different knowledge reserve strategies to obtain multiple mediation groups; The input of the structured case information to the parties' intelligent agents and the mediator's intelligent agent further includes: Step 404: Input the structured case information into each mediation group so that the parties' agents and the mediator's agents in each mediation group can conduct dialogues in multiple preset mediation stages based on the structured case information.

[0081] In the embodiments described in this specification, the degree of self-assertion reflects the strength with which the party agent insists on its own position and demands during the mediation process, while the degree of cooperation reflects the party agent's tendency to negotiate and compromise with the other party. Different party agents are configured with different combinations of behavioral strategies, such as a strong agent with a high degree of self-assertion and a low degree of cooperation, or a cooperative agent with a moderate degree of self-assertion and a high degree of cooperation. This differentiated strategy configuration allows each agent to exhibit different negotiation styles and behavioral characteristics in the mediation dialogue.

[0082] The knowledge reserve strategy of the mediator's intelligent agent encompasses multiple dimensions, including mastery of legal provisions, ability to apply mediation skills, and accumulation of case experience. Different mediator intelligent agents are configured with different knowledge reserve strategies, such as having external knowledge and not having external knowledge. Having external knowledge means that the mediator intelligent agent can call the retrieval module to search for a set of candidate legal bases to assist in proposing solutions. Not having external knowledge means that the mediator intelligent agent can cite the law by utilizing the legal provisions in the structured information of the case and its own parameter-based knowledge. This differentiated configuration of knowledge reserves ensures that disputes can be addressed from multiple angles and levels during the mediation process.

[0083] Multiple mediation teams are formed by combining party agents with different behavioral strategies and mediator agents with different knowledge reserves. Each mediation team consists of party agents with specific behavioral strategies and mediator agents with specific knowledge reserves, forming a unique mediation combination mode. Through systematic combination and configuration, diverse mediation teams covering various possible mediation scenarios can be generated, providing rich experimental conditions for subsequent mediation simulations.

[0084] Structured case information is input into each mediation team, enabling the parties' and mediators' agents within each team to engage in dialogue across multiple pre-defined mediation stages based on this information. The structured case information includes key elements such as the facts of the dispute, the points of contention, and the legal basis, providing the agents with a unified case background and reference. Each mediation team, using the same case information but employing different strategy combinations, conducts mediation dialogues through multiple rounds of interactive communication, progressively advancing the mediation process and seeking solutions to the dispute.

[0085] During mediation, the parties' intelligent agents, based on their configured behavioral strategy parameters, exhibit corresponding self-assertion and cooperative tendencies in the dialogue. Agents with a high degree of self-assertion are more insistent on their own demands and show greater caution and conditional constraints when making concessions; agents with a high degree of cooperation are more willing to understand the other party's position and proactively seek solutions acceptable to both parties. The mediator agent, utilizing its knowledge reserves, provides professional legal analysis, mediation suggestions, and solution approaches for different points of contention.

[0086] In a preferred embodiment, the agent also has a reflection and strategy adaptation mechanism, which can dynamically adjust the strategy type or parameter intensity according to the dialogue status, the opponent's behavior and the progress of the issue, so that the agent's behavior is closer to the dynamic evolution characteristics in real negotiation scenarios.

[0087] Specifically, the method further includes: After each round of dialogue, reflect on and evaluate the state of the current round of dialogue; Based on the results of the aforementioned reflection and evaluation, the degree of self-assertion and cooperation of the party's intelligent agent are adjusted.

[0088] In the embodiments of this specification, reflecting on and evaluating the current dialogue state includes at least the following: (1) Has the current focus of the dispute changed, or has it been fully discussed? (2) Changes in the degree to which one's own core demands are met; (3) The opponent's behavioral tendencies (aggressive / cooperative / avoidant, etc.) and the trend of conflict intensity; (4) Consensus convergence trend and litigation escalation risk trend.

[0089] Based on the results of the reflection assessment, the behavioral strategies of the agent can be updated, such as adjusting the degree of self-assertion and cooperation, or switching the strategy type within a preset strategy set.

[0090] For example, if the current focus of the dispute changes or is not fully discussed, the level of cooperation of the party's intelligent agent should be appropriately increased, and the intensity of excessive self-assertion should be reduced; if the satisfaction level of one's own core demands is too low, the intensity of self-assertion of one's own party's intelligent agent should be appropriately increased, and the level of cooperation reduced, to ensure that its core interests are reasonably protected. In actual implementation, staff can pre-set the adjustment decision algorithm and adjust the level of self-assertion and cooperation of the party's intelligent agent based on the reflection and evaluation results, thereby ensuring a dynamic balance between the level of self-assertion and cooperation.

[0091] Through this reflective evaluation and strategy adjustment mechanism, the parties involved can dynamically optimize their behavioral strategies based on the actual development of the dialogue, making them more closely resemble the dynamic evolution characteristics of real negotiation behavior. This mechanism effectively solves the problems of goal drift and strategy mismatch that easily occur in long-term interactions with static strategy settings, improves the realism and effectiveness of mediation simulation, and provides important technical support for generating high-quality mediation process data.

[0092] Based on the same inventive concept, embodiments of this specification also provide an apparatus for constructing civil dispute mediation data, such as... Figure 5 As shown, it includes: The raw data acquisition and processing unit 501 is used to acquire raw data of civil dispute cases and perform structured extraction on the raw data of civil dispute cases to obtain structured case information. The agent creation unit 502 is used to create the party agent corresponding to each party and the mediator agent corresponding to the mediator in the case structured information; The dialogue simulation unit 503 is used to input the structured information of the case to each party's intelligent agent and the mediator's intelligent agent, so that the party's intelligent agent and the mediator's intelligent agent can conduct dialogues in multiple preset mediation stages based on the structured information of the case; The mediation stage scheduling unit 504 is used to obtain the execution data of the multi-round dialogue in the mediation stage after receiving the dialogue end message of each mediation stage, and calculate the stage index of the mediation stage based on the execution data of the multi-round dialogue; when there is a subsequent mediation stage in the mediation stage, the subsequent mediation stage of the mediation stage is scheduled according to the stage index and the stage scheduling rules. Mediation data storage unit 505 is used to store mediation data at each stage of mediation.

[0093] The beneficial effects obtained by the above-described device are the same as those obtained by the above-described method, and will not be described in detail in the embodiments of this specification.

[0094] like Figure 6 The diagram shown is a structural schematic of a computer device according to an embodiment of this specification. The methods described in this specification can be applied to the computer device of this embodiment.

[0095] Computer device 602 may include one or more processors 604, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads. Computer device 602 may also include any memory 606 for storing information of any kind, such as code, settings, data, etc. Non-limitingly, for example, memory 606 may include any type of RAM, any type of ROM, flash memory, hard disk, optical disk, etc. More generally, any storage resource can be used to store information using any technology.

[0096] Furthermore, any storage resource can provide volatile or non-volatile retention of information.

[0097] Furthermore, any storage resource can represent a fixed or removable component of the computer device 602. In one case, when the processor 604 executes associated instructions stored in any storage resource or combination of storage resources, the computer device 602 can perform any operation of the associated instructions. The computer device 602 also includes one or more drive mechanisms 608 for interacting with any storage resource, such as a hard disk drive system, an optical disk drive system, etc.

[0098] Computer device 602 may also include an input / output module 610 (I / O) for receiving various inputs (via input device 612) and providing various outputs (via output device 614). A specific output mechanism may include a presentation device 616 and an associated graphical user interface (GUI) 618. In other embodiments, the input / output module 610 (I / O), input device 612, and output device 614 may be omitted, and the device may function solely as a computer device within a network. Computer device 602 may also include one or more network interfaces 620 for exchanging data with other devices via one or more communication links 622. One or more communication buses 624 couple the components described above together.

[0099] Communication link 622 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. Communication link 622 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0100] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0101] This specification also provides computer-readable instructions, wherein when a processor executes the instructions, the program therein causes the processor to perform the above-described method.

[0102] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0103] It should also be understood that, in the embodiments of this specification, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the embodiments of this specification, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0104] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the embodiments in this specification.

[0105] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0106] In the embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0107] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments described in this specification, depending on actual needs.

[0108] Furthermore, the functional units in the various embodiments of this specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0109] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this specification, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this specification. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0110] This specification describes the principles and implementation methods of the embodiments using specific examples. The above descriptions of the embodiments are only for the purpose of helping to understand the methods and core ideas of the embodiments in this specification. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the embodiments in this specification. Therefore, the content of this specification should not be construed as a limitation on the embodiments in this specification.

Claims

1. A method for constructing civil dispute mediation data, characterized in that, The method includes: Obtain the original data of civil dispute cases, and perform structured extraction on the original data of civil dispute cases to obtain structured case information; Create a party agent corresponding to each party and a mediator agent corresponding to each mediator in the structured information of the case; The structured case information is input to the parties' intelligent agents and the mediator's intelligent agent, so that the parties' intelligent agents and the mediator's intelligent agent can conduct dialogues in multiple preset mediation stages based on the structured case information; After receiving the dialogue end message of each mediation stage, the execution data of the multi-round dialogue in that mediation stage is obtained, and the stage index of that mediation stage is calculated based on the execution data of the multi-round dialogue; when there is a subsequent mediation stage in that mediation stage, the subsequent mediation stage is scheduled according to the stage index and the stage scheduling rules. Store mediation data for each stage of the mediation process.

2. The method according to claim 1, characterized in that, Scheduling subsequent mediation stages based on the aforementioned stage indicators and stage scheduling rules further includes: Based on the stage indicators and the stage scheduling rules, determine whether to skip subsequent mediation stages of this mediation stage; or, The execution budget for subsequent mediation phases of the mediation phase is adjusted based on the phase indicators and the phase scheduling rules.

3. The method according to claim 2, characterized in that, The execution data includes: the satisfaction level of the party's intelligent agent and the mediator's intelligent agent, the degree of consensus on key issues in the structured information of the case, the litigation risk of the party's intelligent agent, and the degree of first coverage of the legal basis used by the mediator's intelligent agent with the legal basis in the structured information of the case. The phase indicators for this mediation stage, calculated based on data from multiple rounds of dialogue, further include: The satisfaction level, litigation risk, consensus level, and first coverage level mentioned in the multiple rounds of dialogue were normalized to obtain the satisfaction level index, litigation risk index, consensus level index, and first coverage level index for this mediation stage.

4. The method according to claim 3, characterized in that, Adjusting the execution budget for subsequent mediation phases based on the aforementioned phase indicators and phase scheduling rules further includes: The adjustment amount and direction of the execution budget for the subsequent mediation stage are determined based on the aforementioned satisfaction index, litigation risk index, consensus index, and first coverage index. The execution budget for the subsequent mediation phase is adjusted according to the stated adjustment amount and direction.

5. The method according to claim 3, characterized in that, After calculating the phase indicators for this mediation phase based on the execution data of multiple rounds of dialogue, the method further includes: The decision on whether to revert to enforcement at this mediation stage is based on the satisfaction index, litigation risk index, consensus index, and first coverage index mentioned in the multiple rounds of dialogue.

6. The method according to claim 1, characterized in that, The method further includes: After receiving the end message of the last mediation stage, the mediation success rate is calculated based on the acceptance of the final solution obtained after the end of the last mediation stage by each party's intelligent agent. The alignment degree of the dispute focus between the first set of dispute focus generated by the mediator intelligent agent after the end of each mediation stage and the second set of dispute focus in the case structure information is calculated. The second coverage degree of the legal basis set used by the mediator intelligent agent after the end of each mediation stage to the legal basis set in the case structure information is also calculated. If the mediation success rate does not meet the predetermined success rate requirement, the alignment of the dispute focus does not meet the predetermined alignment requirement, and / or the second coverage does not meet the predetermined coverage requirement, the parties' agents and the mediator's agent are controlled to re-engage in multiple preset mediation stages based on the case structured information.

7. The method according to claim 1, characterized in that, The multiple pre-set mediation stages include at least a solution generation stage and a bargaining stage; Scheduling subsequent mediation stages based on the aforementioned stage indicators and stage scheduling rules further includes: The bargaining phase is scheduled according to the phase indicators and phase scheduling rules of the scheme generation phase.

8. The method according to claim 1, characterized in that, Creating the party agent corresponding to each party and the mediator agent corresponding to the mediator in the structured case information further includes: Create multiple party intelligent agents for each party and multiple mediator intelligent agents for the mediator; The method further includes: Configure the behavioral strategies of each party's intelligent agent. The parameters of the behavioral strategies include the degree of self-assertion and the degree of cooperation. The behavioral strategies of different parties' intelligent agents are different. Configure the knowledge reserve strategy for each mediator agent, and the knowledge reserve strategies for different mediator agents are different; Multiple mediation teams are obtained by combining multiple party agents with different behavioral strategies and multiple mediator agents with different knowledge reserve strategies. The input of the structured case information to the parties' intelligent agents and the mediator's intelligent agent further includes: The structured case information is input into each mediation group so that the parties' agents and the mediator's agents in each mediation group can conduct dialogues in multiple preset mediation stages based on the structured case information.

9. The method according to claim 8, characterized in that, The method further includes: After each round of dialogue, reflect on and evaluate the state of the current round of dialogue; Based on the results of the aforementioned reflection and evaluation, the degree of self-assertion and cooperation of the party's intelligent agent are adjusted.

10. A device for constructing civil dispute mediation data, characterized in that, The device includes: The raw data acquisition and processing unit is used to acquire raw data of civil dispute cases and perform structured extraction on the raw data of civil dispute cases to obtain structured case information. The agent creation unit is used to create the party agent corresponding to each party in the case structured information and the mediator agent corresponding to the mediator. The dialogue simulation unit is used to input the structured information of the case to each party's intelligent agent and the mediator's intelligent agent, so that the party's intelligent agent and the mediator's intelligent agent can conduct dialogues in multiple preset mediation stages based on the structured information of the case; The mediation stage scheduling unit is used to obtain the execution data of the multi-round dialogue in each mediation stage after receiving the dialogue end message of each mediation stage, and calculate the stage index of the mediation stage based on the execution data of the multi-round dialogue; when there is a subsequent mediation stage in the mediation stage, the subsequent mediation stage of the mediation stage is scheduled according to the stage index and the stage scheduling rules. The mediation data storage unit is used to store mediation data at each stage of the mediation process.

Images