Contract evidence storage and trial assistance method and system based on artificial intelligence and blockchain

Abstract

The application provides a contract evidence storage and trial assistance method and system based on artificial intelligence and blockchain. The method comprises the following steps: storing a multi-modal contract on a blockchain and constructing a clause-level hash subchain, and deploying a performance condition as a smart contract; monitoring the performance state in real time, and performing abnormal storage and updating the subchain when an abnormality occurs; constructing a dispute graph based on the stored data; searching for missing evidence paths in combination with a complaint text, predicting the dispute focus; generating evidence reinforcement suggestions by associating with similar cases; and generating trial suggestions after the supplementary evidence is uploaded. Through the deep integration of artificial intelligence and blockchain technology, the application realizes the credibility of contract evidence storage, the automation of performance monitoring, and the intelligence of trial assistance, and can effectively solve the problems of easy tampering of contract evidence storage, lagging performance monitoring, and low trial efficiency.

Description

Technical Field

[0001] This application relates to the field of contract evidence preservation and judicial trial assistance technology, specifically to contract evidence preservation and trial assistance methods and systems based on artificial intelligence and blockchain. Background Technology

[0002] In current contract management and judicial practice, contract evidence preservation and dispute resolution face numerous challenges. On the one hand, traditional contract evidence preservation often relies on paper archives or centralized electronic storage, which suffers from issues such as susceptibility to tampering, loss, and difficulty in tracing. Once a contract dispute arises, the authenticity and legality of the evidence cannot be effectively guaranteed, increasing the difficulty of resolving the dispute. On the other hand, the lack of an effective real-time monitoring mechanism during contract performance means that abnormal performance often goes undetected and evidence cannot be secured in a timely manner, resulting in an incomplete chain of evidence when a dispute occurs. In judicial proceedings, judges need to spend a significant amount of time reviewing contract clauses, verifying evidence, and searching for similar cases, leading to low trial efficiency. Existing technologies, while some attempts to apply blockchain technology to contract evidence preservation, mostly remain at the level of overall contract preservation, failing to achieve refined management at the clause level and automated execution of performance conditions. The application of artificial intelligence technology in the judicial field is also largely concentrated on case retrieval or text parsing, without deep integration with blockchain evidence preservation technology, making it difficult to form a complete closed-loop process of "evidence preservation - monitoring - dispute resolution - trial assistance." Furthermore, existing trial support systems lack accurate identification of missing evidence and targeted suggestions for evidence reinforcement, failing to effectively assist the party presenting evidence in perfecting the chain of evidence, and also failing to provide judges with scientific and efficient trial references, thus failing to meet the current needs for intelligent and efficient contract management and judicial trials. Summary of the Invention

[0003] The purpose of this application is to provide a method and system for contract evidence storage and trial assistance based on artificial intelligence and blockchain, in order to overcome the shortcomings of existing technologies such as insecure contract evidence storage, lagging performance monitoring, and insufficient intelligence in trial assistance, so as to realize the credibility of contract evidence storage, automation of performance monitoring, accuracy of dispute prediction, and intelligence in trial assistance, thereby improving the efficiency of contract management and the quality of judicial trials, and reducing judicial costs.

[0004] This application also provides a method for contract notarization and trial assistance based on artificial intelligence and blockchain, including the following steps: Multimodal contracts are stored on the blockchain, a clause-level hash subchain is constructed, and the performance conditions are deployed as executable smart contracts. The system monitors the triggering status of performance conditions in real time. If an abnormality is detected, it performs abnormal evidence storage and updates the corresponding clause-level hash subchain on the blockchain. Construct a dispute graph based on the evidence data in the clause-level hash subchain; Obtain the text of the complaint, and search for missing evidence paths in the dispute graph based on the text of the complaint to obtain the preliminary results of the focus of the dispute; Based on the predicted points of contention, related cases are linked to generate suggestions for strengthening evidence. The party providing evidence adds supplementary evidence to the blockchain based on the evidence strengthening suggestions, and the system generates trial suggestions.

[0005] Optionally, in the contract notarization and trial assistance method based on artificial intelligence and blockchain described in this application, the step of notarizing multimodal contracts on the blockchain, constructing a clause-level hash subchain, and deploying the performance conditions as an executable smart contract includes: Acquire multimodal contract data, parse the multimodal contract data using a preset multimodal fusion model, and generate structured contract text; The structured contract text is divided into multiple semantic units using a pre-defined semantic segmentation algorithm, with each semantic unit corresponding to an independent contract clause; Each semantic unit independently calculates its hash value and stores it on the blockchain. Based on the hash value, on-chain timestamp and block height of each clause, a corresponding clause-level hash sub-chain is generated. Extract the set of performance condition parameters from each semantic unit, including triggering conditions and completion standard data; Triggering conditions include event triggering, time triggering, or state triggering; completion standard data includes time thresholds, numerical thresholds, and agreed behavior completion indicators. Each semantic unit's set of performance condition parameters is independently compiled into an executable smart contract instance; Deploy smart contract instances to the blockchain and establish an anchor mapping relationship between each smart contract instance and the hash value of the corresponding terms.

[0006] Optionally, in the contract notarization and trial assistance method based on artificial intelligence and blockchain described in this application, the real-time monitoring of the triggering status of performance conditions, and if an abnormal performance is detected, performing abnormal notarization and updating the corresponding clause-level hash subchain on the blockchain, includes: Obtain contract performance data, compare the contract performance data with the performance condition parameter set, and generate a performance status vector, including time deviation value, numerical deviation rate and status matching code; Determine whether the time deviation value exceeds the preset time deviation threshold, determine whether the numerical deviation rate exceeds the preset deviation rate threshold, and determine whether the status matching code meets the preset performance completion conditions. If any judgment result is yes, then it is determined to be an abnormal performance; Anomaly types are identified based on an anomaly classification model, which is trained on a pre-set historical default case database. Based on the type of exception, the corresponding exception evidence storage template is matched from the pre-configured evidence storage strategy library; Using the abnormal evidence template and based on the abnormality type, performance status vector, and deviation judgment criteria, an abnormal evidence data packet is generated and a hash value of the abnormal evidence is obtained by hash calculation. The abnormal evidence hash value is submitted to the blockchain for evidence storage, and the corresponding clause-level hash subchain is updated based on the abnormal evidence hash value.

[0007] Optionally, in the contract evidence storage and trial assistance method based on artificial intelligence and blockchain described in this application, the construction of a dispute graph based on the evidence storage data in the clause-level hash subchain includes: The evidence storage data in the clause-level hash subchain is transformed into on-chain evidence storage nodes, including contract clause nodes, performance data nodes, abnormal evidence storage nodes, and legal requirement nodes; Based on the on-chain evidence-stored nodes, a dispute graph is constructed using the semantic similarity and logical causal relationship between the associated original semantic content as edges.

[0008] Optionally, in the contract evidence storage and trial assistance method based on artificial intelligence and blockchain described in this application, the step of obtaining the complaint text and searching for missing evidence paths in the dispute graph based on the complaint text to obtain a preliminary judgment result of the focus of the dispute includes: Semantic analysis of the complaint text was performed to extract core elements, including the claims, factual assertions, evidence, and legal basis. The core elements are structured to generate standardized litigation element vectors; Extract the original evidence semantic information associated with the on-chain evidence storage nodes, and calculate the similarity between the semantic information and the standardized litigation element vector. Determine the corresponding target node based on the semantic similarity threshold. Starting from the target node, based on the semantic similarity edges and logical causal relationship edges between nodes in the dispute graph, a search for missing evidence paths is conducted to identify missing links that are related to the litigation claims and factual assertions but for which no corresponding on-chain evidence storage node has been found to support them. Cluster analysis was performed on the missing links to extract the core clauses, types of performance anomalies, and legal gaps corresponding to the missing links; The core clauses, types of performance anomalies, and gaps in legal requirements are input into a pre-set knowledge base of dispute points for semantic mapping and feature matching, generating a preliminary judgment result of the dispute points that includes the type of dispute point, related clauses, explanation of missing evidence, and the basis for the judgment.

[0009] Optionally, in the contract evidence preservation and trial assistance method based on artificial intelligence and blockchain described in this application, the step of generating evidence strengthening suggestions based on the predicted results of disputed issues and related similar cases includes: The predicted focus of the dispute is input into a pre-set multimodal contract dispute case database for matching and identification to obtain similar cases; Extract the key points of judgment, rules of evidence acceptance, and judgment results from similar cases, put them on the blockchain and transform them into judgment nodes, evidence rule nodes, and judgment nodes, and integrate them into the dispute graph, and establish a connection with the corresponding target nodes in the dispute graph. Based on the comparison results between the contract terms, performance data, and legal requirements associated with the evidence rule nodes and the target nodes, suggestions for strengthening evidence are generated.

[0010] Optionally, in the contract evidence storage and trial assistance method based on artificial intelligence and blockchain described in this application, the party presenting the evidence supplements evidence on the blockchain according to the evidence strengthening suggestions, and the system generates trial suggestions, including: The party providing evidence supplements the evidence based on the evidence strengthening suggestions, calculates the hash value, submits it to the clause-level hash subchain for on-chain storage, and generates a supplementary storage node; The disputed graph is updated based on the supplementary evidence storage nodes; Based on all node information in the updated dispute graph and the key points of judgment, rules of evidence acceptance and judgment results in similar cases, a trial analysis model is constructed using a preset logistic regression algorithm; Based on the trial analysis model, the focus of the current case is analyzed, the completeness of the evidence chain and the accuracy of the application of law are assessed, and trial recommendations are output. The trial recommendations include the tendency of the judgment on the points of contention, the conclusion on the admissibility of evidence, and suggestions on the application of legal provisions.

[0011] Secondly, this application provides a contract evidence storage and trial assistance system based on artificial intelligence and blockchain. The system includes a memory and a processor. The memory stores a program for a contract evidence storage and trial assistance method based on artificial intelligence and blockchain. When the program for this method is executed by the processor, it performs the following steps: Multimodal contracts are stored on the blockchain, a clause-level hash subchain is constructed, and the performance conditions are deployed as executable smart contracts. The system monitors the triggering status of performance conditions in real time. If an abnormality is detected, it performs abnormal evidence storage and updates the corresponding clause-level hash subchain on the blockchain. Construct a dispute graph based on the evidence data in the clause-level hash subchain; Obtain the text of the complaint, and search for missing evidence paths in the dispute graph based on the text of the complaint to obtain the preliminary results of the focus of the dispute; Based on the predicted points of contention, related cases are linked to generate suggestions for strengthening evidence. The party providing evidence adds supplementary evidence to the blockchain based on the evidence strengthening suggestions, and the system generates trial suggestions.

[0012] Optionally, in the contract notarization and trial assistance system based on artificial intelligence and blockchain described in this application, the step of notarizing multimodal contracts on the blockchain, constructing a clause-level hash subchain, and deploying the performance conditions as executable smart contracts includes: Acquire multimodal contract data, parse the multimodal contract data using a preset multimodal fusion model, and generate structured contract text; The structured contract text is divided into multiple semantic units using a pre-defined semantic segmentation algorithm, with each semantic unit corresponding to an independent contract clause; Each semantic unit independently calculates its hash value and stores it on the blockchain. Based on the hash value, on-chain timestamp and block height of each clause, a corresponding clause-level hash sub-chain is generated. Extract the set of performance condition parameters from each semantic unit, including triggering conditions and completion criteria; Triggering conditions include event triggering, time triggering, or state triggering; completion criteria include time thresholds, numerical thresholds, and agreed-upon behavior completion indicators. Each semantic unit's set of performance condition parameters is independently compiled into an executable smart contract instance; Deploy smart contract instances to the blockchain and establish an anchor mapping relationship between each smart contract instance and the hash value of the corresponding terms.

[0013] Optionally, in the contract notarization and trial assistance system based on artificial intelligence and blockchain described in this application, the real-time monitoring of the triggering status of performance conditions, and if an abnormal performance is detected, performing abnormal notarization and updating the corresponding clause-level hash subchain on the blockchain, includes: Obtain contract performance data, compare the contract performance data with the performance condition parameter set, and generate a performance status vector, including time deviation value, numerical deviation rate and status matching code; Determine whether the time deviation value exceeds the preset time deviation threshold, determine whether the numerical deviation rate exceeds the preset deviation rate threshold, and determine whether the status matching code meets the preset performance completion conditions. If any judgment result is yes, then it is determined to be an abnormal performance; Anomaly types are identified based on an anomaly classification model, which is trained on a pre-set historical default case database. Based on the type of exception, the corresponding exception evidence storage template is matched from the pre-configured evidence storage strategy library; Using the abnormal evidence template and based on the abnormality type, performance status vector, and deviation judgment criteria, an abnormal evidence data packet is generated and a hash value of the abnormal evidence is obtained by hash calculation. The abnormal evidence hash value is submitted to the blockchain for evidence storage, and the corresponding clause-level hash subchain is updated based on the abnormal evidence hash value.

[0014] As can be seen from the above, the contract evidence storage and trial assistance method and system based on artificial intelligence and blockchain provided in this application are designed to overcome the shortcomings of existing technologies, such as insecure contract evidence storage, lagging performance monitoring, and insufficient intelligent trial assistance. This enables the trustworthiness of contract evidence storage, automation of performance monitoring, accurate prediction of disputes, and intelligent trial assistance, thereby improving the efficiency of contract management and the quality of judicial trials, and reducing judicial costs.

[0015] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing embodiments of this application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating the contract notarization and trial assistance method based on artificial intelligence and blockchain provided in this application embodiment; Figure 2 A flowchart illustrating the construction of a clause-level hash subchain in the contract notarization and trial assistance method based on artificial intelligence and blockchain provided in this application embodiment; Figure 3 A flowchart illustrating the contract notarization and trial assistance method based on artificial intelligence and blockchain provided in this application embodiment; Figure 4 This is a schematic diagram of a contract storage and trial assistance system based on artificial intelligence and blockchain provided in an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0019] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0020] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a contract evidence storage and trial assistance method based on artificial intelligence and blockchain, as described in some embodiments of this application. This method is used in terminal devices, such as computers and mobile phones. The method includes the following steps: S11. Blockchain notarization of multimodal contracts, construction of clause-level hash subchain, and deployment of performance conditions as executable smart contracts; S12. Monitor the triggering status of performance conditions in real time. If a performance abnormality is detected, perform abnormal evidence storage and update the corresponding clause-level hash subchain on the blockchain. S13. Construct a dispute graph based on the evidence data in the clause-level hash subchain; S14. Obtain the text of the complaint and search for missing evidence paths in the dispute graph based on the text of the complaint to obtain the preliminary results of the focus of the dispute. S15. Based on the predicted results of the points of contention, associate similar cases and generate suggestions for strengthening evidence; S16. The party providing evidence adds supplementary evidence to the blockchain based on the evidence strengthening suggestions, and the system generates trial suggestions.

[0021] It should be noted that this application deeply integrates artificial intelligence and blockchain technology, employing a clause-level hash subchain to achieve refined evidence preservation of multimodal contracts. It combines smart contracts to automatically monitor performance conditions and secure abnormal evidence on the blockchain in real time. Relying on dispute graphs, evidence missing path searches, and similar case matching, it achieves accurate prediction of disputed issues and intelligent evidence reinforcement. Furthermore, it uses an LSTM time-series model to complete clause-level default risk warnings, ultimately forming a closed-loop process of "trustworthy evidence preservation - automatic monitoring - risk warning - dispute analysis - trial assistance." This not only solves the problems of easy tampering, incomplete evidence chains, and lagging performance supervision in traditional contract evidence preservation, but also significantly improves the efficiency of dispute resolution and the level of intelligent judicial adjudication, reduces judicial costs and subjective bias, and provides safe, efficient, and accurate technical support for full lifecycle contract management and judicial assistance.

[0022] Please refer to Figure 2 , Figure 2This is a flowchart illustrating the construction of a clause-level hash subchain in a contract notarization and trial assistance method based on artificial intelligence and blockchain, as described in some embodiments of this application. According to embodiments of the present invention, the step of blockchain notarization of multimodal contracts, construction of a clause-level hash subchain, and deployment of performance conditions as executable smart contracts includes: S21. Obtain multimodal contract data, parse the multimodal contract data using a preset multimodal fusion model, and generate structured contract text; S22. Use a preset semantic segmentation algorithm to split the structured contract text into multiple semantic units, each semantic unit corresponding to an independent contract clause; S23. Semantic units independently calculate hash values ​​and store them on the blockchain. Based on the hash value, blockchain timestamp, and block height of each clause, a corresponding clause-level hash sub-chain is generated. S24. Extract the performance condition parameter set from each semantic unit, including triggering conditions and completion standard data; S25. Triggering conditions include event triggering, time triggering, or state triggering. Completion standard data includes time threshold, numerical threshold, and agreed behavior completion identifier. S26. Compile the performance condition parameter set of each semantic unit independently into an executable smart contract instance; S27. Deploy smart contract instances to the blockchain and establish an anchor mapping relationship between each smart contract instance and the hash value of the corresponding terms.

[0023] It should be noted that, in this embodiment, the multimodal contract specifically includes contract-related data in various forms such as text, images, audio, video, and electronic signatures. The multimodal fusion model is specifically a multimodal fusion parsing model based on an improved Transformer architecture. This model includes a text parsing branch, an image recognition branch, and an audio transcription branch. The text parsing branch uses the BERT model to extract semantics from the contract text; the image recognition branch uses a CNN convolutional neural network to extract features from contract-related images and convert them into text descriptions; and the audio transcription branch uses an ASR speech recognition model to transcribe audio information into text. Then, an attention mechanism is used to fuse and align the features extracted from each branch, ultimately outputting a structured contract text. This application decomposes the structured contract text into semantic units and constructs a clause-level hash subchain, achieving refined evidence storage and traceability of contract clauses. It can accurately locate the evidence storage information of any clause, avoiding the problem of clause association confusion caused by overall evidence storage. Meanwhile, the performance condition parameter set was compiled into a smart contract and anchored to the clause hash value, realizing the automated execution and monitoring of performance conditions. This laid the foundation for subsequent performance anomaly monitoring and ensured the accurate correspondence between the smart contract and the corresponding contract clauses, thereby improving the accuracy and reliability of performance monitoring.

[0024] According to an embodiment of the present invention, the real-time monitoring of the triggering status of performance conditions, and if a performance anomaly is detected, performing anomaly notarization and updating the corresponding clause-level hash subchain on the blockchain, includes: Obtain contract performance data, compare the contract performance data with the performance condition parameter set, and generate a performance status vector, including time deviation value, numerical deviation rate and status matching code; Determine whether the time deviation value exceeds the preset time deviation threshold, determine whether the numerical deviation rate exceeds the preset deviation rate threshold, and determine whether the status matching code meets the preset performance completion conditions. If any judgment result is yes, then it is determined to be an abnormal performance; Anomaly types are identified based on an anomaly classification model, which is trained on a pre-set historical default case database. Based on the type of exception, the corresponding exception evidence storage template is matched from the pre-configured evidence storage strategy library; Using the abnormal evidence template and based on the abnormality type, performance status vector, and deviation judgment criteria, an abnormal evidence data packet is generated and a hash value of the abnormal evidence is obtained by hash calculation. The abnormal evidence hash value is submitted to the blockchain for evidence storage, and the corresponding clause-level hash subchain is updated based on the abnormal evidence hash value.

[0025] It should be noted that this application achieves quantitative assessment of performance status and accurate identification of anomalies by refining the monitoring, evidence preservation, and updating of clause-level hash subchains. Based on the anomaly classification model trained on the historical default case database and the pre-configured evidence preservation template, the application ensures the standardization, completeness, and immutability of the evidence preservation of performance anomalies. At the same time, by updating the corresponding clause-level hash subchain, the application achieves accurate association between abnormal evidence and contract clauses, effectively solving the problems of lagging performance monitoring, untimely fixation of abnormal evidence, and incomplete evidence chains in traditional methods. This provides reliable data support for subsequent dispute resolution and trial assistance.

[0026] According to an embodiment of the present invention, constructing a dispute graph based on evidence data in a clause-level hash subchain includes: The evidence storage data in the clause-level hash subchain is transformed into on-chain evidence storage nodes, including contract clause nodes, performance data nodes, abnormal evidence storage nodes, and legal requirement nodes; Based on the on-chain evidence-stored nodes, a dispute graph is constructed using the semantic similarity and logical causal relationship between the associated original semantic content as edges.

[0027] It should be noted that, in this embodiment, the semantic similarity edge is constructed as follows: for any two nodes, the system extracts their text feature vectors (such as the main text of the contract terms, the description of performance data, and the definition of legal requirements), and uses a pre-trained language model (such as BERT) to calculate the cosine similarity between the two. If the similarity exceeds a preset threshold (such as 0.7), a semantic similarity edge is established between the two nodes, and the similarity value is recorded as the edge weight. The logical causal relationship edge is constructed as follows: based on time sequence, legal rules, and domain knowledge base, logical causal relationship edges are established, including: time causal edge: if the timestamp of the performance data node is after the agreed time of the contract terms node and constitutes a breach of agreement, a "violation" relationship edge is established; legal reasoning edge: according to the preset legal rule base, if a certain performance fact node can prove or refute a certain legal requirement node, a "proof" or "refute" relationship edge is established; requirement association edge: according to the legal knowledge graph, if there is a legal logical association between two legal requirement nodes (such as "force majeure" and "waiver of liability for breach of contract"), an "association" relationship edge is established.

[0028] According to an embodiment of the present invention, the step of obtaining the complaint text and performing an evidence missing path search in the dispute graph based on the complaint text to obtain a preliminary judgment result of the focus of the dispute includes: Semantic analysis of the complaint text was performed to extract core elements, including the claims, factual assertions, evidence, and legal basis. The core elements are structured to generate standardized litigation element vectors; Extract the original evidence semantic information associated with the on-chain evidence storage nodes, and calculate the similarity between the semantic information and the standardized litigation element vector. Determine the corresponding target node based on the semantic similarity threshold. Starting from the target node, based on the semantic similarity edges and logical causal relationship edges between nodes in the dispute graph, a search for missing evidence paths is conducted to identify missing links that are related to the litigation claims and factual assertions but for which no corresponding on-chain evidence storage node has been found to support them. Cluster analysis was performed on the missing links to extract the core clauses, types of performance anomalies, and legal gaps corresponding to the missing links; The core clauses, types of performance anomalies, and gaps in legal requirements are input into a pre-set knowledge base of dispute points for semantic mapping and feature matching, generating a preliminary judgment result of the dispute points that includes the type of dispute point, related clauses, explanation of missing evidence, and the basis for the judgment.

[0029] It should be noted that this application locates evidence gaps through evidence missing path search, and combines cluster analysis and a knowledge base of disputed points to achieve accurate prediction of disputed points, clarify the core of the case dispute, provide a clear direction for evidence strengthening and trial analysis, effectively avoid trial delays caused by evidence omissions, and improve the pertinence and efficiency of trial assistance. In the embodiments of this application, the process of cluster analysis and related information extraction for missing links is as follows: First, all missing evidence links found are collected, and the missing links are grouped using the contract terms, performance anomaly characteristics, and legal requirement types associated with the links as clustering dimensions, and the K-means clustering algorithm is used to group the missing links, and missing links with similar characteristics are grouped into the same cluster; Second, features are extracted from each cluster, and the contract terms commonly associated with the missing links within the cluster are selected as core terms, and the performance anomaly manifestations that frequently occur in the links within the cluster are extracted to determine the performance anomaly type; Finally, the legal requirement nodes associated with the clusters are compared with the legal requirements required for complete litigation, and the legal requirement content that is not supported by the evidence nodes is identified, which is the legal requirement gap.

[0030] According to an embodiment of the present invention, the step of generating evidence strengthening suggestions based on the predicted results of the disputed issues and related similar cases includes: The predicted focus of the dispute is input into a pre-set multimodal contract dispute case database for matching and identification to obtain similar cases; Extract the key points of judgment, rules of evidence acceptance, and judgment results from similar cases, put them on the blockchain and transform them into judgment nodes, evidence rule nodes, and judgment nodes, and integrate them into the dispute graph, and establish a connection with the corresponding target nodes in the dispute graph. Based on the comparison results between the contract terms, performance data, and legal requirements associated with the evidence rule nodes and the target nodes, suggestions for strengthening evidence are generated.

[0031] It should be noted that this application transforms the key points of similar cases' judgments, rules of evidence acceptance, and judgment results into nodes integrated into the dispute graph, achieving a deep integration of evidence data and similar case experience. This makes the suggestions for evidence strengthening more aligned with the actual needs of the case. Based on the comparison between evidence rule nodes and target nodes, suggestions are generated, which can accurately guide the party presenting evidence to supplement missing evidence, help improve the chain of evidence, lay the foundation for subsequent trial work, reduce invalid evidence, and improve the efficiency of evidence presentation.

[0032] According to an embodiment of the present invention, the party providing evidence supplements evidence onto the blockchain based on the evidence strengthening suggestion, and the system generates a trial suggestion, including: The party providing evidence supplements the evidence based on the evidence strengthening suggestions, calculates the hash value, submits it to the clause-level hash subchain for on-chain storage, and generates a supplementary storage node; The disputed graph is updated based on the supplementary evidence storage nodes; Based on all node information in the updated dispute graph and the key points of judgment, rules of evidence acceptance and judgment results in similar cases, a trial analysis model is constructed using a preset logistic regression algorithm; Based on the trial analysis model, the focus of the current case is analyzed, the completeness of the evidence chain and the accuracy of the application of law are assessed, and trial recommendations are output. The trial recommendations include the tendency of the judgment on the points of contention, the conclusion on the admissibility of evidence, and suggestions on the application of legal provisions.

[0033] It should be noted that supplementing evidence on the blockchain and updating the dispute graph ensures the integrity and immutability of the evidence chain, providing comprehensive and credible evidentiary support for the trial. Utilizing logistic regression algorithms combined with dispute graph information and similar case experience to construct a trial analysis model enables the intelligent generation of trial recommendations, reducing judges' subjective judgments and enhancing the objectivity and scientific rigor of these recommendations. The trial recommendations cover the tendencies of judgments on the points of contention, conclusions on evidence acceptance, and suggestions on the application of legal provisions, providing judges with comprehensive trial references, effectively shortening the trial cycle, and improving trial efficiency and quality. Specifically, the updated dispute graph includes node information for the current case's evidence storage node, supplementary evidence storage nodes, judgment nodes from similar cases, evidence rule nodes, and judgment nodes. The system structurally encodes all node information in the updated dispute graph (including the shortest path distance between nodes, the number of paths, the cumulative edge weights, and isograph structure features) with the key judgment points, evidence acceptance rules, and judgment results in similar cases. It extracts Boolean feature vectors covering legal elements, and constructs multidimensional feature vectors based on the structural similarity and node feature similarity between the case graph and the similar case graph calculated using a graph embedding algorithm. These multidimensional feature vectors are then used as training data to obtain a trial analysis model through supervised learning using logistic regression. During the analysis, this model is used to quantitatively assess the current case: it outputs a judgment tendency prediction based on the matching degree between the disputed focus features and the features of similar cases; it outputs an evidence chain integrity score based on the Boolean vectors of legal elements and graph structure association features; and it outputs a confidence level of the accuracy of legal application based on the comparison results between the Boolean vectors of legal elements and the legal provisions accepted in similar cases. Finally, based on the probability distribution, scoring results, and preset decision thresholds, it automatically generates trial recommendations including the judgment tendency of the disputed focus, the evidence chain integrity evaluation, the evidence acceptance conclusion, and recommended applicable legal provisions.

[0034] According to an embodiment of the present invention, it further includes: Extract historical performance data of the corresponding contract terms from the clause-level hash subchain, including historical performance time, performance value, performance status data and historical default record data; Historical performance data is preprocessed to remove outliers and standardize the data to generate a time-series performance dataset. A time-series prediction model is generated by training a pre-defined long short-term memory network model based on a time-series performance dataset. Obtain recent timeline performance data for the corresponding contract terms, including recent performance time, recent performance value, recent performance status data, and recent breach-related records. The recent time-series performance data is input into the time-series prediction model for processing to obtain the predicted probability of default for the corresponding terms within a future preset time period; Determine whether the predicted default probability is greater than a preset default probability threshold; If so, the system will automatically generate a clause-level risk warning report and send it to the contract performer. The risk warning report will be hashed to obtain the warning hash value, submitted to the blockchain for notarization, and the clause-level hash subchain of the corresponding clause will be updated.

[0035] It should be noted that historical breach records refer to objective records of breaches that occurred during the past performance of the corresponding contract terms. Specifically, they include: the time of the breach, the type of breach (such as time breach, numerical breach, status breach), the degree of breach (such as minor breach, serious breach), the rectification status of the breach, and the outcome of the breach handling (such as the amount of compensation and the rectification period), and all other records related to past breaches.

[0036] This invention also discloses a contract evidence storage and trial assistance system 4 based on artificial intelligence and blockchain, including a memory 41 and a processor 42. The memory stores a method program for contract evidence storage and trial assistance based on artificial intelligence and blockchain. When the processor executes the method program for contract evidence storage and trial assistance based on artificial intelligence and blockchain, it performs the following steps: Multimodal contracts are stored on the blockchain, a clause-level hash subchain is constructed, and the performance conditions are deployed as executable smart contracts. The system monitors the triggering status of performance conditions in real time. If an abnormality is detected, it performs abnormal evidence storage and updates the corresponding clause-level hash subchain on the blockchain. Construct a dispute graph based on the evidence data in the clause-level hash subchain; Obtain the text of the complaint, and search for missing evidence paths in the dispute graph based on the text of the complaint to obtain the preliminary results of the focus of the dispute; Based on the predicted points of contention, related cases are linked to generate suggestions for strengthening evidence. The party providing evidence adds supplementary evidence to the blockchain based on the evidence strengthening suggestions, and the system generates trial suggestions.

[0037] It should be noted that this application deeply integrates artificial intelligence and blockchain technology, employing a clause-level hash subchain to achieve refined evidence preservation of multimodal contracts. It combines smart contracts to automatically monitor performance conditions and secure abnormal evidence on the blockchain in real time. Relying on dispute graphs, evidence missing path searches, and similar case matching, it achieves accurate prediction of disputed issues and intelligent evidence reinforcement. Furthermore, it uses an LSTM time-series model to complete clause-level default risk warnings, ultimately forming a closed-loop process of "trustworthy evidence preservation - automatic monitoring - risk warning - dispute analysis - trial assistance." This not only solves the problems of easy tampering, incomplete evidence chains, and lagging performance supervision in traditional contract evidence preservation, but also significantly improves the efficiency of dispute resolution and the level of intelligent judicial adjudication, reduces judicial costs and subjective bias, and provides safe, efficient, and accurate technical support for full lifecycle contract management and judicial assistance.

[0038] According to an embodiment of the present invention, the step of performing blockchain notarization on multimodal contracts, constructing a clause-level hash subchain, and deploying the performance conditions as an executable smart contract includes: Acquire multimodal contract data, parse the multimodal contract data using a preset multimodal fusion model, and generate structured contract text; The structured contract text is divided into multiple semantic units using a pre-defined semantic segmentation algorithm, with each semantic unit corresponding to an independent contract clause; Semantic units independently calculate hash values ​​and store them on the blockchain. Based on the hash value, on-chain timestamp, and block height of each clause, a corresponding clause-level hash subchain is generated. Extract the set of performance condition parameters from each semantic unit, including triggering conditions and completion standard data; Triggering conditions include event triggering, time triggering, or state triggering; completion standard data includes time thresholds, numerical thresholds, and agreed behavior completion indicators. Each semantic unit's set of performance condition parameters is independently compiled into an executable smart contract instance; Deploy smart contract instances to the blockchain and establish an anchor mapping relationship between each smart contract instance and the hash value of the corresponding terms.

[0039] It should be noted that, in this embodiment, the multimodal contract specifically includes contract-related data in various forms such as text, images, audio, video, and electronic signatures. The multimodal fusion model is specifically a multimodal fusion parsing model based on an improved Transformer architecture. This model includes a text parsing branch, an image recognition branch, and an audio transcription branch. The text parsing branch uses the BERT model to extract semantics from the contract text; the image recognition branch uses a CNN convolutional neural network to extract features from contract-related images and convert them into text descriptions; and the audio transcription branch uses an ASR speech recognition model to transcribe audio information into text. Then, an attention mechanism is used to fuse and align the features extracted from each branch, ultimately outputting a structured contract text. This application decomposes the structured contract text into semantic units and constructs a clause-level hash subchain, achieving refined evidence storage and traceability of contract clauses. It can accurately locate the evidence storage information of any clause, avoiding the problem of clause association confusion caused by overall evidence storage. Meanwhile, the performance condition parameter set was compiled into a smart contract and anchored to the clause hash value, realizing the automated execution and monitoring of performance conditions. This laid the foundation for subsequent performance anomaly monitoring and ensured the accurate correspondence between the smart contract and the corresponding contract clauses, thereby improving the accuracy and reliability of performance monitoring.

[0040] According to an embodiment of the present invention, the real-time monitoring of the triggering status of performance conditions, and if a performance anomaly is detected, performing anomaly notarization and updating the corresponding clause-level hash subchain on the blockchain, includes: Obtain contract performance data, compare the contract performance data with the performance condition parameter set, and generate a performance status vector, including time deviation value, numerical deviation rate and status matching code; Determine whether the time deviation value exceeds the preset time deviation threshold, determine whether the numerical deviation rate exceeds the preset deviation rate threshold, and determine whether the status matching code meets the preset performance completion conditions. If any judgment result is yes, then it is determined to be an abnormal performance; Anomaly types are identified based on an anomaly classification model, which is trained on a pre-set historical default case database. Based on the type of exception, the corresponding exception evidence storage template is matched from the pre-configured evidence storage strategy library; Using the abnormal evidence template and based on the abnormality type, performance status vector, and deviation judgment criteria, an abnormal evidence data packet is generated and a hash value of the abnormal evidence is obtained by hash calculation. The abnormal evidence hash value is submitted to the blockchain for evidence storage, and the corresponding clause-level hash subchain is updated based on the abnormal evidence hash value.

[0041] It should be noted that this application achieves quantitative assessment of performance status and accurate identification of anomalies by refining the monitoring, evidence preservation, and clause-level hash subchain updates. The anomaly classification model trained on a historical default case database and the pre-configured evidence preservation template ensure the standardization, completeness, and immutability of the evidence preservation of performance anomalies. At the same time, by updating the corresponding clause-level hash subchain, the application achieves accurate association between abnormal evidence and contract clauses, effectively solving the problems of lagging performance monitoring, untimely fixation of abnormal evidence, and incomplete evidence chains in traditional methods. This provides reliable data support for subsequent dispute resolution and trial assistance.

[0042] According to an embodiment of the present invention, constructing a dispute graph based on evidence data in a clause-level hash subchain includes: The evidence storage data in the clause-level hash subchain is transformed into on-chain evidence storage nodes, including contract clause nodes, performance data nodes, abnormal evidence storage nodes, and legal requirement nodes; Based on the on-chain evidence-stored nodes, a dispute graph is constructed using the semantic similarity and logical causal relationship between the associated original semantic content as edges.

[0043] It should be noted that, in this embodiment, the semantic similarity edge is constructed as follows: for any two nodes, the system extracts their text feature vectors (such as the main text of the contract terms, the description of performance data, and the definition of legal requirements), and uses a pre-trained language model (such as BERT) to calculate the cosine similarity between the two. If the similarity exceeds a preset threshold (such as 0.7), a semantic similarity edge is established between the two nodes, and the similarity value is recorded as the edge weight. The logical causal relationship edge is constructed as follows: based on time sequence, legal rules, and domain knowledge base, logical causal relationship edges are established, including: time causal edge: if the timestamp of the performance data node is after the agreed time of the contract terms node and constitutes a breach of agreement, a "violation" relationship edge is established; legal reasoning edge: according to the preset legal rule base, if a certain performance fact node can prove or refute a certain legal requirement node, a "proof" or "refute" relationship edge is established; requirement association edge: according to the legal knowledge graph, if there is a legal logical association between two legal requirement nodes (such as "force majeure" and "waiver of liability for breach of contract"), an "association" relationship edge is established.

[0044] According to an embodiment of the present invention, the step of obtaining the complaint text and performing an evidence missing path search in the dispute graph based on the complaint text to obtain a preliminary judgment result of the focus of the dispute includes: Semantic analysis of the complaint text was performed to extract core elements, including the claims, factual assertions, evidence, and legal basis. The core elements are structured to generate standardized litigation element vectors; Extract the original evidence semantic information associated with the on-chain evidence storage nodes, and calculate the similarity between the semantic information and the standardized litigation element vector. Determine the corresponding target node based on the semantic similarity threshold. Starting from the target node, based on the semantic similarity edges and logical causal relationship edges between nodes in the dispute graph, a search for missing evidence paths is conducted to identify missing links that are related to the litigation claims and factual assertions but for which no corresponding on-chain evidence storage node has been found to support them. Cluster analysis was performed on the missing links to extract the core clauses, types of performance anomalies, and legal gaps corresponding to the missing links; The core clauses, types of performance anomalies, and gaps in legal requirements are input into a pre-set knowledge base of dispute points for semantic mapping and feature matching, generating a preliminary judgment result of the dispute points that includes the type of dispute point, related clauses, explanation of missing evidence, and the basis for the judgment.

[0045] It should be noted that this application locates evidence gaps through evidence missing path search, and combines cluster analysis and a knowledge base of disputed points to achieve accurate prediction of disputed points, clarify the core of the case dispute, provide a clear direction for evidence strengthening and trial analysis, effectively avoid trial delays caused by evidence omissions, and improve the pertinence and efficiency of trial assistance. In the embodiments of this application, the process of cluster analysis and related information extraction for missing links is as follows: First, all searched evidence missing links are collected, and the missing links are grouped using the contract terms, performance anomaly characteristics, and legal requirement types associated with the links as clustering dimensions, and the K-means clustering algorithm is used to group the missing links, and missing links with similar characteristics are grouped into the same cluster; Second, features are extracted from each cluster, and the contract terms commonly associated with the missing links within the cluster are selected as core terms, and the performance anomaly manifestations that frequently occur in the links within the cluster are extracted to determine the performance anomaly type; Finally, the legal requirement nodes associated with the clusters are compared with the legal requirements required for complete litigation, and the legal requirement content that is not supported by the evidence nodes is identified, which is the legal requirement gap.

[0046] According to an embodiment of the present invention, the step of generating evidence strengthening suggestions based on the predicted results of the disputed issues and related similar cases includes: The predicted focus of the dispute is input into a pre-set multimodal contract dispute case database for matching and identification to obtain similar cases; Extract the key points of judgment, rules of evidence acceptance, and judgment results from similar cases, put them on the blockchain and transform them into judgment nodes, evidence rule nodes, and judgment nodes, and integrate them into the dispute graph, and establish a connection with the corresponding target nodes in the dispute graph. Based on the comparison results between the contract terms, performance data, and legal requirements associated with the evidence rule nodes and the target nodes, suggestions for strengthening evidence are generated.

[0047] It should be noted that this application transforms the key points of similar cases' judgments, rules of evidence acceptance, and judgment results into nodes integrated into the dispute graph, achieving a deep integration of evidence data and similar case experience. This makes the suggestions for evidence strengthening more aligned with the actual needs of the case. Based on the comparison between evidence rule nodes and target nodes, suggestions are generated, which can accurately guide the party presenting evidence to supplement missing evidence, help improve the chain of evidence, lay the foundation for subsequent trial work, reduce invalid evidence, and improve the efficiency of evidence presentation.

[0048] According to an embodiment of the present invention, the party providing evidence supplements evidence onto the blockchain based on the evidence strengthening suggestion, and the system generates a trial suggestion, including: The party providing evidence supplements the evidence based on the evidence strengthening suggestions, calculates the hash value, submits it to the clause-level hash subchain for on-chain storage, and generates a supplementary storage node; The disputed graph is updated based on the supplementary evidence storage nodes; Based on all node information in the updated dispute graph and the key points of judgment, rules of evidence acceptance and judgment results in similar cases, a trial analysis model is constructed using a preset logistic regression algorithm; Based on the trial analysis model, the focus of the current case is analyzed, the completeness of the evidence chain and the accuracy of the application of law are assessed, and trial recommendations are output. The trial recommendations include the tendency of the judgment on the points of contention, the conclusion on the admissibility of evidence, and suggestions on the application of legal provisions.

[0049] It should be noted that supplementing evidence on the blockchain and updating the dispute graph ensures the integrity and immutability of the evidence chain, providing comprehensive and credible evidentiary support for the trial. Utilizing logistic regression algorithms combined with dispute graph information and similar case experience to construct a trial analysis model enables the intelligent generation of trial recommendations, reducing judges' subjective judgments and enhancing the objectivity and scientific rigor of these recommendations. The trial recommendations cover the tendencies of judgments on the points of contention, conclusions on evidence acceptance, and suggestions on the application of legal provisions, providing judges with comprehensive trial references, effectively shortening the trial cycle, and improving trial efficiency and quality. Specifically, the updated dispute graph includes node information for the current case's evidence storage node, supplementary evidence storage nodes, judgment nodes from similar cases, evidence rule nodes, and judgment nodes. The system structurally encodes all node information in the updated dispute graph (including the shortest path distance between nodes, the number of paths, the cumulative edge weights, and isograph structure features) with the key judgment points, evidence acceptance rules, and judgment results in similar cases. It extracts Boolean feature vectors covering legal elements, and constructs multidimensional feature vectors based on the structural similarity and node feature similarity between the case graph and the similar case graph calculated using a graph embedding algorithm. These multidimensional feature vectors are then used as training data to obtain a trial analysis model through supervised learning using logistic regression. During the analysis, this model is used to quantitatively assess the current case: it outputs a judgment tendency prediction based on the matching degree between the disputed focus features and the features of similar cases; it outputs an evidence chain integrity score based on the Boolean vectors of legal elements and graph structure association features; and it outputs a confidence level of the accuracy of legal application based on the comparison results between the Boolean vectors of legal elements and the legal provisions accepted in similar cases. Finally, based on the probability distribution, scoring results, and preset decision thresholds, it automatically generates trial recommendations including the judgment tendency of the disputed focus, the evidence chain integrity evaluation, the evidence acceptance conclusion, and recommended applicable legal provisions.

[0050] According to an embodiment of the present invention, it further includes: Extract historical performance data of the corresponding contract terms from the clause-level hash subchain, including historical performance time, performance value, performance status data and historical default record data; Historical performance data is preprocessed to remove outliers and standardize the data to generate a time-series performance dataset. A time-series prediction model is generated by training a pre-defined long short-term memory network model based on a time-series performance dataset. Obtain recent timeline performance data for the corresponding contract terms, including recent performance time, recent performance value, recent performance status data, and recent breach-related records. The recent time-series performance data is input into the time-series prediction model for processing to obtain the predicted probability of default for the corresponding terms within a future preset time period; Determine whether the predicted default probability is greater than a preset default probability threshold; If so, the system will automatically generate a clause-level risk warning report and send it to the contract performer. The risk warning report will be hashed to obtain the warning hash value, submitted to the blockchain for notarization, and the clause-level hash subchain of the corresponding clause will be updated.

[0051] It should be noted that historical breach records refer to objective records of breaches that occurred during the past performance of the corresponding contract terms. Specifically, they include: the time of the breach, the type of breach (such as time breach, numerical breach, status breach), the degree of breach (such as minor breach, serious breach), the rectification status of the breach, and the outcome of the breach handling (such as the amount of compensation and the rectification period), and all other records related to past breaches.

[0052] The contract evidence storage and trial assistance method and system based on artificial intelligence and blockchain disclosed in this invention aims to overcome the defects of existing technologies, such as insecure contract evidence storage, lagging performance monitoring, and insufficient intelligent trial assistance. It realizes the credibility of contract evidence storage, automation of performance monitoring, accuracy of dispute prediction, and intelligent trial assistance, thereby improving the efficiency of contract management and the quality of judicial trials, and reducing judicial costs.

[0053] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0054] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0055] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

Claims

1. A method for contract notarization and trial assistance based on artificial intelligence and blockchain, characterized in that: Includes the following steps: Multimodal contracts are stored on the blockchain, a clause-level hash subchain is constructed, and the performance conditions are deployed as executable smart contracts. The system monitors the triggering status of performance conditions in real time. If an abnormality is detected, it performs abnormal evidence storage and updates the corresponding clause-level hash subchain on the blockchain. Construct a dispute graph based on the evidence data in the clause-level hash subchain; Obtain the text of the complaint, and search for missing evidence paths in the dispute graph based on the text of the complaint to obtain the preliminary results of the focus of the dispute; Based on the predicted points of contention, related cases are linked to generate suggestions for strengthening evidence. The party providing evidence adds supplementary evidence to the blockchain based on the evidence strengthening suggestions, and the system generates trial suggestions.

2. The method for contract storage and trial assistance based on artificial intelligence and blockchain according to claim 1, characterized in that, The process of storing multimodal contracts on the blockchain, constructing a clause-level hash subchain, and deploying performance conditions as executable smart contracts includes: Acquire multimodal contract data, parse the multimodal contract data using a preset multimodal fusion model, and generate structured contract text; The structured contract text is divided into multiple semantic units using a pre-defined semantic segmentation algorithm, with each semantic unit corresponding to an independent contract clause; Each semantic unit independently calculates its hash value and stores it on the blockchain. Based on the hash value, on-chain timestamp and block height of each clause, a corresponding clause-level hash sub-chain is generated. Extract the set of performance condition parameters from each semantic unit, including triggering conditions and completion standard data; Triggering conditions include event triggering, time triggering, or state triggering; completion standard data includes time thresholds, numerical thresholds, and agreed behavior completion indicators. Each semantic unit's set of performance condition parameters is independently compiled into an executable smart contract instance; Deploy smart contract instances to the blockchain and establish an anchor mapping relationship between each smart contract instance and the hash value of the corresponding terms.

3. The method for contract storage and trial assistance based on artificial intelligence and blockchain according to claim 2, characterized in that, The real-time monitoring of the triggering status of performance conditions, if a performance anomaly is detected, involves anomaly notarization and updating the corresponding clause-level hash subchain on the blockchain, including: Obtain contract performance data, compare the contract performance data with the performance condition parameter set, and generate a performance status vector, including time deviation value, numerical deviation rate and status matching code; Determine whether the time deviation value exceeds the preset time deviation threshold, determine whether the numerical deviation rate exceeds the preset deviation rate threshold, and determine whether the status matching code meets the preset performance completion conditions. If any judgment result is yes, then it is determined to be an abnormal performance; Anomaly types are identified based on an anomaly classification model, which is trained on a pre-set historical default case database. Based on the type of exception, the corresponding exception evidence storage template is matched from the pre-configured evidence storage strategy library; Using the abnormal evidence template and based on the abnormality type, performance status vector, and deviation judgment criteria, an abnormal evidence data packet is generated and a hash value of the abnormal evidence is obtained by hash calculation. The abnormal evidence hash value is submitted to the blockchain for evidence storage, and the corresponding clause-level hash subchain is updated based on the abnormal evidence hash value.

4. The method for contract storage and trial assistance based on artificial intelligence and blockchain according to claim 3, characterized in that, The construction of the dispute graph based on the evidence storage data in the clause-level hash subchain includes: The evidence storage data in the clause-level hash subchain is transformed into on-chain evidence storage nodes, including contract clause nodes, performance data nodes, abnormal evidence storage nodes, and legal requirement nodes; Based on the on-chain evidence-stored nodes, a dispute graph is constructed using the semantic similarity and logical causal relationship between the associated original semantic content as edges.

5. The method for contract storage and trial assistance based on artificial intelligence and blockchain according to claim 4, characterized in that, The process of obtaining the complaint text and searching for missing evidence paths in the dispute graph based on the complaint text to obtain a preliminary judgment of the focus of the dispute includes: Semantic analysis of the complaint text was performed to extract core elements, including claims, factual assertions, evidence, and legal basis. The core elements are structured to generate standardized litigation element vectors; Extract the original evidence storage semantic information associated with the on-chain evidence storage nodes, and calculate the similarity between the semantic information and the semantic content of the standardized litigation element vector. Determine the corresponding target node based on the semantic similarity threshold. Starting from the target node, based on the semantic similarity edges and logical causal relationship edges between nodes in the dispute graph, a search for missing evidence paths is conducted to identify missing links that are related to the litigation claims and factual assertions but for which no corresponding on-chain evidence storage node has been found to support them. Cluster analysis was performed on the missing links to extract the core clauses, types of performance anomalies, and legal gaps corresponding to the missing links; The core clauses, types of performance anomalies, and gaps in legal requirements are input into a pre-set knowledge base of dispute points for semantic mapping and feature matching, generating a preliminary judgment result of the dispute points that includes the type of dispute point, related clauses, explanation of missing evidence, and the basis for the judgment.

6. The method for contract storage and trial assistance based on artificial intelligence and blockchain according to claim 5, characterized in that, Based on the predicted results of the points of contention and related similar cases, suggestions for strengthening evidence are generated, including: The predicted focus of the dispute is input into a pre-set multimodal contract dispute case database for matching and identification to obtain similar cases; Extract the key points of judgment, rules of evidence acceptance, and judgment results from similar cases, put them on the blockchain and transform them into judgment nodes, evidence rule nodes, and judgment nodes, and integrate them into the dispute graph, and establish a connection with the corresponding target nodes in the dispute graph. Based on the comparison results between the contract terms, performance data, and legal requirements associated with the evidence rule nodes and the target nodes, suggestions for strengthening evidence are generated.

7. The method for contract storage and trial assistance based on artificial intelligence and blockchain according to claim 6, characterized in that, The party providing evidence supplements the evidence on the blockchain based on the evidence strengthening suggestions, and the system generates trial suggestions, including: The party providing evidence supplements the evidence based on the evidence strengthening suggestions, calculates the hash value, submits it to the clause-level hash subchain for on-chain storage, and generates a supplementary storage node; The disputed graph is updated based on the supplementary evidence storage nodes; Based on all node information in the updated dispute graph and the key points of judgment, rules of evidence acceptance and judgment results in similar cases, a trial analysis model is constructed using a preset logistic regression algorithm; Based on the trial analysis model, the focus of the current case is analyzed, the completeness of the evidence chain and the accuracy of the application of law are assessed, and trial recommendations are output. The trial recommendations include the tendency of the judgment on the points of contention, the conclusion on the admissibility of evidence, and suggestions on the application of legal provisions.

8. A contract storage and trial assistance system based on artificial intelligence and blockchain, characterized in that: The system includes a memory and a processor. The memory stores a program for a contract notarization and trial assistance method based on artificial intelligence and blockchain. When the program for this method is executed by the processor, it performs the following steps: Multimodal contracts are stored on the blockchain, a clause-level hash subchain is constructed, and the performance conditions are deployed as executable smart contracts. The system monitors the triggering status of performance conditions in real time. If an abnormality is detected, it performs abnormal evidence storage and updates the corresponding clause-level hash subchain on the blockchain. Construct a dispute graph based on the evidence data in the clause-level hash subchain; Obtain the text of the complaint, and search for missing evidence paths in the dispute graph based on the text of the complaint to obtain the preliminary results of the focus of the dispute; Based on the predicted points of contention, related cases are linked to generate suggestions for strengthening evidence. The party providing evidence adds supplementary evidence to the blockchain based on the evidence strengthening suggestions, and the system generates trial suggestions.

9. The contract storage and trial assistance system based on artificial intelligence and blockchain according to claim 8, characterized in that, The process of storing multimodal contracts on the blockchain, constructing a clause-level hash subchain, and deploying performance conditions as executable smart contracts includes: Acquire multimodal contract data, parse the multimodal contract data using a preset multimodal fusion model, and generate structured contract text; The structured contract text is divided into multiple semantic units using a pre-defined semantic segmentation algorithm, with each semantic unit corresponding to an independent contract clause; Each semantic unit independently calculates its hash value and stores it on the blockchain. Based on the hash value, on-chain timestamp and block height of each clause, a corresponding clause-level hash sub-chain is generated. Extract the set of performance condition parameters from each semantic unit, including triggering conditions and completion standard data; Triggering conditions include event triggering, time triggering, or state triggering; completion standard data includes time thresholds, numerical thresholds, and agreed behavior completion indicators. Each semantic unit's set of performance condition parameters is independently compiled into an executable smart contract instance; Deploy smart contract instances to the blockchain and establish an anchor mapping relationship between each smart contract instance and the hash value of the corresponding terms.

10. The contract storage and trial assistance system based on artificial intelligence and blockchain according to claim 9, characterized in that, The real-time monitoring of the triggering status of performance conditions, if a performance anomaly is detected, involves anomaly notarization and updating the corresponding clause-level hash subchain on the blockchain, including: Obtain contract performance data, compare the contract performance data with the performance condition parameter set, and generate a performance status vector, including time deviation value, numerical deviation rate and status matching code; Determine whether the time deviation value exceeds the preset time deviation threshold, determine whether the numerical deviation rate exceeds the preset deviation rate threshold, and determine whether the status matching code meets the preset performance completion conditions. If any judgment result is yes, then it is determined to be an abnormal performance; Anomaly types are identified based on an anomaly classification model, which is trained on a pre-set historical default case database. Based on the type of exception, the corresponding exception evidence storage template is matched from the pre-configured evidence storage strategy library; Using the abnormal evidence template and based on the abnormality type, performance status vector, and deviation judgment criteria, an abnormal evidence data packet is generated and a hash value of the abnormal evidence is obtained by hash calculation. Submit the abnormal evidence hash value to the blockchain for evidence storage, and update the corresponding clause-level hash subchain based on the abnormal evidence hash value.

Images