A method and system for chain consistency and explainability of evidence

Abstract

This invention relates to the field of authenticity verification technology, proposing a method and system for authenticity verification based on the consistency and interpretability of the evidence chain. The authenticity verification method includes the following steps: data acquisition and preprocessing, evidence chain construction and association mapping, multi-dimensional consistency verification, interpretability verification, result feedback and dynamic updating. The authenticity verification system includes a data acquisition and preprocessing module, an evidence chain construction module, a multi-dimensional consistency verification module, an interpretability verification module, a dynamic updating and storage module, and a user interaction module. In summary, this invention achieves full-dimensional consistency verification of report documents and associated evidence chains, while outputting an interpretable verification report that conforms to judicial norms, clearly defining the basis for verification, points of contradiction, and reasoning logic. It aligns with the requirements of judicial blockchain evidence storage and cross-chain collaboration, improves the judicial credibility of the verification results, and balances universality, scalability, and efficiency.

Description

Technical Field

[0001] This invention relates to the field of authenticity identification technology, and in particular to a method and system for authenticity identification that provides consistency and interpretability of the evidence chain. Background Technology

[0002] In judicial practice, reports serve as the core basis for determining the facts of a case and applying the law. Their authenticity directly affects judicial fairness, case outcomes, and the legitimate rights and interests of the parties involved. They are crucial for maintaining judicial authority and advancing the modernization of the trial system and trial capabilities. For example, legal reports such as forensic appraisal reports, notarized documents, and inspection records all require a complete, coherent, and consistent chain of evidence to support their legal validity.

[0003] In existing technologies, the authentication of legal industry reports mainly focuses on detecting tampering of single documents (such as hash verification and digital signature verification) or verifying the authenticity of isolated evidence, which has the following core shortcomings: First, there is a lack of overall consistency verification of legal reports and related evidence chains (such as case records, original expert data, witness testimonies, legal provisions, judicial blockchain evidence data, etc.), making it difficult to identify covert forgery behaviors such as "the content of the report is out of touch with the facts of the case", "contradictions between pieces of evidence", and "conflicts between the report and the application of legal provisions".

[0004] Secondly, the identification process suffers from a "black box" problem. Most identification algorithms only output "true / false" results, failing to explain the basis for identification, the location of contradictions, and the reasoning logic, making it difficult for judicial authorities to accept the identification results.

[0005] Third, existing systems are mostly designed for general fields and do not take into account the characteristics of the legal industry (such as the legality, relevance, and legality requirements of the chain of evidence). They have poor universality and are difficult to adapt to the scenario of dynamic updates of the chain of evidence in legal cases. They also lack scalability and cannot meet the needs of cross-chain collaborative applications of judicial blockchain.

[0006] Furthermore, with the development of generative AI and tampering technologies, the methods for forging legal reports and related evidence are becoming increasingly sophisticated. Traditional methods relying on manual review are inefficient, highly subjective, and limited by professional capabilities, making it difficult to comprehensively investigate covert forgeries. Existing technologies cannot simultaneously ensure accuracy in identification, consistency of the evidence chain, and interpretability of the process. They also fail to fully adapt to the application requirements of judicial blockchain evidence storage and cross-chain collaboration, making it difficult to meet the stringent demands of the legal industry for verifying the authenticity of reports and documents, thus hindering the progress of smart court construction. Therefore, there is an urgent need for a method and system for authenticity verification that is tailored to the characteristics of the legal industry, adapted to judicial blockchain applications, capable of verifying the consistency of the evidence chain in legal reports and documents, providing an interpretable identification process, and being universal and scalable, to address the shortcomings of existing technologies. Summary of the Invention

[0007] In view of this, the present invention proposes a method and system for identifying the authenticity of evidence chains based on consistency and interpretability. It is mainly aimed at various reports and documents in the legal industry, realizing full-dimensional consistency verification between the report documents and the associated legal evidence chains. At the same time, it outputs an interpretability identification report that conforms to judicial norms, clearly defining the identification basis, contradictions and reasoning logic, meeting the requirements of judicial blockchain evidence storage and cross-chain collaboration, improving the judicial credibility of the identification results, and taking into account universality, scalability and efficiency.

[0008] The technical solution of this invention is implemented as follows: On the one hand, this invention provides a method for verifying the authenticity of evidence chains based on consistency and interpretability, comprising the following steps: Step 1: Data Acquisition and Preprocessing Obtain legal documents to be identified and evidence chain data, preprocess the documents to be identified and evidence chain data to obtain standardized data.

[0009] The aforementioned evidence chain data includes the original data source of the report documents (such as original test data of judicial appraisal, original notarization records), the generation process records (such as appraisal process records, notarization process records), signature information (such as the signatures of appraisers, notaries, and judges), related supporting materials (such as case records, witness testimonies, inspection and examination records, legal basis, judicial blockchain evidence storage data), and legal industry background knowledge (such as judicial appraisal standards, notarization procedure rules, and relevant legal and regulatory provisions).

[0010] The preprocessing described above includes format standardization, noise removal, and key information extraction. Key information extraction employs Natural Language Processing (NLP) technology, combined with a legal industry dictionary, to extract the core conclusions, key data, signing information, legal citations, and evidence citation relationships from the report documents. Simultaneously, it extracts corresponding key information from the evidence chain data (such as legal provisions, core content of witness testimonies, and hash values ​​of judicial blockchain evidence storage), forming a key information set to ensure the accuracy of subsequent consistency verification. Referring to the key information extraction logic of the BiLSTM-CRF model, the model parameters are optimized based on the characteristics of legal texts to improve the completeness and accuracy of legal information extraction. Furthermore, it connects to the judicial blockchain platform to achieve efficient retrieval and parsing of evidence storage data.

[0011] Step Two: Evidence Chain Construction and Association Mapping Based on the standardized data mentioned above, a legal evidence chain network for the aforementioned report documents to be identified is constructed, and a mapping relationship is established between the key information of the aforementioned report documents to be identified and the nodes of the evidence chain.

[0012] The aforementioned legal evidence chain network takes the core conclusions of the aforementioned reports to be identified (such as expert opinions and notarization conclusions) as its vertices, related legal evidence (such as original test data, witness testimonies, legal provisions, and judicial blockchain evidence data) as its nodes, and the logical and legal relationships between evidence and conclusions, and between different pieces of evidence, as its edges.

[0013] The system clearly defines the supporting evidence for each core conclusion, the report citation location for each piece of evidence, and the legal basis for each piece of evidence, forming a four-dimensional association system of "conclusion-evidence-legal provision-existing evidence". Referring to the evidence chain construction mechanism of the DeepRare system, it introduces evidence confidence parameters (combining the legality, authenticity, and relevance of evidence) to provide a quantitative basis for subsequent consistency verification. At the same time, it adapts to the cross-chain collaborative application requirements of judicial blockchain and realizes the association mapping between the evidence chain and judicial blockchain data.

[0014] Step 3: Multi-dimensional Consistency Verification Based on the aforementioned legal evidence chain network and associated mapping relationships, consistency verification is performed from three core dimensions: content consistency verification, temporal consistency verification, and association consistency verification, and verification results are generated.

[0015] Content consistency verification: The core conclusions, key data, legal citations, and information of corresponding evidence chain nodes in the identification report are compared with the judicial blockchain evidence storage data to determine whether there are numerical contradictions, logical conflicts, or errors in the application of legal provisions. The cosine similarity algorithm is combined with a legal logic rule engine to quantify the content matching degree, ensure the integrity of evidence data transmission and storage, and prevent evidence from being tampered with.

[0016] Temporal consistency verification: Compare the generation time and signing time of the identification report with the generation time, circulation time and judicial blockchain storage time of the evidence chain data to determine whether there are temporal inconsistencies. Combine timestamp technology, digital signature verification and judicial blockchain timestamp verification to ensure the rationality of the temporal logic.

[0017] Association Consistency Verification: Verify the legal and logical connections between evidence nodes in the legal evidence chain network, determine whether there are evidence gaps, contradictions, or breaks in the evidence chain, construct a legal evidence association network using knowledge graph technology, incorporate legal logical reasoning rules, and identify association anomalies through graph neural network algorithms.

[0018] Step 4: Interpretability Identification Based on the above verification results and in accordance with the legal industry's evidence identification standards, the authenticity of the evidence is determined, and an interpretability identification report is generated.

[0019] If all dimensions pass the verification and the chain of evidence meets the legal requirements, the report document is deemed genuine; if any dimension fails the verification, specific contradictions, abnormal evidence, conflict types, and issues related to the application of legal provisions are identified, and an interpretable identification report that conforms to judicial norms is generated.

[0020] The aforementioned interpretable authentication report includes: authenticity determination results, details of contradictions, authentication basis, and reasoning logic chain; referencing the traceable reasoning mechanism of the DeepRare system, it achieves full transparency of the authentication process, while introducing legal document authentication technology to provide additional support for the authentication of document forgery, ensuring that the authentication report can be directly used as auxiliary evidence in judicial proceedings and adapting to the verification requirements of the judicial blockchain verification platform.

[0021] Step 5: Results Feedback and Dynamic Updates The aforementioned interpretability identification report is fed back to the user, while supplementary evidence, objection information, or newly stored evidence data from the judicial blockchain are received from the user. The evidence chain network is dynamically updated, and consistency verification is re-executed to achieve dynamic optimization of the identification results.

[0022] The aforementioned users include judges, lawyers, forensic experts, and notaries.

[0023] In addition, the identification data will be stored in a distributed database and simultaneously uploaded to the judicial blockchain platform for evidence preservation. This will be used for model optimization and rapid identification of subsequent similar legal reports. By referring to the logic of judicial evidence preservation and dynamic updates of the judicial blockchain, the system's adaptability will be improved.

[0024] Based on the above technical solutions, preferably, in step one above, the key information extraction also includes the extraction of hidden features from the report documents to be identified and the evidence chain data, including dot matrix marks on printed legal documents, handwriting features of handwritten signatures, physical and chemical features of seal impressions, and anti-counterfeiting marks on notarized documents.

[0025] Based on the above technical solutions, preferably, in step three above, the multi-dimensional consistency verification adopts a quantitative scoring mechanism, combined with the legal industry evidence recognition weight, and sets a verification score (0-100 points) for each dimension. The score of each dimension is calculated based on the content matching degree, temporal rationality, correlation completeness and legal compliance. The total score = content consistency score × 45% + temporal consistency score × 30% + correlation consistency score × 25%. If the total score is ≥ 85 points, it is judged as initially qualified, and then the contradiction points are reviewed in combination with legal norms. If the total score is < 85 points, it is directly judged as unqualified, and the contradiction points and relevant legal provisions corresponding to the dimension with the lowest score are marked first.

[0026] Based on the above technical solutions, preferably, in step four above, the reasoning logic chain is presented in the form of "tree structure + legal citation". The result of the authenticity judgment is the root node, the verification results of each dimension are the first-level branches, and the contradiction points, evidence information, and legal basis are the second-level branches. It clearly shows the basis for the formation of each judgment conclusion and the corresponding legal support. At the same time, it supports users to click on the branch nodes to view the detailed verification process, algorithm principle and relevant legal provisions, judicial blockchain evidence details, and solve the "black box" problem of existing identification technology.

[0027] On the other hand, the present invention also provides a system for verifying the authenticity of evidence chain consistency and interpretability, including: a data acquisition and preprocessing module, an evidence chain construction module, a multi-dimensional consistency verification module, an interpretability verification module, a dynamic update and storage module, and a user interaction module.

[0028] Data acquisition and preprocessing module: used to acquire the report documents to be identified and the evidence chain data, connect to the judicial blockchain platform to retrieve the stored evidence data, perform format standardization, noise removal, key information extraction operations, and output standardized data.

[0029] Evidence Chain Construction Module: Based on the above standardized data, a legal evidence chain network is constructed, establishing a mapping relationship between key information of the report document to be identified and evidence chain nodes and judicial blockchain evidence storage data, and outputting a four-dimensional association system of "conclusion-evidence-legal provisions-existing evidence".

[0030] Multi-dimensional consistency verification module: Used to verify the consistency of content, time sequence, and association of evidence chains, output verification scores and details of contradictions, and obtain verification results.

[0031] Explanability Authentication Module: Based on the verification results, it determines the authenticity of the data and generates an explainability authentication report that conforms to judicial standards.

[0032] Dynamic update and storage module: Used to receive user feedback, dynamically update the evidence chain network, verification parameters and judicial blockchain related data, and store relevant data for this identification.

[0033] User interaction module: Provides users with functions such as file upload, authentication request submission, authentication report viewing, feedback input, historical data query, and retrieval of judicial blockchain evidence data. It supports users in raising objections to contradictions in the authentication report and supplementing evidence, and also provides access control functions.

[0034] Based on the above technical solutions, preferably, the system also includes a model optimization module. The model optimization module iteratively optimizes the consistency verification algorithm and the key information extraction algorithm based on the historical identification data, judicial practice feedback and judicial blockchain evidence storage data in the dynamic update and storage module. It also dynamically adjusts the algorithm parameters in combination with new laws and regulations in the legal industry to improve the accuracy and efficiency of identification.

[0035] Based on the above technical solutions, the preferred multi-dimensional consistency verification module also integrates an anomaly detection unit to identify outliers in evidence chain data and judicial blockchain evidence storage data.

[0036] Based on the above technical solutions, the preferred interpretability identification module also supports the export and sharing of identification reports, and generates standardized identification opinions.

[0037] The method and system for verifying the consistency and interpretability of the chain of evidence in this invention have the following advantages over the prior art: 1. It aligns with the characteristics of the legal industry, enabling full-dimensional consistency verification of the legal evidence chain and breaking through the limitations of existing technologies such as "single file verification" and "generalized adaptation".

[0038] 2. Solve the "black box" problem in the identification process and improve the judicial credibility of the results.

[0039] 3. It has strong versatility and good scalability, and is suitable for multiple scenarios in the legal industry and judicial blockchain applications.

[0040] 4. Highly efficient, convenient, and intelligent, reducing the barrier to entry for users while balancing efficiency and ease of use.

[0041] 5. Data security is traceable and complies with judicial data management standards. Detailed Implementation

[0042] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0043] Example 1

[0044] A method for verifying the consistency and interpretability of a chain of evidence includes the following steps: 1. Data Acquisition and Preprocessing The system obtains the DNA forensic report and related legal evidence chain data for the criminal case to be identified. It then connects to the judicial blockchain platform to retrieve the report and related evidence's stored data. The evidence chain data includes: original DNA testing data, qualification certificates of the forensic personnel, testing equipment operation logs, related case records, relevant legal provisions from the "General Rules of Forensic Identification Procedures" and the "Criminal Procedure Law," and the judicial blockchain's stored evidence hash value. The preprocessing process includes: converting the PDF-format appraisal report into editable text, removing redundant formatting and noisy data; using a BiLSTM-CRF model optimized with a legal industry dictionary to extract the core conclusions, key data, signature information, legal citations, and evidence citation relationships from the report; simultaneously extracting corresponding information from the evidence chain data, including the matching degree value, testing time, equipment number, appraisal personnel qualification information, submission time in the case record, original text of legal provisions, and hash value of judicial blockchain evidence storage; using microscopic analysis algorithms to extract the handwritten signature features and seal features of the appraisal institution on the report's signature page to form a set of key information; in addition, calculating the SM3 hash value of all data and conducting a preliminary comparison with the hash value of judicial blockchain evidence storage for subsequent integrity verification, referencing hash verification techniques in judicial video tracing and judicial blockchain evidence storage, in line with the Supreme People's Court's requirements for the reliable protection of electronic evidence.

[0045] 2. Evidence chain construction and association mapping Combining the standards for evidence identification in criminal cases, this method uses the core conclusion of the forensic report, "The DNA match between the submitted sample and the suspect's sample is 99.99%," as the apex. It uses the original DNA testing data, the forensic personnel's qualification certificates, equipment operation logs, case records, relevant legal provisions, and judicial blockchain-stored evidence data as evidence nodes. It constructs a legal evidence chain network with logical edges: "Testing data supports the conclusion," "Equipment logs corroborate the validity of the testing," "The legality of the forensic personnel's qualification certificate report," "Case records relate to the testing background," "Legal provisions support the compliance of the forensic procedure," and "Judicial blockchain-stored evidence corroborates the authenticity of the data." It establishes a mapping relationship, such as the correspondence between "DNA match rate of 99.99%" in the report and... The evidence chain includes the "matching index of the original DNA test data" and the matching data in the judicial blockchain evidence storage. The report signing time corresponds to the "test completion time + reasonable review period" and the judicial blockchain evidence storage time. The legal provisions cited in the report correspond to the relevant original text of the "General Rules for Judicial Appraisal Procedures". At the same time, the confidence level of each evidence node is calculated (such as 95% confidence level of the original test data, 99% confidence level of the judicial blockchain evidence storage data, and 98% confidence level of the appraiser's qualifications), forming a four-dimensional association system of "conclusion-evidence-legal provisions-evidence storage". The evidence chain construction and confidence level calculation logic of the DeepRare system are referenced and optimized in combination with the evidence requirements of criminal cases to adapt to the cross-chain collaborative application needs of judicial blockchain.

[0046] 3. Multi-dimensional consistency verification Content consistency verification: The "DNA matching degree 99.99%" in the comparison report matches the original test data (99.99%) in the evidence chain and the matching degree (99.99%) in the judicial blockchain evidence storage data. The matching degree calculated using the cosine similarity algorithm is 100%, with no numerical contradictions. The consistency of the "DNA matching" conclusion with the sample information submitted in the case record and the compliance with the identification process norms in the "General Rules for Judicial Appraisal Procedures" are verified through the legal logic rule engine. It is confirmed that the legal provisions are correctly cited and the identification process is compliant, and the content consistency score is determined to be 98 points.

[0047] Time sequence consistency verification: The report signing time (March 10, 2026) was compared with the test completion time (March 8, 2026), the test time of the equipment operation log (March 8, 2026), and the judicial blockchain evidence storage time (March 8, 2026). The review period was 2 days, which complies with the reasonable review period stipulated in the "General Rules for Judicial Appraisal Procedures" and there was no time sequence contradiction. At the same time, the timestamp of the digital signature, the timestamp of the judicial blockchain and the report signing time were verified to be consistent, and the time sequence consistency score was determined to be 96 points.

[0048] Association Consistency Verification: Using a graph neural network algorithm combined with the rules for association of evidence in criminal proceedings, the association between nodes in the evidence chain is analyzed. The association between the detection data and the equipment logs, the association between the detection data and the case records, the association between the signed information and the qualifications of the appraisers, the association between the report and the judicial blockchain evidence storage data, and the fact that all evidence is related to the facts of the case with no disconnection or contradiction, the association consistency score is determined to be 97 points.

[0049] The total score = 98×45%+96×30%+97×25%=97.15 points ≥ 85 points, which is initially judged as qualified. The contradictions were reviewed in conjunction with the Criminal Procedure Law and the General Rules of Judicial Appraisal Procedure, and no abnormal contradictions or procedural violations were found.

[0050] 4. Interpretability identification and result generation Based on the verification results and in accordance with the standards for the identification of evidence in criminal cases, the DNA forensic identification report to be identified is determined to be genuine. An interpretable identification report conforming to judicial norms is generated, including: the result of authenticity (genuine), details of contradictions (none), identification basis (content consistency score 98, temporal consistency score 96, correlation consistency score 97, total score 97.15, meeting the passing standard, based on Article 50 of the Criminal Procedure Law and Article 28 of the General Rules for Forensic Identification Procedures), and a logical chain of reasoning (a tree structure, with the root node being "Report is True," first-level branches being the verification results of the three dimensions, and second-level branches being the verification details of each dimension, corresponding evidence nodes, legal basis, and details of judicial blockchain evidence storage, such as the content consistency branch marked "The report's DNA matching degree is consistent with the original test data and judicial blockchain evidence storage data, and the legal citation is compliant"). Simultaneously, SM3 hash verification results, judicial blockchain evidence storage verification results, and signature verification results are attached to enhance the report's judicial authority and ensure that the report can serve as supplementary evidence for cross-examination in criminal cases, adapting to the verification requirements of the judicial blockchain verification platform.

[0051] 5. Results Feedback and Dynamic Updates The interpretability identification report is fed back to the presiding judge. If the judge has no objection, the identification data (identification report, evidence chain data, verification results, and identification report) is stored in a distributed database and simultaneously uploaded to the judicial blockchain platform for evidence preservation. If supplementary evidence (such as new DNA retest data) or new evidence information is added to the judicial blockchain, the system will receive the supplementary evidence, dynamically update the evidence chain network, and re-perform consistency verification to ensure the accuracy of the identification results. This process, referencing the judicial evidence preservation and dynamic updating of the judicial blockchain, aligns with the Supreme People's Court's requirements for on-chain storage and dynamic collaboration of judicial data.

[0052] If the report to be identified is found to be forged, for example, if the report states "DNA matching rate 99.99%", while the matching rate of the original test data is 80% and the matching rate of the judicial blockchain evidence storage data is 80%, then the content consistency verification score will be lower than 60 points and the total score will be lower than 85 points. The system will determine that the report is forged. The interpretable identification report will clearly mark the contradictions and identification basis, present the reasoning logic chain, and mark the results of hidden feature extraction, providing sufficient judicial support for the forgery determination. It can be directly used to identify forged evidence in cases.

[0053] Example 2

[0054] A system for verifying the consistency and interpretability of the chain of evidence includes: Data acquisition and preprocessing module It supports uploading multiple file formats, including PDF, images, text, and judicial blockchain evidence documents, and connects to the judicial blockchain platform to achieve efficient retrieval and parsing of evidence data. It integrates an NLP key information extraction unit and a hidden feature extraction unit. The NLP unit uses a BiLSTM-CRF model optimized based on the characteristics of legal texts to extract key information from reports and evidence chains. The hidden feature extraction unit uses microscopic analysis algorithms and CV technology to extract features such as signatures, imprints, and dot matrix marks, optimized in accordance with legal document authentication standards. Simultaneously, it performs format standardization and noise removal operations, outputting standardized data. It also integrates a hash calculation unit to calculate the SM3 hash value of all data and conduct a preliminary comparison with the hash value of judicial blockchain evidence to ensure data integrity. Referencing technical solutions for document authentication, judicial video tracing, and judicial blockchain evidence storage, it aligns with the Supreme People's Court's requirements for the reliable protection of electronic evidence.

[0055] Evidence Chain Construction Module The system comprises a network construction unit, an association mapping unit, and a judicial blockchain docking unit. The network construction unit utilizes knowledge graph technology, integrating the logical rules of evidence in criminal cases with relevant norms from the Criminal Procedure Law and the General Rules of Judicial Appraisal Procedures. It constructs a legal evidence chain network with the core conclusions of the report as vertices, evidence as nodes, and legal logical relationships as edges. The association mapping unit establishes precise associations between key information in the report and evidence nodes, legal provisions, and judicial blockchain-stored evidence data, while simultaneously calculating the confidence level of each evidence node. The judicial blockchain docking unit achieves seamless integration with the judicial blockchain platform, supporting the retrieval, association, and verification of stored evidence data. It supports the addition, deletion, and modification of evidence chain nodes, adapting to dynamic update requirements. Referencing the evidence chain construction architecture of the DeepRare system, and combining criminal case evidence requirements with judicial blockchain application optimizations, it enhances the completeness, rationality, and judicial adaptability of the evidence chain network.

[0056] 3. Multi-dimensional Consistency Verification Module: This module includes a content verification unit, a time-series verification unit, a correlation verification unit, an anomaly detection unit, a hash chain verification unit, and a judicial blockchain evidence storage verification unit. The content verification unit employs a cosine similarity algorithm and a legal logic rule engine (incorporating laws and regulations such as the Criminal Procedure Law and the General Rules for Judicial Appraisal Procedures). The time-series verification unit combines timestamp, digital signature verification, and judicial blockchain timestamp verification technology. The correlation verification unit uses a graph neural network algorithm and incorporates evidence correlation reasoning rules for criminal cases. The anomaly detection unit uses the 3σ principle to identify abnormal evidence and false evidence storage data. The hash chain verification unit and the judicial blockchain evidence storage verification unit ensure that the evidence data has not been tampered with and that the evidence storage is valid. Each unit outputs a verification score, and the system automatically calculates the total score, annotates the details of contradictions and the corresponding legal provisions and evidence storage anomaly information, and references the multi-dimensional verification logic of electronic evidence judicial recognition and judicial blockchain anomaly data identification to improve the accuracy and legal compatibility of the verification.

[0057] 4. Explanability Authentication Module: This module includes a authenticity determination unit, a report generation unit, and a reasoning logic presentation unit. The authenticity determination unit determines the authenticity of an application based on the total score, details of contradictions, and the standards for evidence identification in criminal cases. The report generation unit generates an interpretability authentication report that conforms to judicial norms, supports export in PDF and Word formats, and generates a standardized authentication opinion that can be directly used as supporting evidence in court. It also supports uploading the authentication report to the judicial blockchain platform for evidence storage. The reasoning logic presentation unit displays the reasoning process in a "tree structure + legal citation" format, linked to the judicial blockchain evidence storage details. Users can click to view the detailed verification process, algorithm principles, legal provisions, and evidence storage data. By referencing the reasoning presentation method of interpretable AI and the logic of judicial judgments, the authentication process is made transparent, improving the judicial credibility of the authentication results.

[0058] 5. Dynamic Update and Storage Module: This module integrates a distributed database unit, a dynamic optimization unit, and a judicial blockchain evidence storage unit. The database unit uses IPFS distributed storage technology to store report files, evidence chain data, verification results, identification reports, and other data, ensuring data security. The judicial blockchain evidence storage unit enables the on-chain storage of identification-related data, ensuring data traceability and verifiability, in line with the Supreme People's Court's requirements for on-chain storage of judicial data. The dynamic optimization unit receives user feedback, updates the evidence chain network, verification parameters, and judicial blockchain-related data, while also storing historical identification data to support model optimization. By referencing the evidence storage and dynamic update technologies of the judicial blockchain, the system's scalability and adaptability are enhanced.

[0059] 6. User Interaction Module: Provides functions such as file upload, authentication request submission, authentication report viewing, feedback input, historical data query, and retrieval of judicial blockchain evidence storage data; supports users to raise objections to authentication reports and supplement evidence; integrates a permission management unit to differentiate the operation permissions of different roles such as judges, judicial appraisers, and administrators, ensuring judicial data security and operational standardization; the interface design conforms to the operating habits of the smart court platform, improves the system's usability, and adapts to the usage needs of legal industry users.

[0060] 7. Model Optimization Module: Based on historical identification data, judicial practice feedback, and judicial blockchain evidence data from the dynamic update and storage module 5, deep learning algorithms are used to iteratively optimize the key information extraction algorithm and consistency verification algorithm. The legal logic rule engine parameters are adjusted in conjunction with the revisions to the Criminal Procedure Law and the General Rules for Judicial Appraisal Procedures. User-defined verification rules and weight allocations are supported (e.g., increasing the correlation consistency weight in criminal cases), adapting to the needs of different case types such as criminal and civil cases. Referring to the model iteration logic of the DeepRare system, and combining the characteristics of the legal industry with judicial blockchain application optimization, the system's versatility and scalability are improved, contributing to the construction of smart courts.

[0061] The workflow of the authenticity verification system based on the consistency and interpretability of the chain of evidence is as follows: Users (such as presiding judges and forensic experts) upload legal reports and related evidence chain data to be authenticated through the user interaction module. The system retrieves relevant evidence data through the judicial blockchain docking unit. The data acquisition and preprocessing module preprocesses the data and outputs standardized data. The evidence chain construction module builds a legal evidence chain network and establishes a mapping relationship between key information in the report and evidence nodes, legal provisions, and judicial blockchain evidence data. The multi-dimensional consistency verification module performs multi-dimensional verification and outputs verification scores and details of contradictions. The interpretability authentication module generates an interpretable authentication report, feeds it back to the user, and uploads the report to the judicial blockchain platform for evidence storage. Users can input feedback through the user interaction module, and the dynamic update and storage module updates the evidence chain network, verification parameters, and evidence data, and stores historical authentication data. The model optimization module iteratively optimizes the algorithm based on historical data, judicial practice feedback, and evidence data to improve the system's authentication performance and adapt to the dynamic needs of smart court construction and judicial blockchain applications.

[0062] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method of forensic chain consistency and explainability, characterized in that: Includes the following steps: Step 1: Data Acquisition and Preprocessing Obtain the report document to be identified and the evidence chain data, and preprocess the report document to be identified and the evidence chain data to obtain standardized data; Step Two: Evidence Chain Construction and Association Mapping Based on the standardized data, a legal evidence chain network for the report documents to be identified is constructed, and a mapping relationship between the key information of the report documents to be identified and the nodes of the evidence chain is established. Step 3: Multi-dimensional Consistency Verification Based on the legal evidence chain network and associated mapping relationships, consistency verification is performed from three core dimensions: content consistency verification, temporal consistency verification, and association consistency verification, and verification results are generated. Step 4: Interpretability Identification Based on the verification results, the authenticity is determined and an interpretability verification report is generated. Step 5: Results Feedback and Dynamic Updates The interpretability identification report is fed back to the user, and supplementary evidence, objection information or newly stored evidence data in the judicial blockchain are received from the user. The evidence chain network is dynamically updated and consistency verification is re-executed to achieve dynamic optimization of the identification results.

2. The method for verifying the authenticity of evidence chain consistency and interpretability as described in claim 1, characterized in that: In step one, the evidence chain data includes the original data source of the report document, the generation process record, signing information, related supporting materials, and legal industry background knowledge.

3. The method for verifying the authenticity of evidence chain consistency and interpretability as described in claim 1, characterized in that: In step one, preprocessing includes format standardization, noise removal, and key information extraction.

4. The method for verifying the authenticity of evidence chain consistency and interpretability as described in claim 1, characterized in that: In step two, the content consistency verification involves comparing the core conclusions, key data, legal citations, and information of the corresponding evidence chain nodes in the identification report with the judicial blockchain evidence data to determine whether there are numerical contradictions, logical conflicts, or errors in the application of legal provisions.

5. The method for verifying the consistency and interpretability of the chain of evidence as described in claim 1, characterized in that: In step two, the time sequence consistency verification is as follows: compare the generation time and signing time of the identification report document with the generation time, circulation time and judicial blockchain storage time of the evidence chain data to determine whether there is a time sequence contradiction.

6. The method for verifying the authenticity of evidence chain consistency and interpretability as described in claim 1, characterized in that: In step two, the consistency verification is to verify the legal and logical connections between the evidence nodes in the legal evidence chain network, and to determine whether there are any disconnects, contradictions, or breaks in the evidence chain.

7. The method for verifying the authenticity of evidence chain consistency and interpretability as described in claim 1, characterized in that: In step three, the multi-dimensional consistency verification adopts a quantitative scoring mechanism, which combines the evidence recognition weight of the legal industry. Each dimension is set with a verification score, and the scores of each dimension are calculated based on the content matching degree, the reasonableness of the time sequence, the completeness of the association, and the legal compliance.

8. A system for verifying the consistency and interpretability of the chain of evidence, applied in the method for verifying the consistency and interpretability of the chain of evidence as described in claim 1, characterized in that: It includes a data acquisition and preprocessing module, an evidence chain construction module, a multi-dimensional consistency verification module, an interpretability identification module, a dynamic update and storage module, and a user interaction module.