Memetic method structure responsibility-accounting labeling system and labeling method

Abstract

The invention discloses a memetic method structure responsibility-accounting marking system and a marking method, legal text structure analysis is carried out through a memetic path, and a responsibility escape path hidden in a formal legal language is revealed. In a judgment generation / compliance auditing stage, a potential responsibility escape structure can be identified to assist a judge or auditor in judging the effectiveness of a judgment; in an AI legal language system, the responsibility classifying capability of word judgment can be improved, and the AI is prevented from mistakenly recognizing that a sentence pattern seemingly closed is legal; ability training of'slave language recognition structure 'can be formed in law education, and the structural gap of judicial semantic training is filled; in the social governance field, the behavior of legal appearance responsibility shielding can be revealed, and a quantitative basis is provided for legislation and supervision.

Description

Technical Field

[0001] This application belongs to the field of intersectional technology of artificial intelligence and law, and in particular relates to a memetic law structural attribution annotation system and annotation method. Background Art

[0002] Existing AI legal systems can generate legal documents (such as judgments and agreements), but they often fail to determine whether the attribution logic within these documents is truly closed. Traditional logical deduction models focus on the causal relationship between facts and conclusions, but ignore the fact that in complex situations with multiple parties and multiple chains of responsibility, the question of "who is responsible" is often obscured, misplaced, or omitted by structural language, leading to judicial misjudgments and the drift of responsibility.

[0003] Existing legal NLP systems typically rely on keyword recognition and rule matching, but they are unable to penetrate the semantic surface of formal language like "settlement," "agreement," and "no further pursuit of responsibility" to capture the underlying structural paths of evasion. Currently, no mature system can label, categorize, and track structural responsibility imbalances in text using a memetic path approach. Summary of the Invention

[0004] The purpose of this application is to provide a memetic law structural attribution annotation system and annotation method, to analyze the legal text structure through the "Memetic Path", to reveal the responsibility escape path hidden in formal legal language, and to solve the above-mentioned technical problems.

[0005] To achieve the above application objectives, the technical solutions adopted in this application are as follows: The embodiment of the present application provides a meme-based structural attribution annotation system, including a meme path annotation library and an annotation engine; The meme path annotation library includes M path, S path and C path; M path includes the missing / ambiguous responsible subject category, which includes M0001 missing subject module, M0024 labor relationship masking module, and M0042 pseudo causal chain module; S path includes the systemic language failure category, which includes S001 ambiguous structure module, S002 closed loop module, and S003 semantic drift module; C path includes the structure legal disguise category, which includes C0001 language format masking module, C0004 de-masking structure module, and C0006 multi-path overflow module; among them, each module can be compositely connected and the path can be expanded; The annotation engine combines grammatical analysis, syntactic dependency trees, and meme path matching models to automatically annotate potential attribution breakpoints in digitized corpora with meme numbers. This can be embedded in AI adjudication systems and regulatory compliance processes, for example.

[0006] In some embodiments, the memetic law structural attribution annotation system also includes an application interface module for integrating the system into the government compliance supervision sandbox as a plug-in; or integrating it into the legal NLP model as a plug-in to improve AI attribution capabilities; or integrating it into the legal training system as a plug-in to assist students and judges in structural language recognition training.

[0007] In some embodiments, the memetic structural attribution annotation system further includes a multi-language compatibility module for compatibility with multiple languages.

[0008] In some embodiments, the meme path annotation library also includes: P path, the P path includes the structural damage identification class caused by power language, and the structural damage identification class caused by power language includes: manipulative language module and humiliating language module.

[0009] The present application also provides a method for labeling the above-mentioned memetic structural attribution labeling system, including: Import digitized corpus; Sentence the data-based corpus and construct a syntactic tree to analyze the subject, predicate, and causal logical relationships; Through the annotation engine, the meme path annotation library is called to perform matching annotation; Output the annotated text and meme structure analysis report, and determine whether to issue an early warning based on the meme structure analysis report.

[0010] In some embodiments, outputting the annotated text and the meme structure analysis report, and determining whether to issue a warning based on the meme structure analysis report includes: By matching the annotated modules, it is determined whether a complete meme chain exists. If so, an early warning is issued. For example, a complete meme chain indicates a risk of responsibility evasion, and the meme chain path can be used as early warning content to prompt users or managers.

[0011] In some embodiments, after outputting the annotated text and the meme structure analysis report and determining whether to issue a warning based on the meme structure analysis report, the method further includes: Synchronously visualize responsibility paths for educational presentations or system attribution scoring.

[0012] Compared with the prior art, the beneficial effects of the embodiments of the present application are: The embodiments of the present application provide a memetic law structural attribution annotation system and annotation method, which can identify potential attribution evasion structures in the judgment generation / compliance audit stage to assist judges or auditors in judging the validity of judgments; in the AI ​​legal language system, it can improve the attribution ability of judgments to prevent AI from misidentifying "seemingly closed" sentences as legal case closures; in legal education, it can form the ability training of "identifying structures from language" to fill the structural gap in judicial semantic training; in the field of social governance, it can reveal "legal appearance responsibility concealment" behavior and provide a quantitative basis for legislation and supervision. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0014] Figure 1 A schematic diagram of the meme path structure classification provided by an embodiment of the present application is shown.

[0015] Figure 2 A flow chart of attribution logic fracture detection provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0016] In order to make the technical problems, technical solutions and beneficial effects to be solved by this application more clearly understood, this application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0017] It should be noted that when an element is referred to as being “fixed on” or “disposed on” another element, it may be directly on the other element or indirectly on the other element. When an element is referred to as being “connected to” another element, it may be directly connected to the other element or indirectly connected to the other element.

[0018] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. Throughout the description of this application, "plurality" means two or more, unless otherwise specifically defined.

[0019] This application relates to the intersection of artificial intelligence and law, specifically a system and method for identifying and annotating attribution logic gaps and structural obscured paths in digitized corpus. It belongs to the technical category of semantic annotation engines and legal decision-making assistance tools, and can be widely used in judicial decision assistance, compliance review, AI judgment structure verification, and legal education and training.

[0020] Specifically, the embodiments of the present application provide a memetic structural attribution annotation system and annotation method, which conducts structural analysis of digitized corpus through "Memetic Path" to reveal the responsibility evasion paths hidden in formal legal language, including but not limited to: subject missing, causal break, systematic concealment (gray industry), emotional substitution, disguised case-closing language, etc.

[0021] It should be noted that digital corpus includes but is not limited to text, audio, and video data. Examples include: company new employee training manuals, employee labor contracts, interview audio and video, internal notices; mediation documents issued by police stations, arbitration institutions, mediation agencies, etc.; or legal documents, procedural documents, and judgments from courts.

[0022] In order to illustrate the technical solution described in this application, specific embodiments are provided below.

[0023] A memetic structural attribution annotation system, including a memetic path annotation library and an annotation engine.

[0024] See also Figure 1 As shown, the meme path annotation library includes M path, S path and C path; M path includes the responsible subject missing / ambiguous class, and the responsible subject missing / ambiguous class includes M0001 subject missing module, M0024 labor relationship masking module, and M0042 pseudo causal chain module; S path includes systemic language failure class, and the systemic language failure class includes S001 ambiguous structure module, S002 closed loop module, and S003 semantic drift module; C path includes structural legal disguise class, and the structural legal disguise class includes C0001 language format masking module, C0004 de-masking structure module, and C0006 multi-path overflow module.

[0025] As an example, the above classes and modules can be understood as follows: Memetic Path Annotation Library: A collection for storing and managing different types of memetic paths, which are used to annotate and analyze attribution-related structures in legal texts.

[0026] M path: It can include the missing / ambiguous responsible party category, which focuses on problems in identifying the responsible party and covers situations where the responsible party is missing or ambiguous.

[0027] S-path: This can include the systemic language failure category, which mainly marks situations where systemic language failure occurs, such as language structure problems in legal texts that make it impossible to accurately convey responsibility information.

[0028] C-path: can include a structural legal disguise class, which is used to identify situations that appear to be legal but are actually structurally disguised.

[0029] M0001 Missing Subject Module: This module belongs to the missing / fuzzy responsible subject category (M path), and is used to detect the missing responsible subject in legal texts.

[0030] M0024 Labor Relations Obscuration Module: This module belongs to the category of missing / ambiguous responsible parties (M path), and is used to identify situations where labor relations are obscured in legal texts, resulting in unclear responsible parties.

[0031] M0042 Pseudo-causal chain module: This module belongs to the category of missing / fuzzy responsible parties (M path), and is used to detect the existence of false causal chains in legal texts, which affects the accurate judgment of the responsible parties.

[0032] S001 Ambiguous Structure Module: This module belongs to the category of Systemic Language Failures (S-path) and is used to identify ambiguous language structures in legal texts that may lead to biased understanding of liability.

[0033] S002 Closed Loop Module: This module belongs to the category of systemic language failure (S path) and is used to detect closed loop structures appearing in legal texts. This structure may cause the determination of liability to fall into an infinite loop and make it impossible to reach a clear conclusion.

[0034] S003 Semantic Drift Module: This module belongs to the category of systematic language failure (S-path) and is used to identify situations in which semantics in legal texts drift during transmission, affecting the accurate understanding of liability information.

[0035] C0001 Language Format Masking Module: This module belongs to the structural legal disguise category (C path) and is used to detect situations in which specific language formats are used in legal texts to mask responsibilities, making the text appear legal on the surface but concealing the actual responsibilities.

[0036] C0004 Settlement Concealing Structure Module: This module belongs to the category of structural legal disguise (C path), and is used to identify situations in which liability is concealed through settlement-related structures in legal texts. It seems that a settlement has been reached, but in fact the liability has not been properly handled.

[0037] C0006 Multi-path Overflow Module: This module belongs to the category of structural legal disguise (C-path). It is used to detect situations in which there are multiple paths in the legal text and responsibility information overflows in these paths and is difficult to clearly define. It looks like a legal structure, but in fact the responsibility is unclear.

[0038] It should be noted that the above-mentioned M0001, M0024, M0042, S001, S002, S003, C0001, C0004, C0006, etc. represent the numbers of specific modules.

[0039] Among them, the modules can be compositely connected and the paths can be expanded.

[0040] For example, the concept of "composite linking between modules" means that within the memetic law structural attribution annotation system, modules under different paths do not exist in isolation but can be combined and used in conjunction with each other. For example, when processing complex legal texts, a single text segment may exhibit both ambiguity regarding the responsible party (involving modules under the M path) and semantic drift (involving modules under the S path). In this case, modules such as the M0001 Missing Subject Module, the M0024 Labor Relations Obscuration Module (from the M path), and the S003 Semantic Drift Module (from the S path) can be combined to jointly annotate and analyze the text segment, comprehensively and accurately identifying any attribution-related issues. This composite linking approach allows for flexible handling of a wide range of complex legal text scenarios, improving the accuracy and practicality of the annotation system. For example, a complete memetic chain can be formed to indicate the risk of a certain type of responsibility evasion.

[0041] For example, "path extensibility" means that the M-path, S-path, and C-path in the meme path annotation library are not fixed and unchangeable. As the legal field develops, new legal issues emerge, or research on attribution deepens, the system can add new modules under these paths based on actual needs. For example, if a new type of ambiguity in the responsible party is discovered in the future, a corresponding module can be added to the missing / ambiguous category of the responsible party in the M-path; or if a new systemic language failure occurs, a corresponding module can be added to the systemic language failure category of the S-path. This path extensibility enables the memetic law structural attribution annotation system to keep pace with the times, constantly adapt to new legal environments and research needs, and maintain its effectiveness and advancement.

[0042] Therefore, it should be noted that, according to actual needs, each path (M path, S path, C path) can also include other classes, and other classes can also include multiple other modules.

[0043] As an example, the meme path annotation library also includes: P path, which includes the structural damage identification class caused by power language, and the structural damage identification class caused by power language includes P0001 manipulative language module and P0002 humiliating language module.

[0044] For example, the P-path could include the identification of structural harm caused by the use of power language, which refers to a category or category specifically used to identify and assess the structural harm caused by the use of power language. This type of identification may involve the analysis of digitized corpus, the examination of the context of language use, and the assessment of the impact on victims, aiming to reveal how power language causes harm to individuals or groups through its structure and expression.

[0045] The P0001 Manipulative Language Module belongs to the category of structural damage identification caused by power language (P path). It is used to identify and analyze manipulative elements in digitized corpus, such as detecting whether there are elements in the language that attempt to manipulate others' thoughts, behaviors or emotions.

[0046] The P0002 humiliating language module belongs to the category of structural harm identification caused by power language (P path). It is used to identify and analyze humiliating elements in digitized corpus, such as detecting whether there are expressions in the language that belittle, insult, or harm the dignity of others.

[0047] As an example, each module can form a meme chain to point to a specific meme structure analysis (see the gym gray market operation script below).

[0048] The annotation engine is used to combine grammatical analysis, syntactic dependency trees and meme path matching models to automatically label potential attribution breakpoints in digitized corpus with meme numbers, which can be embedded in AI judgment systems and compliance processes.

[0049] For example, "AI adjudication system and compliance process" refers to the application scenario or usage environment in which the system of this application plays a role.

[0050] AI Judgment System: This system uses artificial intelligence technology to generate legal judgments (the core content of court judgments, including the determination of case facts, application of law, and the verdict outcome). Embedded within this system, the AI-powered system can annotate potential attribution breakpoints within the data-based corpus (such as case fact descriptions and legal text citations) used in judgment generation. This helps the AI ​​more accurately identify and address liability-related issues during judgment generation, improving the quality and accuracy of judgments.

[0051] Compliance Process: In the legal field, compliance process generally refers to the series of operations and review processes that companies or organizations undertake to comply with laws, regulations, and industry norms. Once embedded in this compliance process, this application system can identify potential liability breakpoints in relevant documents (such as contracts, agreements, and internal rules and regulations). This helps companies or organizations identify potential liability risks during document review, allowing them to make timely revisions and improvements to ensure compliance and avoid legal disputes arising from unclear responsibilities.

[0052] As an example, a meme path matching model can be trained using a relevant AI big data model (e.g., a GPT model). For example, by training with training data (including various digitized corpora), the corresponding relationships between different digitized corpora and the aforementioned modules (i.e., the meme path matching model) are obtained. It should be understood that different digitized corpora can form corresponding relationships with a single module or with multiple modules, and multiple modules may form a meme chain.

[0053] When new digitized corpus is input, the annotation engine is used to analyze the new digitized corpus by combining grammatical analysis, syntactic dependency tree and meme path matching model, and annotate it with meme numbers, that is, to determine which modules the new digitized corpus corresponds to, so as to output the annotated text and meme structure analysis report, and determine whether to issue an early warning based on the meme structure analysis report.

[0054] For example, the system outputs annotated text and a meme structure analysis report. The meme structure analysis report is then used to determine whether a warning should be issued. Specifically, the system matches the annotated modules to determine whether a complete meme chain exists. If so, a warning is issued for manual verification and confirmation. For example, a complete meme chain indicates a risk of responsibility absconding, and the meme chain path can be used as a warning to alert users or administrators.

[0055] In some embodiments of the present application, the memetic law structural attribution annotation system also includes an application interface module for integrating the system into government compliance regulatory sandboxes such as GovTech and PDPC as a plug-in; or integrating it into legal NLP models as a plug-in to enhance AI attribution capabilities; or integrating it into legal training systems as a plug-in to assist students and judges in structural language recognition training.

[0056] In some embodiments of the present application, the memetic structural attribution annotation system further includes a multi-language compatibility module for compatibility with multiple languages.

[0057] It should be noted that this system is designed with structure as the core and language as the shell. It has been tested in Chinese-English bilingual scenarios and is not affected by language changes. Structural masking features can be recognized across semantic spaces.

[0058] See also Figure 2 As shown, in some embodiments of the present application, a tagging method of the above-mentioned memetic structural attribution tagging system is also provided, which specifically includes the following steps: S10, input layer: import digitized corpus; for example, judgments or legal texts, both in Chinese and English.

[0059] S20, analysis layer: divide the digitized corpus into sentences and construct a syntactic tree to analyze the subject, predicate, and causal logical relationships.

[0060] S30, Matching Layer: The annotation engine uses the meme path annotation library to perform matching annotations. (If "settlement" is detected but no responsibility is attributed, the annotation is C0004.) Specifically, the annotation engine combines grammatical analysis, syntactic dependency trees, and meme path matching models to automatically annotate potential attribution breakpoints in the digitized corpus with meme numbers. For example, this can be embedded in AI adjudication systems and compliance processes.

[0061] S40, Output Layer: Outputs the annotated text and meme structure analysis report. The meme structure analysis report is used to determine whether a warning should be issued. For example, by matching annotated modules, it is determined whether a complete meme chain exists. If so, a warning is issued for manual verification and confirmation. For example, if a meme chain corresponding to a gym's gray market operation script exists, and if it demonstrates a significant "capability to continuously evade non-legal liability," a warning is issued.

[0062] In some embodiments of the present application, after step S40, the method further includes the following steps: S50, Extension Layer: Synchronous visualization of responsibility paths for educational demonstrations or system attribution scoring.

[0063] As a specific embodiment of the present application, if the meme structure analysis report output in step S40 includes the meme chain: M0011 → M0022 → M0031 → M0041 → M0036 → M0050, it can be determined that this is a "typical structural path of a gym gray market scenario." It should be understood that the M path in the above meme path annotation library also includes meme modules numbered M0011, M0022, M0031, M0041, M0036, M0050, etc.

[0064] The following is a detailed description of the [Example of Meme Chain Expansion] - the gym gray industry operation script.

[0065] Taking the "C-Fit" chain fitness platform in a certain place as an example, its core operator completed a round of gray market script transfer through the following chain: M0011-Evidence Recovery Subject Repair Meme Initially, C-Fit faced public backlash (such as unpaid employee wages and unexplained membership card freezes). The team didn't rush to respond, instead immediately collecting key documents like employee ID cards, membership contracts, and chat logs, in a "responsible information cleansing" exercise. This was ostensibly a "unified review," but in reality, it was a means of erasing the facts.

[0066] M0022-Hidden Threat Meme The company then began to engage in "informal communication and structural pressure" against internal employees, using messages like "You're not suited to stay in this industry for too long" and "Beware of legal letters if you talk nonsense to the media," creating a "leak of intimidation." Activists generally gave up their complaints, leading outside observers to mistakenly believe the issue was resolved.

[0067] M0031-Script Translation Meme Before the risks at headquarters became apparent, the operators "transferred" the same brand to another city (for example, "C-Fit Yuhang" was quickly deregistered, and "C-Fit Binjiang" was established 10 days later). The operating model, logo, and pricing structure remained the same, and the actual shareholders / legal representatives were rebranded, but the plot was repeated.

[0068] M0041-Script Meme Infection Structure The translated brand continues to use the same routine: "low-price prepayment → limited refunds → high employee turnover → delayed mediation → consumer withdrawal". The script runs like a highly toxic replicator, constantly infecting new cities, new members, and new victims.

[0069] M0036-Funding Path Translation Meme During this process, funds from the original platform were transferred from the new company to shell companies affiliated with the controlling shareholder through "training service fees" and "brand licensing." For example, "Hangzhou Libo Consulting" collected licensing fees, while "Hangzhou Yufeng Culture" collected brand promotion fees, effectively transferring funds.

[0070] M0050-Script Meme Migration Structure Meme Ultimately, when a city's regulator takes action, the operator has already completed the overall brand migration to the next target city (such as from Hangzhou to Suzhou), "registering a new platform, attracting new investors, and rebuilding the old story," forming a "national migratory gray industry replication chain."

[0071] Summary from the perspective of meme law: This chain shows the complete script structure of how a typical gray market gym "conceals → threatens → replicates → spreads → translates → migrates", and has a significant "ability to continuously evade non-legal responsibility", providing a standard template for the meme law to identify structural operations.

[0072] Therefore, when the meme structure attribution labeling system provided in the embodiment of the present application analyzes and identifies the above-mentioned meme chain, it can issue an early warning to the user (or manager) so that the user (or manager) can take corresponding measures in time to avoid causing greater damage.

[0073] In summary, the memetic structural attribution labeling system and labeling method provided in the embodiments of the present application can transform the concept of "lack of responsibility" into an observable structural path; use memetic labeling paths to track system responsible language behavior; create a cross-language modeling system with structural language masking; and construct an attribution diagnostic labeling system that is adapted to the AI ​​judgment engine.

[0074] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. The memetic structural attribution annotation system is characterized by: Includes meme path annotation library and annotation engine; The meme path annotation library includes M path, S path and C path; the M path includes the responsible subject missing / ambiguous class, which includes the subject missing module, labor relationship masking module, and pseudo causal chain module; the S path includes the systemic language failure class, which includes the ambiguous structure module, closed loop module, and semantic drift module; the C path includes the structure legal disguise class, which includes the language format masking module, reconciliation masking structure module, and multi-path overflow module; wherein, each module can be compositely connected and the path can be expanded; The annotation engine is used to combine grammatical analysis, syntactic dependency tree and meme path matching model to automatically annotate potential attribution breakpoints in the digitized corpus with meme numbers.

2. The memetic structural attribution annotation system according to claim 1 is characterized in that: It also includes an application interface module for integrating the system into the government compliance supervision sandbox as a plug-in; or integrating it into the legal NLP model as a plug-in to improve AI attribution capabilities; or integrating it into the legal training system as a plug-in to assist students and judges in conducting structured language recognition training.

3. The memetic structural attribution annotation system according to claim 1, characterized in that: It also includes a multi-language compatibility module for compatibility with multiple languages.

4. The memetic structural attribution annotation system according to claim 1, characterized in that: The meme path annotation library also includes: a P path, the P path includes a structural damage identification class caused by power language, and the structural damage identification class caused by power language includes: a manipulative language module and a humiliating language module.

5. The annotation method of the memetic structural attribution annotation system according to any one of claims 1 to 4, characterized in that: include: Import digitized corpus; Sentence the digitized corpus and construct a syntactic tree to analyze the subject, predicate, and causal logical relationships; Through the annotation engine, the meme path annotation library is called to perform matching annotation; Output the annotated text and meme structure analysis report, and determine whether to issue an early warning based on the meme structure analysis report.

6. The marking method according to claim 5, characterized in that: Outputting the annotated text and meme structure analysis report, and determining whether to issue an early warning based on the meme structure analysis report, includes: By matching the marked modules, it is determined whether there is a complete meme chain, and if so, an early warning is issued.

7. The marking method according to claim 5, characterized in that: After outputting the annotated text and the meme structure analysis report, and determining whether to issue an early warning based on the meme structure analysis report, the method further includes: Synchronously visualize responsibility paths for educational presentations or system attribution scoring.

Images