Document review method and legal document review method

Abstract

Embodiments of the present specification provide a document auditing method and a legal document auditing method, wherein the document auditing method comprises: obtaining a to-be-audited document; performing type identification on the to-be-audited document to obtain a document type of the to-be-audited document; determining a document auditing list of the to-be-audited document according to the document type, wherein the document auditing list comprises a plurality of auditing rules corresponding to the document type, and the plurality of auditing rules are different in auditing dimensions; and respectively utilizing the plurality of auditing rules to perform auditing on the to-be-audited document to obtain a document auditing result of the to-be-audited document. By adopting the idea of divide and conquer, the document auditing list is disassembled into a plurality of auditing rules different in auditing dimensions, precise document auditing for different auditing rules is realized, and when the document is audited by using each auditing rule, the complete to-be-audited document is audited, so that the auditing process can be flexibly analyzed according to the context of the to-be-audited document, and the accuracy of the document auditing is further improved.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of Internet technology, and in particular to document review methods and legal document review methods. Background Technology

[0002] With the development of computer technology, automated document review has become increasingly sophisticated. Document review refers to the process of checking and evaluating document content to ensure it meets specific standards, requirements, or regulations. Taking contract document review as an example, contract review is crucial for both businesses and individuals, serving as the primary line of defense in preventing legal risks and protecting legitimate rights.

[0003] Currently, automated document review can typically be performed based on preset regular expression rules. However, in the face of complex and changing document review environments, the above solutions cannot capture subtle or complex document issues, such as deep-seated legal conflicts, ambiguous wording or semantics, new legal regulations, etc., resulting in poor document review accuracy. Therefore, there is an urgent need for a document review solution with high accuracy. Summary of the Invention

[0004] In view of this, embodiments of this specification provide a document review method. One or more embodiments of this specification also relate to a legal document review method, a document review device, a legal document review apparatus, a computing device, a computer-readable storage medium, and a computer program product, to address the technical deficiencies existing in the prior art.

[0005] According to a first aspect of the embodiments of this specification, a document review method is provided, comprising:

[0006] Retrieve documents pending review;

[0007] Type identification is performed on the documents to be reviewed to obtain the document type of the documents to be reviewed;

[0008] Based on the document type, a document review list is determined for the documents to be reviewed. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0009] Multiple review rules are used to review the documents to be reviewed, and the review results of the documents to be reviewed are obtained.

[0010] According to a second aspect of the embodiments of this specification, a legal document review method is provided, comprising:

[0011] Obtain the legal documents pending review;

[0012] Type identification is performed on the legal documents to be reviewed to obtain the document type of the legal documents to be reviewed;

[0013] Based on the document type, a document review list for legal documents to be reviewed is determined. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0014] Multiple review rules are used to review the legal documents to be reviewed, and the review results of the legal documents to be reviewed are obtained.

[0015] According to a third aspect of the embodiments of this specification, a legal document review method is provided, including:

[0016] Receive legal documents pending review sent by the client;

[0017] Using the document review checklist of legal documents to be reviewed, the document review results are obtained and fed back to the client. The document review checklist is determined based on the document type of the legal documents to be reviewed. The document type is obtained by identifying the type of the legal documents to be reviewed. The document review checklist includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0018] According to a fourth aspect of the embodiments of this specification, a document review apparatus is provided, comprising:

[0019] The first acquisition module is configured to acquire documents pending review.

[0020] The first identification module is configured to identify the type of the document to be reviewed and obtain the document type of the document to be reviewed;

[0021] The first determination module is configured to determine a document review list for documents to be reviewed based on the document type. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0022] The first review module is configured to review the documents to be reviewed using multiple review rules and obtain the document review results.

[0023] According to a fifth aspect of the embodiments of this specification, a legal document review apparatus is provided, comprising:

[0024] The second acquisition module is configured to acquire legal documents to be reviewed.

[0025] The second identification module is configured to identify the type of the legal document to be reviewed and obtain the document type of the legal document to be reviewed;

[0026] The second determination module is configured to determine a document review list of legal documents to be reviewed based on the document type. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0027] The second review module is configured to review the legal documents to be reviewed using multiple review rules, and obtain the document review results of the legal documents to be reviewed.

[0028] According to a sixth aspect of the embodiments of this specification, a legal document review apparatus is provided, comprising:

[0029] The receiving module is configured to receive pending legal documents sent by the client;

[0030] The third review module is configured to review the legal documents to be reviewed using a document review checklist, obtain the document review results, and feed the results back to the client. The document review checklist is determined based on the document type of the legal documents to be reviewed. The document type is obtained by identifying the type of the legal documents to be reviewed. The document review checklist includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0031] According to a seventh aspect of the embodiments of this specification, a computing device is provided, comprising:

[0032] Memory and processor;

[0033] The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the methods provided in the first, second, or third aspects described above.

[0034] According to an eighth aspect of the embodiments of this specification, an electronic device is provided, comprising:

[0035] The memory and processor are connected via a bus;

[0036] The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the methods provided in the first, second, or third aspects described above.

[0037] According to a ninth aspect of the embodiments of this specification, a computer-readable storage medium is provided that stores a computer program / instructions that, when executed by a processor, implement the steps of the methods provided in the first, second, or third aspects described above.

[0038] According to a tenth aspect of the embodiments of this specification, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the steps of the methods provided in the first, second, or third aspects described above.

[0039] This specification provides a document review method according to one embodiment, comprising: acquiring a document to be reviewed; identifying the document type of the document to be reviewed; determining a document review list for the document to be reviewed based on the document type, wherein the document review list includes multiple review rules corresponding to the document type, and the multiple review rules have different review dimensions; and reviewing the document to be reviewed using the multiple review rules respectively to obtain the document review result. By adopting a divide-and-conquer approach to decompose the document review list into multiple review rules with different review dimensions, precise document review for different review rules is achieved. Furthermore, when reviewing documents using each review rule, the document to be reviewed includes the complete context, allowing the review process to flexibly analyze the context content of the document to be reviewed, further improving the accuracy of document review. Attached Figure Description

[0040] Figure 1 This is an architecture diagram of a document review system provided in one embodiment of this specification;

[0041] Figure 2 This is a flowchart illustrating a document review method provided in one embodiment of this specification;

[0042] Figure 3 This is a flowchart illustrating a legal document review method provided in one embodiment of this specification;

[0043] Figure 4 This is a flowchart illustrating another legal document review method provided in one embodiment of this specification;

[0044] Figure 5 This is a flowchart illustrating the processing procedure of a legal document review method provided in one embodiment of this specification;

[0045] Figure 6 This is a schematic diagram of the structure of a document review device provided in one embodiment of this specification;

[0046] Figure 7 This is a schematic diagram of the structure of a legal document review device provided in one embodiment of this specification;

[0047] Figure 8 This is a schematic diagram of another legal document review device provided in one embodiment of this specification;

[0048] Figure 9This is a structural block diagram of a computing device provided in one embodiment of this specification;

[0049] Figure 10 This is a structural block diagram of an electronic device provided in one embodiment of this specification. Detailed Implementation

[0050] Many specific details are set forth in the following description to provide a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0051] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of this specification. The singular forms “a,” “described,” and “the” as used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0052] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this specification, and similarly, second may also be referred to as first. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0053] Furthermore, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0054] In one or more embodiments of this specification, a large model refers to a deep learning model with a large number of model parameters, typically containing hundreds of millions, tens of billions, hundreds of billions, trillions, or even tens of trillions of model parameters. A large model can also be called a foundational model (Foundation Model 1). It is pre-trained using large-scale unlabeled corpora to produce a pre-trained model with hundreds of millions of parameters. Such models can adapt to a wide range of downstream tasks and have good generalization ability. Examples include Large Language Models (LLMs) and Multi-modal Pre-training Models (MLMs).

[0055] In practical applications, large models only require a small number of samples to fine-tune the pre-trained model before they can be applied to different tasks. Large models can be widely used in fields such as Natural Language Processing (NLP) and Computer Vision. Specifically, they can be applied to computer vision tasks such as Visual Question Answering (VQA), Image Captioning (IC), and Image Generation, as well as NLP tasks such as text-based sentiment classification, text summarization, and machine translation. The main application scenarios for large models include digital assistants, intelligent robots, search, online education, office software, e-commerce, and intelligent design.

[0056] First, the terms and concepts used in one or more embodiments of this specification will be explained.

[0057] Bidirectional Encoder Representations from Transformers (BERT) is a pre-trained NLP model. By learning from large amounts of unlabeled text data, this model can capture deep semantic information in text and achieves significant performance improvements on numerous NLP tasks.

[0058] The Text-to-Text Transfer Transformer (T5) is a deep self-attention model for text-to-text communication. A key feature of the T5 model is that it unifies all NLP tasks into a text-to-text format, meaning both input and output are text sequences. This design allows the model to more easily adapt to various tasks, such as translation, question answering, and summarization.

[0059] Large Language Model: Refers to deep learning models with a large number of parameters, especially in the field of NLP. These models can understand and generate natural language by learning a vast amount of text data, demonstrating strong context understanding, generation, and reasoning capabilities. Large language models include, but are not limited to, BERT, T5, etc. Due to their scale and complexity, large language models can often better capture the nuances and complex structures of language.

[0060] Prompt: In the field of NLP, especially in the application of large models, Prompt refers to a piece of text used to guide the model to generate a specific type of response. Simply put, it provides the model with an initial input or problem scenario to prompt and stimulate the model to produce the desired output. For example, in the contract review task, the Prompt may be an excerpt of a contract text plus clear instructions, such as "Please check whether the legal terms in this contract are compliant."

[0061] Agent: Refers to an entity that can autonomously perceive the environment, make decisions, and take actions to achieve goals. It can be a software program, a robot, or other forms of automated systems. The core characteristics of an Agent include autonomy, reactivity, social ability, and learning or adaptation ability.

[0062] Contract Review: Refers to the process of carefully examining a contract document to ensure its legal validity and commercial reasonableness. This includes verifying the clarity, compliance, logical consistency of the terms, as well as risk assessment, ensuring that the contract complies with relevant laws and regulations, and protecting the rights and interests of both parties.

[0063] Problem of token exceeding the length limit: Token refers to the basic unit for an artificial intelligence model to process text. For example, "hamburger" is divided into three tokens: "ham", "bur", and "ger", while "pear" is one token. One token is approximately equivalent to 4 characters or 0.75 English words. Moreover, Chinese uses more tokens compared to English. For example, "你好,世界" can be divided into 11 tokens, while "He l lo Wor ld" has only two tokens. Taking the contract review task as an example, in the contract review task, lawyers have a large number of detailed regulations for each type of review. For a single contract, the full - scale review usually exceeds 8k tokens. Usually, an input Prompt contains: the original contract text, review regulations, and specific review task instructions. Currently, in common open - source models on the market, the token limit for the input of large models is around 3w tokens, which cannot meet the complete review of the full - scale rules of long - text contracts over 20,000 words.

[0064] Taking contract review as an example, it is crucial for both businesses and individuals, serving as the primary line of defense against legal risks and the protection of legitimate rights. For businesses, rigorous contract review ensures transaction security, avoids economic losses and reputational damage caused by unfavorable or illegal terms, and maintains stable development and market competitiveness. For individuals, contract review prevents falling into unfair agreements, protects personal property and privacy, ensures that responsibilities and rights are clearly and equitably matched, and avoids potential future legal disputes and personal losses.

[0065] In traditional work models, contract review processes typically encounter the following difficulties: Highly specialized: Contract review involves the comprehensive application of legal knowledge, industry regulations, and practical experience. It requires lawyers not only to possess solid legal foundations but also to continuously accumulate experience in handling industry characteristics and common issues. Therefore, while newly hired legal assistants may quickly grasp the basics, achieving efficient and accurate contract review requires considerable time and practical experience. Significant time and energy expenditure: For law firms, the large number of contracts is commonplace due to a broad client base and diverse project types. During contract review, each contract requires careful consideration of every clause to ensure the identification and correction of potential problems, guaranteeing the contract's rigor and legality. Furthermore, the complexity and variability of contract terms necessitate adherence to constantly updated laws and regulations. Therefore, contract review is not only a technical task but also a test of a professional's patience and meticulousness, typically requiring substantial time and energy.

[0066] Rule-based analysis is an early method for automated contract review. This method automates review based on pre-defined regular expression rules, improving efficiency in some rule-based reviews. However, it lacks flexibility in the face of complex and changing legal environments. The main steps of this method are as follows: Rule Definition: Legal experts define a set of rules based on experience and legal knowledge. These rules may include keyword and phrase identification, existence checks for specific clauses, and standardization checks of contract structure and content. For example, common rules include "all contracts must include confidentiality clauses" or "the word 'permanent' cannot appear in the contract." Data Input and Preprocessing: Contract documents are input into the system and, after optical character recognition conversion and preprocessing (such as cleaning, formatting, and standardization), are transformed into structured data that the system can understand. Rule Matching: The rule engine matches the preprocessed contract content against the predefined rules. Through text search, pattern matching, clause comparison, and other technologies, the rule engine analyzes each clause in the contract, checking for rule violations or discrepancies with existing rules. Problem Identification and Warnings: Once the analysis engine finds a mismatch between the contract content and the rules, it issues a warning. The warning may relate to a specific clause or the entire content or structure of the contract. While rule-based analysis is highly useful for examining standard clauses and expected content, it cannot always capture more subtle or complex legal issues, such as deep-seated legal conflicts, ambiguous wording or semantics, or new legal provisions. It suffers from a lack of flexibility and adaptability, resulting in poor accuracy in contract review results.

[0067] To address the aforementioned issues, this specification proposes a document review method. This method breaks down complex contract review requirements into task decomposition, forming a document review list containing multiple review rules with different review dimensions. Then, the document to be reviewed is reviewed in parallel using these multiple review rules, improving document review performance while supporting more comprehensive and accurate document review. Specifically, the method involves: acquiring the document to be reviewed; identifying the document type; determining the document review list based on the document type, where the review list includes multiple review rules corresponding to the document type, with each rule having different review dimensions; and reviewing the document using these multiple review rules to obtain the document review result.

[0068] This specification provides a document review method, and also relates to a legal document review method, a document review device, a legal document review device, a computing device, a computer-readable storage medium, and a computer program product, which are described in detail in the following embodiments.

[0069] See Figure 1 , Figure 1This specification illustrates an architecture diagram of a document review system according to one embodiment of the present specification. The document review system may include a client 100 and a server 200.

[0070] Client 100 is used to send documents to be reviewed to server 200;

[0071] Server 200 is used to identify the type of the document to be reviewed and obtain the document type of the document to be reviewed; based on the document type, it determines the document review list of the document to be reviewed, wherein the document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different; the document to be reviewed is reviewed using the multiple review rules respectively to obtain the document review result of the document to be reviewed; and the document review result is sent to client 100.

[0072] Client 100 is also used to receive document review results sent by server 200.

[0073] The solution implemented in this specification breaks down the document review checklist into multiple review rules with different review dimensions using a divide-and-conquer approach. This achieves precise document review for different review rules. Furthermore, when reviewing documents using each review rule, the complete document to be reviewed is reviewed, allowing the review process to be flexibly analyzed based on the context of the document to be reviewed, further improving the accuracy of document review.

[0074] In practical applications, a document review system may include multiple clients 100 and a server 200. The multiple clients 100 can establish communication connections through the server 200. In the document review scenario, the server 200 provides document review services among the multiple clients 100. Each client 100 can act as a sender or receiver, communicating through the server 200. Users can interact with the server 200 through client 100 to receive data sent by other clients 100, or send data to other clients 100, etc. In the document review scenario, a user can publish a data stream to the server 200 through client 100. The server 200 generates document review results based on this data stream and pushes the document review results to other clients with established communication. See also Figure 1 The client 100 can also interact with the user through a graphical user interface to invoke the document review service, thereby implementing the document review method provided in the embodiments of this specification. The client 100 and the server 200 establish a connection via a network. The network provides a communication link medium between the client 100 and the server 200. The network can include various connection types, such as wired or wireless communication links or fiber optic cables. The data transmitted by the client 100 may need to undergo encoding, transcoding, compression, or other processing before being published to the server 200.

[0075] Client 100 can be a browser, application (APP), web application such as HyperText Markup Language 5 (H5), lightweight application (also known as a mini-program), or cloud application, etc. Client 100 can be developed based on the software development kit (SDK) of the corresponding service provided by server 200, such as based on the Real-Time Communication (RTC) SDK. Client 100 can be deployed in electronic devices and depends on the device to run or some APP on the device to run. Electronic devices may have a display screen and support information browsing, such as personal mobile terminals such as mobile phones, tablets, and personal computers. Various other types of applications can also be configured in electronic devices, such as human-computer interaction applications, model training applications, text processing applications, web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0076] Server 200 may include servers providing various services, such as servers providing communication services to multiple clients, servers supporting backend training of models used on clients, and servers processing data sent by clients. It should be noted that server 200 can be implemented as a distributed server cluster composed of multiple servers, or as a single server. The server can also be a server in a distributed system, or a server integrated with blockchain. The server can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms, or an intelligent cloud computing server or intelligent cloud host with artificial intelligence technology.

[0077] It is worth noting that the document review method provided in the embodiments of this specification is generally executed by the server. However, in other embodiments of this specification, the client may also have similar functions to the server, thereby executing the document review method provided in the embodiments of this specification. In other embodiments, the document review method provided in the embodiments of this specification may also be executed jointly by the client and the server. Next, taking the execution of the document review method provided in the embodiments of this specification by the server as an example, the document review method will be described in detail.

[0078] See Figure 2 , Figure 2 This specification illustrates a flowchart of a document review method according to an embodiment, which specifically includes the following steps:

[0079] Step 202: Obtain the document to be reviewed.

[0080] It should be noted that documents awaiting review refer to files or materials that require a review process. This review process may include checking the accuracy, compliance, and completeness of the document's content to ensure that the document meets relevant standards or regulations. Documents awaiting review can be documents from various scenarios, such as internal corporate process documents, academic papers, legal documents, etc. Documents awaiting review can be in different document formats, such as Portable Document Format (PDF), DOCX, etc. Documents awaiting review can also be in different languages, such as Chinese documents, English documents, etc.

[0081] In practical applications, there are various ways to obtain documents awaiting review, and the specific method should be selected according to the actual situation. This specification does not impose any limitations on these methods in its embodiments. In one possible implementation, documents awaiting review can be received from a user via a client. In another possible implementation, documents awaiting review can be read from other databases or data acquisition devices.

[0082] Step 204: Perform type identification on the document to be reviewed to obtain the document type of the document to be reviewed.

[0083] In one or more embodiments of this specification, after obtaining the document to be reviewed, it is further necessary to obtain a document review list for the document to be reviewed. To ensure that the document review list can accurately review the document to be reviewed, the document type of the document to be reviewed can be obtained first, and then the document review list for the document to be reviewed can be determined based on the document type of the document to be reviewed.

[0084] It should be noted that type identification refers to the process of determining the category to which a document to be reviewed belongs by analyzing its content, structure, format, and other characteristics. Document type refers to classifying a document into a specific category based on its content, structure, format, and other characteristics. Document types include, but are not limited to, technical documents, legal documents, and academic papers. Taking legal documents as an example, legal documents can be further subdivided into various types, such as contracts, agreements, powers of attorney, etc.

[0085] In practical applications, there are various ways to identify the document type of the document to be reviewed, and the specific method chosen depends on the actual situation. This specification does not impose any limitations on these methods in its embodiments. One possible implementation of this specification is to directly utilize a type identification model to identify the document type of the document to be reviewed.

[0086] In another possible implementation of this specification, the type recognition model can learn from large-scale training data across multiple domains, acquiring expertise and language habits specific to those domains. The knowledge and experience of experts can then guide the model in a targeted manner to improve its performance in particular domains. Therefore, type recognition of the document to be reviewed can be performed based on multiple candidate document types to obtain the document type of the document to be reviewed. That is, the above-mentioned type recognition of the document to be reviewed to obtain the document type can include the following steps:

[0087] Retrieve multiple candidate document types;

[0088] The document type of the document to be reviewed is obtained by identifying multiple candidate document types and input type of the document to be reviewed.

[0089] It should be noted that the multiple candidate document types can be document types summarized in advance based on prior knowledge. The type recognition model can be a large model, a deep learning model trained on multiple sample documents and the sample document types of each sample document, or a deep learning model trained on multiple candidate sample documents, multiple sample documents, and the sample document types of each sample document.

[0090] In practical applications, there are various ways to obtain multiple candidate document types, and the specific method should be selected according to the actual situation. This specification does not impose any limitations on this method in its embodiments. In one possible implementation of this specification, multiple candidate document types can be received from a user via a client. In another possible implementation of this specification, multiple candidate document types can be read from other data acquisition devices or databases.

[0091] Furthermore, when inputting multiple candidate document types and the document to be reviewed into the type recognition model, one possible implementation is to directly input the candidate document types and the document to be reviewed into the type recognition model. Another possible implementation is to add multiple candidate document types to the type recognition prompt template to obtain type recognition prompt information, and then input the type recognition prompt information and the document to be reviewed into the type recognition model to obtain the document type of the document to be reviewed.

[0092] The solution implemented in this specification improves the efficiency and accuracy of document type determination by utilizing a type recognition model to identify the document type of the document to be reviewed from multiple candidate document types summarized based on prior knowledge.

[0093] Step 206: Based on the document type, determine the document review list for the document to be reviewed. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0094] It should be noted that a document review checklist refers to a collection of document review rules or standards. The checklist may also include key considerations for each rule. Review dimensions refer to different aspects or standards to be considered during the review process. Each dimension represents a specific review perspective. By using multiple review rules across different dimensions, the quality, compliance, and applicability of a document can be comprehensively evaluated, ensuring that the document under review meets all review requirements. Review dimensions include, but are not limited to, content accuracy, compliance, completeness, readability, and format consistency, etc., and are selected based on the actual situation; this specification does not impose any limitations on these aspects. Review rules describe the broad logic of document review, allowing for dynamic analysis and identification of the document's context during the review process, improving the flexibility of rule-based review. The document review checklist can be manually compiled by experts or generated semi-automatically using a combination of expert knowledge and generative models to match the document type.

[0095] In practical applications, there are multiple ways to determine the document review list of documents to be reviewed based on document type. The specific method should be selected according to the actual situation. This specification does not limit the specific methods used in this embodiment.

[0096] In one optional embodiment of this specification, a document review list of documents to be reviewed can be selected from a pre-generated list of multiple candidate review lists based on document type. That is, the above-mentioned determination of the document review list of documents to be reviewed based on document type may include the following steps:

[0097] Retrieve multiple candidate review lists and candidate document types for the multiple candidate review lists;

[0098] The document types are matched with the candidate document types in multiple candidate review lists to obtain multiple matching results, where each matching result corresponds one-to-one with a candidate review list.

[0099] Based on multiple matching results, a document review list is selected from multiple candidate review lists to identify documents to be reviewed.

[0100] It should be noted that multiple candidate review lists refer to pre-generated review lists. The candidate document types in each candidate review list can be the same or different. The candidate review lists can be generated manually by experts or semi-automatically through a combination of expert knowledge and a generative model. The matching result describes the degree of match between the document type of the document to be reviewed and the candidate document types in any candidate review list. The matching result can be a specific matching degree, such as 80%, or a matching level, such as "very good match." The document review list of the document to be reviewed refers to the candidate review list with the highest degree of matching between the candidate document types from among the multiple candidate review lists.

[0101] In practical applications, there are multiple ways to match document types with candidate document types in each candidate review list. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on this approach. In one possible implementation, a first similarity between the document type and the candidate document type can be calculated, and the corresponding matching result is determined based on the first similarity. In another possible implementation, features can be extracted from both the document type and the candidate document type, and a second similarity between the document type features and the candidate document type features can be calculated. The corresponding matching result is then determined based on the second similarity. The similarity calculation methods include, but are not limited to, cosine similarity calculation and Euclidean distance calculation.

[0102] The solution implemented in this specification ensures the accuracy of the document review list by selecting a document review list from multiple candidate documents based on the document type of the document to be reviewed and the candidate document types of multiple candidate documents.

[0103] In another possible implementation of this specification, manually compiling a list of all possible candidate documents for review by experts is a significant challenge. Furthermore, the document review checklist is quite flexible, depending on the characteristics of different documents. For example, the key points of review for sales contracts and rental contracts are completely different. For sales contracts, different subject matters have different characteristics, directly affecting the focus of the review. Therefore, review dimensions can be determined based on expert knowledge, and a rule-based model can be used to generate a corresponding document review checklist based on these dimensions and the documents to be reviewed. By integrating expert knowledge and model capabilities, a document review checklist that considers the characteristics of different documents can be generated, ultimately improving its cross-domain adaptability. That is, the above-mentioned document review checklist, determined based on document type, can include the following steps:

[0104] Obtain multiple review dimensions corresponding to the document type;

[0105] The model is generated by inputting multiple review dimensions and documents to be reviewed into rules, resulting in multiple review rules, where each review rule corresponds one-to-one with a review dimension.

[0106] Based on multiple review rules, construct a document review list of documents to be reviewed.

[0107] It should be noted that the rule generation model can be a large model, or it can be a deep learning model trained based on sample review rules, sample review documents, and multiple sample review dimensions corresponding to the sample review rules. Multiple review dimensions can constitute the review framework of the document review checklist.

[0108] In practical applications, there are various ways to obtain multiple review dimensions corresponding to a document type, and the specific method should be selected according to the actual situation. This specification does not impose any limitations on these methods in its embodiments. In one possible implementation, multiple review dimensions corresponding to a document type can be received from a user via a client, wherein these multiple review dimensions are derived based on prior knowledge. In another possible implementation, multiple review dimensions corresponding to a document type can be read from other data acquisition devices or databases.

[0109] Furthermore, there are multiple ways to construct a document review list for documents to be reviewed based on multiple review rules. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on this approach. In one possible implementation, multiple review rules can be directly concatenated to obtain the document review list for the documents to be reviewed. In another possible implementation, multiple review rules can be deduplicated, and the deduplicated review rules can be concatenated to obtain the document review list for the documents to be reviewed.

[0110] The solution implemented in this specification automatically generates a document review checklist based on a manually defined review framework, improving the flexibility and accuracy of the document review checklist.

[0111] In one optional embodiment of this specification, in addition to determining the document review list of documents to be reviewed based on document type, a custom review list sent by the user based on actual needs can also be received. The determined document review list and the custom review list are combined as the list of documents to be reviewed, thereby improving the user experience.

[0112] Step 208: Use multiple review rules to review the document to be reviewed and obtain the document review results.

[0113] It should be noted that the document review results describe whether the document under review complies with each review rule. The document review results may also include risk-based modification suggestions for cases where the document does not comply with the review rules.

[0114] In practical applications, when reviewing documents using multiple review rules, these rules can be applied sequentially. To improve efficiency, multiple review rules can be applied in parallel to obtain the final review result. Furthermore, when reviewing a document, a rule engine can be invoked to perform the review based on multiple rules, or a document review model can be invoked to apply multiple rules separately to obtain the final review result. For example, a document review model can be invoked in parallel to review the document based on multiple rules, yielding the final review result.

[0115] In one optional embodiment of this specification, to improve the accuracy of document review, before reviewing the document to be reviewed using multiple review rules and obtaining the document review result, key document information can be extracted from the document to be reviewed. Then, the document to be reviewed using multiple review rules and the key document information to obtain the document review result. Key document information includes, for example, the document publication date, keywords in the document, and the document author.

[0116] In practical applications, document review results include risk detection results and risk modification suggestions. After reviewing the document using multiple review rules and obtaining the document review results, the following steps can be included: In the first possible implementation of this specification, if the risk detection result indicates that the document to be reviewed has a risk, risk modification suggestions can be directly sent to the client so that the user can modify the document based on the risk modification suggestions. In the second possible implementation of this specification, if the risk detection result indicates that the document to be reviewed has a risk, the risk modification suggestions can be verified, and if the verification result is successful, the risk modification suggestions can be sent to the client. In the third possible implementation of this specification, if the risk detection result indicates that the document to be reviewed has a risk, the document to be reviewed can be directly modified according to the risk modification suggestions, and the modified document to be reviewed can be sent to the client. In the fourth possible implementation of this specification, if the risk detection result indicates that the document to be reviewed has a risk, the risk modification suggestions can be verified, and if the verification result is successful, the document to be reviewed can be modified according to the risk modification suggestions, and the modified document to be reviewed can be sent to the client.

[0117] The solution implemented in this specification breaks down the document review checklist into multiple review rules with different review dimensions using a divide-and-conquer approach. This achieves precise document review for different review rules. Furthermore, when reviewing documents using each review rule, the complete document to be reviewed is reviewed, allowing the review process to be flexibly analyzed based on the context of the document to be reviewed, further improving the accuracy of document review.

[0118] In one optional embodiment of this specification, the above-described method of using multiple review rules to review the document to be reviewed and obtain the document review result may include the following steps:

[0119] Input the review rules and the document to be reviewed into the document review model to obtain the review results of the document to be reviewed under the review rules;

[0120] The document review result is determined based on the review results of the document under multiple review rules.

[0121] It's important to note that the document review model can be a machine learning model trained based on sample review rules, sample review documents, and sample review results. Taking contracts as an example, the training method and document review process based on the model are explained as follows: Data Preparation and Preprocessing: The contract text undergoes cleaning and preprocessing steps such as format standardization, noise reduction, and extraction of important features to ensure the machine learning model can understand and learn from it. Building the Machine Learning Model: A suitable machine learning algorithm is selected and trained, such as Support Vector Machines (SVM), Naive Bayes, or neural networks. Model Training: The machine learning model learns from the contract text as a sample set. This process may require multiple iterations, and the model is adjusted to optimize prediction results. Deployment and Implementation: The validated and optimized model can be used in actual contract review. Document review models trained based on machine learning models can understand and judge more complex contract content to a certain extent by automatically learning from a large number of contracts. However, due to its reliance on large amounts of labeled data specific to a particular scenario and the relatively slow training and optimization speed, it cannot adapt and process novel, unlabeled contract scenarios as quickly as larger models. Therefore, document review models can also be large-scale models. Large-scale models provide a more comprehensive, accurate, and efficient solution for document review. They can process large numbers of documents simultaneously, analyze complex review rules, and understand and infer based on past training and experience. This allows for a comprehensive document review to be completed within seconds, significantly saving review time and ensuring accuracy.

[0122] In one optional embodiment of this specification, document review projects typically require outputting comprehensive document review reports for long texts. A complete report usually involves outputting over 10,000 tokens, which can lead to excessively long review times and low performance. To address this issue, this embodiment proposes a method based on parallel invocation of a document review model using multiple review rules, resolving the performance problem caused by excessive output characters. For review requirements of the same magnitude, using a parallel review approach with multiple review rules can improve review performance by more than three times.

[0123] In practical applications, there are multiple ways to determine the document review result of a document under review based on the review results under various review rules. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on this approach. One possible implementation of this specification is to directly concatenate the review results obtained from reviewing the document under each review rule to obtain the document review result. Another possible implementation of this specification is to utilize a document review model to integrate the review results corresponding to each review rule, resulting in a more readable document review result.

[0124] The solution implemented in this specification utilizes a document review model to review documents based on individual review rules. This allows the document review model to perform a more detailed analysis and understanding of each rule, helping to generate accurate single-rule review conclusions. Finally, a more accurate document review result is generated through a summarizing process. Furthermore, by adopting a divide-and-conquer approach, calling the document review model to review documents based on individual rules allows the document review list to be broken down into sections within the acceptable character length for the model, enabling each model call to perform individual review for each rule. This overcomes the issue of token length exceeding limits when calling the model alone, helping the document review agent perform more refined and comprehensive reviews.

[0125] In one optional embodiment of this specification, determining the document review result of the document to be reviewed based on the review results under multiple review rules may include the following steps:

[0126] The review results of the document to be reviewed under multiple review rules are input into the document review model for integration, and the document review result of the document to be reviewed is obtained.

[0127] For example, suppose that review result 1 obtained by reviewing the document under review using review rule 1 is "Reviewed, the document under review is valid," and review result 2 obtained by reviewing the document under review using review rule 2 is "Reviewed, the content of the document under review is accurate." If review result 1 and review result 2 are directly concatenated, the resulting document review result is "Reviewed, the document under review is valid; reviewed, the content of the document under review is accurate." It is evident that the document review result has poor readability. Therefore, review result 1 and review result 2 can be input into the document review model, and the semantic understanding capability of the document review model can be used to integrate the results to obtain the document review result for the document under review: "Reviewed, the document under review is valid and the content is accurate."

[0128] The solution implemented in this specification utilizes the semantic understanding capability of the document review model to integrate multiple single-rule review conclusions into a document review result, thereby improving the readability of the document review result.

[0129] In one optional embodiment of this specification, determining the document review result of the document to be reviewed based on the review results under multiple review rules may include the following steps:

[0130] Get the priority of the review results of the document to be reviewed under multiple review rules;

[0131] Based on the priority of the results, determine the document review result of the document to be reviewed.

[0132] It should be noted that the result priority of the audit results refers to the importance level assigned to the results obtained from auditing the documents under different audit rules. By prioritizing the results, it can be ensured that the most critical issues are addressed promptly. There are multiple ways to obtain the result priority of the audit results; the specific method should be selected based on the actual situation, and this specification does not limit this approach. In one possible implementation of this specification, the rule priorities pre-configured for each audit rule can be obtained, and the rule priorities can be determined as the result priorities of the corresponding audit results. In another possible implementation of this specification, the audit results include the risk detection results under the current audit rules, and the result priority of the audit results can be determined based on the risk detection results. For example, if risk detection result 1 corresponding to audit rule 1 is high risk, and risk detection result 2 corresponding to audit rule 2 is low risk, then the result priority 1 of audit result 1 is higher than the result priority 2 of audit result 2.

[0133] In practical applications, there are multiple ways to determine the document review result of a document based on result priority. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on this approach. In one possible implementation, review results with higher priority can be highlighted or displayed at the top of the document review results. In another possible implementation, review results with a priority lower than a preset priority can be deleted, and the document review result of the document to be reviewed can be determined based on review results with a priority not lower than the preset priority.

[0134] By applying the solutions in the embodiments of this specification, the document review result of the document to be reviewed is determined according to the priority of the review results obtained from the review of the document to be reviewed under each review rule, thereby improving the flexibility of the document review result.

[0135] In one optional embodiment of this specification, before inputting the audit rules and the document to be audited into the document audit model to obtain the audit result of the document to be audited under the audit rules, the following steps may also be included:

[0136] Obtain audit prompt information, wherein the audit prompt information is used to guide the document audit model to audit the document based on the target audit task, the target audit task includes at least one of the following: risk detection task, risk factor determination task, risk type identification task, risk content search task, and risk content modification task;

[0137] Inputting the review rules and the document to be reviewed into the document review model to obtain the review results of the document under the review rules can include the following steps:

[0138] Input the audit prompts, audit rules, and the document to be audited into the document audit model to obtain the audit results of the document under the audit rules.

[0139] It's important to note that document review not only involves detailed reasoning across numerous rules but also allows for targeted review of specific tasks, achieving precise document review. Targeted review tasks can be single or multiple. The risk detection task checks for risks in the document to be reviewed that do not comply with review rules. The risk factor identification task identifies specific risk factors present in the document, such as the lack of data protection clauses in a contract. The risk type identification task identifies the specific type of risk (e.g., high, medium, low) in the document. The risk content search task locates the specific paragraphs in the document where risks exist. The risk content modification task adds appropriate content to the document to address the risks.

[0140] By applying the solution proposed in the embodiments of this specification, specific target review tasks are added to the review prompt information. The review prompt information guides the review process of the document review model, enabling the document review model to execute and integrate multiple tasks to obtain information-rich document review results.

[0141] In one optional embodiment of this specification, after reviewing the document to be reviewed using multiple review rules and obtaining the document review result, the following steps may be further included:

[0142] Send the document review results to the client;

[0143] Receive rule update information sent by the client based on the document review results, wherein the rule update information includes at least one of rule deletion information, rule modification information, and rule addition information;

[0144] Update the document review list of documents pending review based on the rule update information, and obtain the updated document review list.

[0145] It's important to note that rule update information refers to notifications sent by the client regarding changes to the document review checklist. These notifications can include the deletion, modification, or addition of rules within the checklist. The purpose of sending rule update information is to ensure that the document review checklist always complies with the client's latest review standards and requirements. Rule update information typically includes the identifier of the review rule to be updated and may also include the reason for the update. Rule deletion information refers to notifications sent by the client regarding the deletion of an existing review rule from the document review checklist. Rule modification information refers to notifications sent by the client regarding the modification of an existing review rule from the document review checklist. Rule addition information refers to notifications sent by the client regarding the addition of a new review rule to the document review checklist. Rule addition information may include detailed information about the new rule and the reason for its addition.

[0146] The solution implemented in this specification sends document review results to the client, allowing users to filter or adjust existing document review lists to create lists that better meet their review needs. Users can also customize review rules to meet their specific requirements, resulting in more comprehensive and accurate document review lists and enabling iterative updates. Furthermore, during these iterative updates, the document review model can perform highly adaptive reviews based on the new lists, improving the accuracy of document reviews.

[0147] In one optional embodiment of this specification, during the document review process, various review clauses are usually complex and ever-changing, and need to comply with constantly updated laws, regulations, and related data. Therefore, before reviewing the document to be reviewed, it can be determined whether the document review list applicable to the document to be reviewed meets the rule supplementation conditions. If so, the document review list is supplemented to achieve real-time updating of the document review list. That is, before reviewing the document to be reviewed using multiple review rules and obtaining the document review result of the document to be reviewed, the following steps may also be included:

[0148] If the document review checklist meets the supplementary rule conditions, obtain supplementary rule information for the document review checklist;

[0149] The process of reviewing a document using multiple review rules to obtain the review results may include the following steps:

[0150] The document to be reviewed is reviewed using multiple review rules and supplementary information to obtain the document review results.

[0151] It's important to note that the supplementary rule conditions are used to determine whether the document review checklist is relevant to continuously updated rules or information. If the document review checklist meets the supplementary rule conditions, it means the checklist is relevant to the continuously updated rules or information, and needs to be supplemented using the updated supplementary rule information. If the document review checklist does not meet the supplementary rule conditions, it means the document review is unrelated to the continuously updated rules or information, and does not need to be supplemented using the updated supplementary rule information; the document can be reviewed accurately. Supplementary rule information refers to information that supplements and refines the rules of the document review checklist, such as new legal provisions and current company qualifications.

[0152] In practical applications, when the document review checklist meets the supplementary rule conditions, there are multiple ways to obtain supplementary rule information for the document review checklist. The specific method should be selected based on the actual situation, and this specification does not limit this approach. One possible implementation of this specification is to receive supplementary rule information for the document review checklist sent by the user through a client. Another possible implementation of this specification is to call an external interface to obtain supplementary rule information for the document review checklist.

[0153] The solution implemented in this specification obtains supplementary rule information for the document review checklist and uses this continuously updated supplementary rule information as background information during the review process. This guides the document review model to reference real-time knowledge and generate more timely review results. Furthermore, by maintaining only the real-time updated supplementary rule information, the model's review results can be ensured to be based on new knowledge, reducing the time cost of manually modifying the document review checklist or retraining the model, and enabling rapid real-time data updates.

[0154] In one optional embodiment of this specification, obtaining supplementary information on the rules for the document review checklist may include the following steps:

[0155] Call a third-party API to process the documents to be reviewed and obtain the processing results;

[0156] The processing results will be used as supplementary information for the rules of the document review checklist.

[0157] It should be noted that third-party interfaces, also known as external interfaces, refer to Application Programming Interfaces (APIs) provided by external service providers. Third-party interfaces allow document review systems to communicate with external services, obtain data, or use their provided functions without needing to understand or manage the specific implementation details of the service. When processing documents to be reviewed by calling third-party interfaces, the processing result refers to the data or information returned by the interface. These results typically include analysis or judgment of the document content, including but not limited to the following: Content classification: What category does the document belong to, such as news, comments, advertisements, etc.? Sensitive word detection: Does the document contain sensitive words, and what are the specific locations and types of these words? Sentiment analysis: Is the document's sentiment positive, negative, or neutral? Compliance check: Does the document comply with specific laws, regulations, or industry standards? Copyright check: Does the document have any copyright issues, and has it used unauthorized content?

[0158] For example, during the process of reviewing the counterparty risk of a contract to be reviewed, the enterprise's qualifications can be determined in real time by calling an external interface, and the qualification review results can be identified as supplementary information for the rules, thereby helping the document review model generate more comprehensive document review results.

[0159] By applying the solutions in the embodiments of this specification, a third-party interface is called to process the document to be reviewed and obtain the processing results. The processing results are determined as supplementary information for the rules of the document review checklist. Thus, the existing review rules can be updated or improved based on the analysis results provided by the third-party interface, which helps to improve the review process, improve the accuracy and efficiency of the review, and adapt to the ever-changing network environment and legal and regulatory requirements in a timely manner.

[0160] In one optional embodiment of this specification, the risk modification suggestions in the document review results may contain certain errors. Therefore, before modifying the document to be reviewed using the risk remediation suggestions, the rationality and feasibility of the risk modification suggestions can be verified to avoid introducing new risks or problems due to inappropriate modifications. That is, the document review results include risk detection results and risk modification suggestions. After reviewing the document to be reviewed using multiple review rules and obtaining the document review results, the following steps may also be included:

[0161] If the risk detection results indicate that the document to be reviewed has risks, the risk modification suggestions are verified and the verification results are obtained.

[0162] If the verification result is successful, modify the document to be reviewed according to the risk modification suggestions to obtain the target modified document.

[0163] It should be noted that the risk detection result refers to the outcome of the document review process, where the document under review is examined and identified using various review rules across multiple dimensions in the document review checklist to determine if any risks exist. If the risk detection result indicates that the document under review has no risks, the risk modification suggestion can be empty; if the risk detection result indicates that the document under review has risks, the risk modification suggestion can be a proposed correction or improvement measure targeting the identified risks. The target modified document refers to the document under review after modifications based on the risk modification suggestion. The target modified document does not contain any of the risks previously identified during the review.

[0164] In practical applications, there are various ways to verify risk modification suggestions and obtain verification results. The specific method chosen depends on the actual situation, and this specification does not limit the methods used in this embodiment. In one possible implementation, preset error suggestions can be obtained, and it can be determined whether the risk modification suggestions include these preset error suggestions. If so, the verification result is determined to be a failure; if the preset error suggestions are not included, the verification result is determined to be a success. In another possible implementation, a third-party verification tool can be called to verify the risk modification suggestions and obtain the verification results.

[0165] Furthermore, if the verification result is a failure, risk modification suggestions can be regenerated, or adjusted to obtain verification-passing modification suggestions. If the verification result is a pass, one possible implementation is to directly modify the document content suggested in the risk modification suggestions to the corresponding content. Another possible implementation in this specification is to call a large model to modify the document to be reviewed based on the risk modification suggestions, thereby obtaining the target modified document.

[0166] By applying the solution in the embodiments of this specification, when the risk detection result indicates that the document to be reviewed has a risk, the risk modification suggestion is verified, and the document to be reviewed is modified using the verified risk modification suggestion, thereby improving the accuracy of the target document to be modified.

[0167] In one optional embodiment of this specification, the document review result includes risk detection results and risk modification suggestions; after reviewing the document to be reviewed using multiple review rules and obtaining the document review result, the following steps may also be included:

[0168] If the risk detection results indicate that the document to be reviewed has risks, risk modification suggestions will be sent to the client.

[0169] Receive target modification information sent by the client, wherein the target modification information is obtained by the client based on document review requirements and risk modification suggestions;

[0170] Modify the document to be reviewed based on the target modification information to obtain the target modified document.

[0171] It should be noted that the risk modification suggestions generated by the server may not meet the client's needs. Therefore, the risk modification suggestions can be sent to the client, which can then determine the target modification information that meets the document review requirements. The client can then modify the document to be reviewed based on the target modification information to obtain the target modified document.

[0172] For example, suppose the risk modification suggestion is "to clearly stipulate the liability for breach of contract at the end of the contract". After the risk modification suggestion is sent to the client, if the client feels that the liability for breach of contract should correspond to the specific agreement, then the client sends the target modification information to the server: "to clearly stipulate the liability for breach of contract in Article 4 on page 7 of the contract".

[0173] By applying the solution in the embodiments of this specification, risk modification suggestions are sent to the client for confirmation. The target modification information sent by the client is used to modify the document to be reviewed, thereby obtaining the target modified document, which improves the flexibility of document review.

[0174] The following is in conjunction with the appendix Figure 3 Taking the document review method provided in this specification as an example in a legal document processing scenario, the document review method will be further explained. Among other things, Figure 3 This specification illustrates a flowchart of a legal document review method according to an embodiment, which specifically includes the following steps:

[0175] Step 302: Obtain the legal documents to be reviewed.

[0176] Step 304: Identify the type of the legal document to be reviewed to obtain its document type.

[0177] Step 306: Based on the document type, determine the document review list for the legal documents to be reviewed. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0178] Step 308: Use multiple review rules to review the legal documents to be reviewed and obtain the document review results.

[0179] It should be noted that legal documents pending review refer to documents or materials in legally relevant scenarios that require a review process. Legal documents pending review include, but are not limited to, contracts (such as commercial contracts, employment contracts, etc.), agreements (such as shareholder agreements, investment agreements, settlement agreements, etc.), and litigation documents (such as complaints, answers, etc.). The specific selection depends on the actual situation, and this specification does not impose any limitations on this.

[0180] In practical applications, the implementation methods of steps 302 to 308 are the same as those of steps 202 to 208 described above, and will not be repeated in the embodiments of this specification.

[0181] The solution implemented in this specification breaks down the document review checklist into multiple review rules with different review dimensions using a divide-and-conquer approach. This enables precise document review for different review rules. Furthermore, when reviewing legal documents using each review rule, the complete legal document to be reviewed is reviewed, allowing the review process to be flexibly analyzed based on the context of the legal document to be reviewed, further improving the accuracy of legal document review.

[0182] Figure 4 This specification illustrates a flowchart of another legal document review method provided in one embodiment, which specifically includes the following steps:

[0183] Step 402: Receive the legal documents to be reviewed sent by the client.

[0184] Step 404: Using the document review checklist of the legal documents to be reviewed, review the legal documents to be reviewed, obtain the document review results of the legal documents to be reviewed, and feed back the document review results to the client. The document review checklist is determined based on the document type of the legal documents to be reviewed. The document type is obtained by identifying the type of the legal documents to be reviewed. The document review checklist includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0185] In practical applications, the implementation methods of steps 402 to 404 are the same as those of steps 202 to 208 described above, and will not be repeated in the embodiments of this specification.

[0186] By applying the solutions in the embodiments of this specification, a document review checklist is used to achieve precise document review for different review rules. Furthermore, when reviewing legal documents using each review rule, the complete legal document to be reviewed is reviewed, which allows the review process to be flexibly analyzed based on the context of the legal document to be reviewed, further improving the accuracy of legal document review.

[0187] See Figure 5 , Figure 5 This document illustrates a flowchart of a legal document review method according to one embodiment of this specification. Taking a contract document as an example, and the legal document review method being executed by a contract review intelligence agent, the flowchart of the legal document review method is explained. This process includes four stages: contract document uploading, contract review checklist construction, contract document review, and summary application.

[0188] Contract document upload: Users upload the contract document to be reviewed in the contract review agent. The contract review agent calls the document review model to parse the contract document and determine the document type of the contract document to be reviewed.

[0189] Contract review checklist construction: A contract review checklist suitable for the contract documents to be reviewed can be constructed in three ways: First, the default checklist of the contract review agent (which can be manually configured by experts based on expert knowledge or professional books) is determined as the contract review checklist, realizing the cold start and application of contract review; Second, based on the contract documents to be reviewed, the default checklist is selected and modified to obtain a contract review checklist with more precise rules (user-selected checklist), achieving more accurate contract review; Third, user-defined review rules are received, and a contract review checklist that can cover more comprehensive review scenarios is constructed based on the user-defined review rules (custom checklist). In practical applications, the contract review checklist can be constructed in a semi-automatic way: Obtain multiple review dimensions summarized by experts for the document type of the contract document to be reviewed; input the multiple review dimensions and the contract document to be reviewed into the rule generation model to obtain multiple review rules (such as review rule 1, review rule 2, ...), where the review rules correspond one-to-one with the review dimensions; construct the contract review checklist for the contract documents to be reviewed based on the multiple review rules;

[0190] Contract Document Review: The contract review agent can analyze each review rule and generate risk assessment results based on the rules. For example, for the rule "the contract must include termination conditions," if the contract does not contain termination conditions, the contract review agent can determine that there is a risk. Specifically, for each review rule, the contract review agent can call the document review model in parallel to perform review, executing multi-task judgments including risk detection, risk factor determination, risk type identification, risk content search, and risk content modification, obtaining the review results of the contract document under multiple review rules. Optionally, when calling the document review model for review, it can determine whether the contract review checklist meets the rule supplementary conditions. If so, it obtains the rule supplementary information (external real-time knowledge) for the document review checklist, and calls the document review model in parallel based on multiple review rules and rule supplementary information to review the contract document. Typically, the contract review process includes three main types of rule reviews: First, basic review: including confirming the purpose of the contract, verifying the qualifications of the parties involved, and reviewing the legality of the contract; second, detailed review of the contract terms, including rights and obligations, liability for breach of contract, dispute resolution mechanisms, etc., to ensure that the terms are clear, fair, reasonable, and compliant; third, review of specific rules: including review of special rules for this type of contract or customized requirements from the client.

[0191] Summary Application: After obtaining the review results of the contract document under multiple review rules, the Contract Review Agent can integrate these results based on set priorities and importance to obtain the final contract review result for the contract document. For example, higher-risk clauses can be assigned higher scores to make them stand out in the review results. Furthermore, the Contract Review Agent can perform result verification to determine if the risk detection results in the review results indicate that the contract document under review contains risks. If so, it verifies the risk modification suggestions in the review results. If the verification passes, the risk modification suggestions are applied, and the contract document under review is modified accordingly to obtain the target modified contract. Alternatively, if the risk detection results indicate that the contract document under review contains risks, the agent can send risk modification suggestions to the client; receive target modification information from the client, which is derived by the client based on the document review requirements and risk modification suggestions; and modify the contract document under review according to the target modification information to obtain the target modified contract.

[0192] In one optional embodiment of this specification, the contract review agent sends the contract review results to the client; receives rule update information sent by the client based on the contract review results, wherein the rule update information includes at least one of rule deletion information, rule modification information, and rule addition information; and updates the contract review list of contract documents to be reviewed according to the rule update information to obtain an updated contract review list, thereby improving the contract review effect. Thus, a high-quality contract review list is gradually formed during the continuous review process, thereby iterating to become a more accurate and comprehensive contract review agent.

[0193] Applying the solutions in the embodiments of this specification, firstly, different contract review checklists can be flexibly generated based on different document types. A rule generation model is used to generate review rules not covered by expert knowledge, helping contract reviews cover different areas. Secondly, by adopting a divide-and-conquer approach to decompose the review task, the contract review checklist is broken down into multiple review rules with different review dimensions. This supports more comprehensive rules (rules that cannot be covered by regular expressions), achieving precise contract review for different review rules. Simultaneously, it solves the limitation of model input length on contract review, enabling the review of more rules.

[0194] Corresponding to the above-described document review method embodiments, this specification also provides embodiments of a document review device. Figure 6 A schematic diagram of a document review device according to one embodiment of this specification is shown. Figure 6 As shown, the device includes:

[0195] The first acquisition module 602 is configured to acquire documents to be reviewed.

[0196] The first identification module 604 is configured to identify the type of the document to be reviewed and obtain the document type of the document to be reviewed.

[0197] The first determining module 606 is configured to determine a document review list for documents to be reviewed based on the document type. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0198] The first review module 608 is configured to review the document to be reviewed using multiple review rules to obtain the document review result.

[0199] Optionally, the first determining module 606 is further configured to obtain multiple candidate review lists and candidate document types of the multiple candidate review lists; match the document types and candidate document types of the multiple candidate review lists respectively to obtain multiple matching results, wherein the matching results correspond one-to-one with the candidate review lists; and select the document review list of the document to be reviewed from the multiple candidate review lists based on the multiple matching results.

[0200] Optionally, the first determining module 606 is configured to obtain multiple review dimensions corresponding to the document type; generate a model by combining the multiple review dimensions and the input rules of the document to be reviewed, and obtain multiple review rules, wherein the review rules correspond one-to-one with the review dimensions; and construct a document review list of the document to be reviewed based on the multiple review rules.

[0201] Optionally, the first review module 608 is further configured to input the review rules and the document to be reviewed into the document review model to obtain the review result of the document to be reviewed under the review rules; and to determine the document review result of the document to be reviewed based on the review results of the document to be reviewed under multiple review rules.

[0202] Optionally, the device further includes: a third acquisition module configured to acquire audit prompt information, wherein the audit prompt information is used to guide the document audit model to perform document audit based on a target audit task, the target audit task including at least one of a risk detection task, a risk factor determination task, a risk type identification task, a risk content search task, and a risk content modification task; and a first audit module 608 further configured to input the audit prompt information, audit rules, and the document to be audited into the document audit model to obtain the audit result of the document to be audited under the audit rules.

[0203] Optionally, the first review module 608 is further configured to input the review results of the document to be reviewed under multiple review rules into the document review model for integration, so as to obtain the document review result of the document to be reviewed.

[0204] Optionally, the first review module 608 is further configured to obtain the result priority of the review results of the document to be reviewed under multiple review rules; and determine the document review result of the document to be reviewed based on the result priority.

[0205] Optionally, the first identification module 604 is configured to acquire multiple candidate document types; input the multiple candidate document types and the document to be reviewed into a type identification model to obtain the document type of the document to be reviewed.

[0206] Optionally, the device further includes: a first sending module configured to send document review results to a client; receive rule update information sent by the client based on the document review results, wherein the rule update information includes at least one of rule deletion information, rule modification information, and rule addition information; and update the document review list of documents to be reviewed according to the rule update information to obtain the updated document review list.

[0207] Optionally, the device further includes: a fourth acquisition module, configured to acquire supplementary rule information for the document review list when the document review list meets the supplementary rule conditions; and a first review module 608, further configured to review the document to be reviewed using multiple review rules and supplementary rule information respectively, and obtain the document review result of the document to be reviewed.

[0208] Optionally, the fourth acquisition module is further configured to call a third-party interface to process the document to be reviewed and obtain the processing result; and to determine the processing result as supplementary information for the rules of the document review checklist.

[0209] Optionally, the document review result includes risk detection results and risk modification suggestions; the device further includes: a verification module, configured to verify the risk modification suggestions when the risk detection result indicates that the document to be reviewed has risks, and obtain a verification result; if the verification result is that the verification passes, the document to be reviewed is modified according to the risk modification suggestions to obtain the target modified document.

[0210] Optionally, the document review result includes risk detection results and risk modification suggestions; the device further includes: a second sending module, configured to send risk modification suggestions to the client when the risk detection result indicates that the document to be reviewed has risks; receive target modification information sent by the client, wherein the target modification information is obtained by the client based on the document review requirements and risk modification suggestions; modify the document to be reviewed according to the target modification information to obtain the target modified document.

[0211] The solution implemented in this specification breaks down the document review checklist into multiple review rules with different review dimensions using a divide-and-conquer approach. This achieves precise document review for different review rules. Furthermore, when reviewing documents using each review rule, the complete document to be reviewed is reviewed, allowing the review process to be flexibly analyzed based on the context of the document to be reviewed, further improving the accuracy of document review.

[0212] The above is an illustrative scheme of a document review device according to this embodiment. It should be noted that the technical solution of this document review device and the technical solution of the document review method described above belong to the same concept. For details not described in detail in the technical solution of the document review device, please refer to the description of the technical solution of the document review method described above.

[0213] Corresponding to the above embodiments of the legal document review method, this specification also provides embodiments of the legal document review device. Figure 7 A schematic diagram of a legal document review device according to one embodiment of this specification is shown. Figure 7 As shown, the device includes:

[0214] The second acquisition module 702 is configured to acquire legal documents to be reviewed.

[0215] The second identification module 704 is configured to identify the type of the legal document to be reviewed and obtain the document type of the legal document to be reviewed.

[0216] The second determination module 706 is configured to determine a document review list of legal documents to be reviewed based on the document type. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0217] The second review module 708 is configured to review the legal documents to be reviewed using multiple review rules, and obtain the document review results of the legal documents to be reviewed.

[0218] The solution implemented in this specification breaks down the document review checklist into multiple review rules with different review dimensions using a divide-and-conquer approach. This enables precise document review for different review rules. Furthermore, when reviewing legal documents using each review rule, the complete legal document to be reviewed is reviewed, allowing the review process to be flexibly analyzed based on the context of the legal document to be reviewed, further improving the accuracy of legal document review.

[0219] The above is an illustrative scheme of a legal document review device according to this embodiment. It should be noted that the technical solution of this legal document review device and the technical solution of the above-described legal document review method belong to the same concept. For details not described in detail in the technical solution of the legal document review device, please refer to the description of the technical solution of the above-described legal document review method.

[0220] Corresponding to the above embodiments of the legal document review method, this specification also provides embodiments of the legal document review device. Figure 8 A schematic diagram of another legal document review device provided in one embodiment of this specification is shown. Figure 8 As shown, the device includes:

[0221] The receiving module 802 is configured to receive legal documents to be reviewed sent by the client;

[0222] The third review module 804 is configured to review the legal documents to be reviewed using a document review list, obtain the document review results of the legal documents to be reviewed, and feed the document review results back to the client. The document review list is determined based on the document type of the legal documents to be reviewed. The document type is obtained by identifying the type of the legal documents to be reviewed. The document review list includes multiple review rules corresponding to the document type, and the review dimensions of the multiple review rules are different.

[0223] By applying the solutions in the embodiments of this specification, a document review checklist is used to achieve precise document review for different review rules. Furthermore, when reviewing legal documents using each review rule, the complete legal document to be reviewed is reviewed, which allows the review process to be flexibly analyzed based on the context of the legal document to be reviewed, further improving the accuracy of legal document review.

[0224] The above is an illustrative scheme of a legal document review device according to this embodiment. It should be noted that the technical solution of this legal document review device and the technical solution of the above-described legal document review method belong to the same concept. For details not described in detail in the technical solution of the legal document review device, please refer to the description of the technical solution of the above-described legal document review method.

[0225] Figure 9 A structural block diagram of a computing device 900 provided in one embodiment of this specification is shown.

[0226] The computing device 900 includes:

[0227] Memory 910 and processor 920;

[0228] The memory 910 is used to store computer programs / instructions, and the processor 920 is used to execute the computer programs / instructions. When the computer programs / instructions are executed by the processor 920, they implement the steps of the above-mentioned document review method or legal document review method.

[0229] In one or more embodiments of this specification, the computing device can be understood as an integrated smart terminal, including but not limited to a server, desktop computer, personal computer (PC), all-in-one model machine, mobile phone, tablet computer or other portable smart terminal, etc., and the computing device may have the model described in the above embodiments of this application pre-installed.

[0230] Specifically, this computing device can pre-install various types of models, including but not limited to models in natural language processing, visual processing, speech processing, code processing, and multimodal task processing, thus providing diverse model selection. In different product forms, this computing device can support one or more model usage methods, including but not limited to model training, model invocation, model fine-tuning, model deployment, model inference, and application. In some product forms, this computing device also supports model management, including but not limited to multi-type model management (supporting the management of discriminative, generative, and other model types), model version control (supporting the control of different model versions), and model evaluation (evaluating model performance and effectiveness based on model evaluation tools). In other product forms, this computing device can also create applications based on models, providing API calling capabilities, allowing models to be called into created applications through API interfaces, and providing application management tools for application management and monitoring.

[0231] Furthermore, the computing device may also include data management (supporting the creation and management of model tuning datasets), a training center (providing abundant training resources to help users learn and master artificial intelligence technology), and basic control capabilities (providing enterprise-level basic control capabilities to ensure the security and efficient operation of the system). Through the above functions, it provides a comprehensive and integrated device for artificial intelligence development, training, deployment, and application.

[0232] Figure 10 A structural block diagram of an electronic device 1000 provided according to one embodiment of this specification is shown.

[0233] The memory 1010 and the processor 1020 are connected via a bus 1030;

[0234] The memory 1010 is used to store computer programs / instructions, and the processor 1020 is used to execute the computer programs / instructions. When the computer programs / instructions are executed by the processor 1020, they implement the steps of the above-mentioned code processing model training method, code processing model testing method, code task processing method, or information processing method based on the code processing model.

[0235] Specifically, the components of the electronic device 1000 include, but are not limited to, a memory 1010 and a processor 1020. The processor 1020 and the memory 1010 can be connected via a bus 1030.

[0236] Electronic device 1000 may also include access device 1040, which enables electronic device 1000 to communicate with database 1050 storing data via one or more networks 1060. Examples of such networks include Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or combinations of communication networks such as the Internet. Access device 1040 may include one or more of any type of wired or wireless network interface (e.g., a network interface card (NIC)), such as an IEEE 802.11 Wireless Local Area Networks (WLAN) interface, a Wi-MAX (World Interoperability for Microwave Access) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC) interface, and so on.

[0237] In one embodiment of this specification, the above-described components of the electronic device 1000 and Figure 10 Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 10 The block diagram of the electronic device shown is for illustrative purposes only and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0238] Electronic device 1000 can be any type of stationary or mobile electronic device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable electronic devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary electronic devices such as desktop computers or PCs. Electronic device 1000 can also be a mobile or stationary electronic device.

[0239] The above is an illustrative scheme of an electronic device according to this embodiment. It should be noted that the technical solution of this electronic device belongs to the same concept as the document review method and the legal document review method described above. For details not described in detail in the technical solution of the electronic device, please refer to the description of the technical solution of the document review method or the legal document review method described above.

[0240] An embodiment of this specification also provides a computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the above-described document review method or legal document review method.

[0241] The above is an illustrative scheme of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium belongs to the same concept as the technical solutions of the document review method and the legal document review method described above. Details not described in detail in the technical solution of the storage medium can be found in the descriptions of the technical solutions of the document review method or the legal document review method described above.

[0242] An embodiment of this specification also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described document review method or legal document review method.

[0243] The above is an illustrative scheme of a computer program product according to this embodiment. It should be noted that the technical solution of this computer program product belongs to the same concept as the technical solutions of the document review method and the legal document review method described above. For details not described in detail in the technical solution of the computer program product, please refer to the description of the technical solutions of the document review method or the legal document review method described above.

[0244] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0245] The computer instructions include computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium may be appropriately added or removed according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media may not include electrical carrier signals and telecommunication signals.

[0246] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments in this specification are not limited to the described order of actions, because according to the embodiments in this specification, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments in this specification.

[0247] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0248] The preferred embodiments disclosed above are merely illustrative of this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments described herein. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the embodiments, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. A document review method, comprising: Retrieve documents pending review; The document to be reviewed is identified by type to obtain the document type of the document to be reviewed; Based on the document type, a document review list for the document to be reviewed is determined, wherein the document review list includes multiple review rules corresponding to the document type, and the multiple review rules have different review dimensions; The document to be reviewed is reviewed using the various review rules respectively, and the document review result of the document to be reviewed is obtained.

2. The method according to claim 1, wherein determining the document review list of the documents to be reviewed based on the document type includes: Obtain multiple candidate review lists and the candidate document types for the multiple candidate review lists; The document type is matched with the candidate document types in the multiple candidate review lists to obtain multiple matching results, wherein each matching result corresponds one-to-one with the candidate review list; Based on the multiple matching results, a document review list for the documents to be reviewed is selected from the multiple candidate review lists.

3. The method according to claim 1, wherein determining the document review list of the documents to be reviewed based on the document type includes: Obtain multiple review dimensions corresponding to the document type; The multiple review dimensions and the input rules of the document to be reviewed are used to generate a model to obtain multiple review rules, wherein each review rule corresponds one-to-one with the review dimension; Based on the multiple review rules, a document review list of the documents to be reviewed is constructed.

4. The method according to claim 1, wherein the step of reviewing the document to be reviewed using the plurality of review rules respectively to obtain the document review result of the document to be reviewed includes: Input the audit rules and the document to be audited into the document audit model to obtain the audit result of the document to be audited under the audit rules; The document review result of the document to be reviewed is determined based on the review results of the document under the multiple review rules.

5. The method according to claim 4, further comprising, before inputting the audit rules and the document to be audited into the document audit model to obtain the audit result of the document to be audited under the audit rules: Obtain review prompt information, wherein the review prompt information is used to guide the document review model to review the document based on the target review task, and the target review task includes at least one of the following: risk detection task, risk factor determination task, risk type identification task, risk content search task, and risk content modification task; The step of inputting the review rules and the document to be reviewed into the document review model to obtain the review result of the document to be reviewed under the review rules includes: The review prompt information, the review rules, and the document to be reviewed are input into the document review model to obtain the review result of the document to be reviewed under the review rules.

6. The method according to claim 4, wherein determining the document review result of the document to be reviewed based on the review results of the document to be reviewed under the plurality of review rules includes: The review results of the document to be reviewed under the multiple review rules are input into the document review model for integration to obtain the document review result of the document to be reviewed.

7. The method according to claim 4, wherein determining the document review result of the document to be reviewed based on the review results of the document to be reviewed under the plurality of review rules includes: Obtain the result priority of the review results of the document to be reviewed under the multiple review rules; Based on the priority of the results, the document review result of the document to be reviewed is determined.

8. The method according to claim 1, wherein the step of performing type identification on the document to be reviewed to obtain the document type of the document to be reviewed includes: Retrieve multiple candidate document types; By inputting the multiple candidate document types and the document to be reviewed into the document type recognition model, the document type of the document to be reviewed is obtained.

9. The method according to any one of claims 1 to 8, wherein after reviewing the document to be reviewed using the plurality of review rules to obtain the document review result of the document to be reviewed, the method further includes: Send the document review results to the client; The system receives rule update information sent by the client based on the document review result, wherein the rule update information includes at least one of rule deletion information, rule modification information, and rule addition information; Update the document review list of the documents to be reviewed according to the rules, and obtain the updated document review list.

10. The method according to any one of claims 1 to 8, wherein before reviewing the document to be reviewed using the plurality of review rules respectively to obtain the document review result of the document to be reviewed, the method further includes: If the document review checklist meets the supplementary rule conditions, obtain supplementary rule information for the document review checklist; The step of reviewing the document to be reviewed using the multiple review rules to obtain the document review result includes: The document to be reviewed is reviewed using the multiple review rules and supplementary information for those rules, and the document review result is obtained.

11. The method according to claim 10, wherein obtaining supplementary rule information for the document review checklist includes: Call a third-party interface to process the document to be reviewed and obtain the processing result; The processing result is determined as supplementary information for the rules of the document review checklist.

12. The method according to any one of claims 1 to 8, wherein the document review result includes risk detection results and risk modification suggestions; After reviewing the document to be reviewed using the multiple review rules and obtaining the document review result, the process further includes: If the risk detection result indicates that the document to be reviewed has a risk, the risk modification suggestion is verified to obtain the verification result; If the verification result is successful, the document to be reviewed is modified according to the risk modification suggestions to obtain the target modified document.

13. The method according to any one of claims 1 to 8, wherein the document review result includes risk detection results and risk modification suggestions; After reviewing the document to be reviewed using the multiple review rules and obtaining the document review result, the process further includes: If the risk detection result indicates that the document to be reviewed has a risk, the risk modification suggestion will be sent to the client; The system receives target modification information sent by the client, wherein the target modification information is obtained by the client based on document review requirements and the risk modification suggestions; The document to be reviewed is modified according to the target modification information to obtain the target modified document.

14. A method for reviewing legal documents, comprising: Obtain the legal documents pending review; The document type of the legal document to be reviewed is determined by performing type identification. Based on the document type, a document review list for the legal documents to be reviewed is determined, wherein the document review list includes multiple review rules corresponding to the document type, and the multiple review rules have different review dimensions; The legal documents to be reviewed are reviewed using the various review rules respectively, and the document review results of the legal documents to be reviewed are obtained.

15. A method for reviewing legal documents, comprising: Receive legal documents pending review sent by the client; Using the document review list of the legal documents to be reviewed, the legal documents to be reviewed are reviewed to obtain the document review results, and the document review results are fed back to the client. The document review list is determined based on the document type of the legal documents to be reviewed. The document type is obtained by identifying the type of the legal documents to be reviewed. The document review list includes multiple review rules corresponding to the document type, and the multiple review rules have different review dimensions.

16. A computing device, comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the method according to any one of claims 1 to 15.

17. An electronic device comprising: A memory and a processor, the memory and the processor being connected via a bus; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the method according to any one of claims 1 to 15.

18. A computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 15.

19. A computer program product comprising a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 15.

Images