WO2022111548A1

WO2022111548A1 - Contract review method and apparatus, and readable storage medium

Info

Publication number: WO2022111548A1
Application number: PCT/CN2021/132929
Authority: WO
Inventors: 徐青松; 李青
Original assignee: 杭州睿胜软件有限公司
Priority date: 2020-11-26
Filing date: 2021-11-24
Publication date: 2022-06-02
Also published as: CN112330214A

Abstract

The present invention provides a contract review method and apparatus, and a readable storage medium. First, text information of a target contract is identified; then a contract type of the target contract and term types of terms of the target contract are identified by using a text classification model on the basis of the text information of the target contract; according to the contract type of the target contract and the term types of the terms, a matching term template is found in a sample database for term content comparison, and the risk level of each of the terms is output according to the comparison result. That is, by first automatically identifying text of a target contract, then using a text classification model to search a sample database for a term template matching the type of terms of the target contract, carrying out automatic comparative analysis, and outputting risk points in the contract text, a contract can be automatically reviewed and checked, and the contract review efficiency is improved.

Description

Contract review method, device and readable storage medium

technical field

The present invention relates to the technical field of artificial intelligence, and in particular, to a contract review method, device and readable storage medium.

Background technique

At present, legal workers are faced with the drafting and revision of a large number of contracts. In order to reduce the risks in the contracts, it is necessary to manually review the contracts, which consumes a lot of manpower and time. In order to improve the efficiency of the review, the automatic review and proofreading of the contract can improve the efficiency of the contract review, especially for the contract text with many clauses.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a contract review method to automatically review and proofread contracts.

To achieve the above purpose, the present invention provides a contract review method, comprising:

Textual information identifying the target contract;

Based on the text information of the target contract, the text classification model is used to identify the contract type of the target contract and the clause type of each clause of the target contract, and the contract type of the target contract and the clause type of each clause of the target contract According to the clause type, find the matching clause template in the sample database to compare the clause content, and output the risk level of each clause according to the comparison result.

Optionally, in the contract review method, the step of finding a matching clause template in the sample database to compare clause contents, and outputting the risk level of each clause according to the comparison result includes:

Obtain from the sample database a clause template matching the contract type and clause type of each of the clauses;

Compare the similarity of word feature vectors to each of the clauses and the matched clause templates;

According to the obtained similarity and the matching risk level of the clause template, the risk level of each clause is output.

Optionally, in the contract review method, the sample database includes a plurality of clause templates that match the contract type and clause type of each of the clauses, and the matching clause template is found in the sample database. The steps of comparing the contents of the clauses and outputting the risk level of the clauses according to the comparison results include:

obtaining, from the sample database, a plurality of clause templates matching the contract type and clause type of each of the clauses;

Performing a comparison of the similarity of word feature vectors on each of the clauses and a plurality of matched clause templates to confirm the clause template with the greatest similarity with each of the clauses;

According to the maximum similarity obtained by comparison and the risk level of the item template with the maximum similarity, the risk level of each item is output.

Optionally, in the contract review method, the clause templates in the sample database include: positive samples with no risk or a risk level lower than the first threshold, and/or with a risk or a risk level higher than the first threshold. A negative example of the second threshold;

If the similarity between the clause and the positive example is higher, the risk level of the clause is lower, and if the similarity between the clause and the negative example is higher, the risk level of the clause is higher.

Optionally, in the contract review method, the text classification model is obtained by training contract samples in advance.

Optionally, in the contract review method, the step of training contract samples to obtain the text classification model includes:

Annotate the text paragraphs of each contract sample in the sample database to indicate the contract type, clause type and risk level of each text paragraph;

The text paragraphs marked with contract types, clause types and risk levels are trained by using a neural network model to obtain the text classification model.

Optionally, in the contract review method, the contract review method further includes: saving the terms of the target contract to the sample database, so as to update the text classification model with new samples of the sample database .

Optionally, in the contract review method, the contract review method further includes:

outputting the alignment result using a textual term output model;

The outputted comparison results include:

display the differences between the target contract and the matching clause template;

show the favourable parties of said target contract; and/or

Label the terms at risk and the standard template terms that match the terms at risk.

According to the risk level of each said clause, generate a risk report and send it to the client for confirmation by the client, and,

After client confirmation, the risk report is saved.

In another aspect, the present invention also provides a contract review device, comprising:

Text recognition module, used to recognize the text information of the target contract;

The comparison module is used to identify the contract type of the target contract and the clause types of each clause of the target contract based on the text information of the target contract, using a text classification model, and the contract type according to the target contract Find a matching clause template in the sample database to compare the clause content with the clause type of each clause, and output the risk level of each clause according to the comparison result.

Optionally, in the contract review device, the comparison module finds a matching clause template in a sample database to compare clause contents, and the step of outputting the risk level of each clause according to the comparison result includes:

According to the obtained similarity and the matched risk level of the clause template, the risk level of each clause is output.

Optionally, in the contract review device, the sample database includes a plurality of clause templates that match the contract type and clause type of each of the clauses, and the comparison module finds matching clauses in the sample database. The clause template compares clause contents and outputs the risk level of each clause according to the comparison result, including:

Optionally, in the contract review device, the clause templates in the sample database include: positive samples with no risk or a risk level lower than the first threshold, and/or with a risk or a risk level higher than the first threshold. A negative example of the second threshold;

Optionally, in the contract review apparatus, the text classification model is obtained by pre-training contract samples.

Optionally, in the contract review device, the step of training contract samples to obtain the text classification model includes:

Marking the text paragraphs of each contract sample in the sample database to mark the contract type, clause type and risk level of each text paragraph;

Optionally, in the contract review device, the contract review device further comprises: a storage module, configured to save the terms of the target contract to the sample database, so as to be updated with a new sample of the sample database the text classification model.

Optionally, in the contract review device, the contract review device further includes an output module, and the output module is configured to output the comparison result by using a text clause output model;

Wherein, the outputted comparison results include:

show the favourable parties of said target contract; and/or

Optionally, in the contract review apparatus, the comparison module is further configured to generate a risk report according to the risk level of each of the clauses and send it to the client for confirmation by the client.

The present invention also provides a readable storage medium, where a computer program is stored in the readable storage medium, and when the computer program is executed by a processor, the above-mentioned contract review method is implemented.

To sum up, the contract review method, device and readable storage medium provided by the present invention include: identifying text information of a target contract; based on the text information of the target contract, using a text classification model to identify the target The contract type of the contract and the clause type of each clause of the target contract, and according to the contract type of the target contract and the clause type of each clause, find a matching clause template in the sample database to compare the clause content, and The risk level of each of the items is output according to the comparison result. That is, first through the automatic identification of the target contract text, and then use the text classification model to search the sample database for the clause template that matches the type of the clause of the target contract, perform automatic comparison and analysis, and output the risk points in the contract text. In this way, the efficiency of contract review is improved by automatically reviewing and proofreading the contract.

Description of drawings

1 is a flowchart of a contract review method provided by an embodiment of the present invention;

FIG. 2 is a block diagram of the composition of a contract review device provided by an embodiment of the present invention;

Wherein, each reference sign is described as follows:

11-text recognition module; 12-comparison module; 13-output module; 14-storage module.

Detailed ways

The contract review method, device and readable storage medium proposed by the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent from the following description. It should be noted that, the accompanying drawings are all in a very simplified form and in inaccurate scales, and are only used to facilitate and clearly assist the purpose of explaining the embodiments of the present invention. Furthermore, the structures shown in the drawings are often part of the actual structure. In particular, each drawing needs to show different emphases, and sometimes different scales are used.

It should be noted that, in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any relationship between these entities or operations. any such actual relationship or sequence exists. Moreover, the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device comprising a list of elements includes not only those elements, but also includes not explicitly listed or other elements inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in a process, method, article or apparatus that includes the element.

As shown in FIG. 1 , this embodiment provides a contract review method, and the contract review method includes the following steps:

S11, identifying the text information of the target contract;

S12, based on the text information of the target contract, using a text classification model, identify the contract type of the target contract and the clause type of each clause of the target contract, and identify the contract type of the target contract and the clause type of each clause of the target contract According to the clause type of the clause, find the matching clause template in the sample database to compare the clause content, and output the risk level of each clause according to the comparison result.

The contract review method provided in this embodiment firstly identifies the target contract text automatically, and then uses the text classification model to search the sample database for a clause template that matches the type of the clause of the target contract, and performs automatic comparison and analysis. Output the risk points in the contract text, thus improving the efficiency of contract review by automatically reviewing and proofreading the contract.

The above steps are further described in detail below.

In step S11, the text presentation of the target contract may be a Word version, a PDF version, a PPT version, a TXT version or a picture version. The textual information of various versions of the contract can be recognized by the character recognition model.

When identifying the text information of the target contract, the target contract can also be compared with the contract texts of different historical versions, and through the comparison, it can be judged whether the content in the target contract and the content of the original contract (contract of any historical version) are not Consistent to avoid tampering with the target contract text. The target contract and the original contract can be in the same format or in different formats (eg, one or more of Word, PDF, PPT, TXT, etc.).

Specifically, the following methods can be used to compare the target contract with the original contract:

Obtain images of the target contract and electronic documentation of the original contract;

Identify each line of character regions in the image of the target contract based on a pre-trained region identification model, where the region identification model is a neural network-based model;

Based on the pre-trained character recognition model, identify the character content in the character area of each line, and obtain the recognized character, and the character recognition model is a model based on a neural network; obtain the position information of the recognized character;

generating an electronic document of the target contract according to the location information and the recognized characters; comparing the content of the electronic document of the target contract with the electronic document of the original contract; and

According to the comparison result, it is judged whether there is a difference between the target contract and the original contract, and the difference point is located according to the position information of the difference point.

In this way, the comparison between the target contract and the original contract is completed through the above steps. If there is a difference, it means that the target contract may have been tampered with. After verifying the located difference, step S12 is performed.

In step S12, when using the text classification model to find a matching clause template in the sample database to compare clause content, and output the risk level of each clause according to the comparison result, the following steps can be specifically adopted:

Obtain a clause template matching the contract type and clause type of each of the clauses from the sample database; compare each of the clauses and the matched clause template for the similarity of word feature vectors; according to the comparison The obtained similarity and the matching risk level of the clause template are used to output the risk level of each clause.

In addition, preferably, in the establishment stage of the sample database, a plurality of clause templates with the same contract type and clause type are collected and saved to the sample database, so that the text classification obtained by training using the sample database When the model compares clauses, the sample database includes a plurality of clause templates matching the contract type and clause type of each clause. On this basis, in step S12, when using the text classification model to find a matching clause template in the sample database to compare clause contents, and outputting the risk level of each clause according to the comparison result, the specific method can be as follows: step:

Obtain a plurality of clause templates matching the contract type and clause type of each clause from the sample database; compare the similarity of word feature vectors for each clause and the matched clause templates respectively Yes, to confirm the clause template with the maximum similarity with each of the clauses; output the risk level of each clause according to the maximum similarity obtained from the comparison and the risk level of the clause template with the maximum similarity.

Wherein, the text classification model is obtained by training contract samples in advance. The contract samples are contract samples in the sample database. Specifically, the following steps can be used to train contract samples to obtain the text classification model:

Mark the text paragraphs of each contract sample in the sample database to mark the contract type, clause type and risk level of each text paragraph as shown in Table 1; use the neural network model to mark the contract type, clause type and the text passages of the risk level are trained to obtain the text classification model.

In this way, after identifying the text information of the target contract, the text classification model can be used to identify the contract type of the target contract, the clause type of each clause, and the risk level of each clause.

In this embodiment, the presentation of the risk levels can be diversified, for example, the identifiers a, b, c, d are used to distinguish different levels, or the text content is high, high, low, low, etc. to distinguish different levels level.

Table 1

合同类型type of contract	条款类型Clause Type	内容content	风险等级Risk level
WWWWWW	wwwwww	AAAAAAAA	aaaaaa
XXXXXX	xxxxxx	BBBBBBBB	bbbbbb
YYYYYY	yyyyyy	CCCCCCCCCC	cccccc
ZZZZZZ	zzzzzz	DDDDDDDDDDDD	dddddd
。。。. . .	。。。. . .	。。。. . .	。。。. . .

In some other embodiments, after the contract type, clause type and risk level of each text paragraph are marked, the neural network model can be used to train the contract type to obtain the first classification model (also called contract type identification). model), use the neural network model to train the clause types to obtain the second classification model (also called the clause classification model), then use the first classification model to identify the contract type of the target contract, and use the second classification model to identify the target contract. The clause type of each clause, and then, using the character matching recognition model, according to the contract type of the target contract and the clause type of each clause, find a matching clause template in the sample database to compare the clause content, and according to the comparison As a result, the risk level for each of the stated items is output.

As can be seen from the above description, the text classification model may be a general classification model, and may also include several sub-classification models. The specific presentation form of the text classification model does not constitute a limitation to the present application, and only needs to make the model identifiable by using the model. The contract type of the target contract, the clause type of each of the clauses, and the risk level of each of the clauses are sufficient.

In addition, during sample training, the samples in the sample database can also be classified according to the risk degree of the clause, and divided into positive samples and negative samples. If the clause text in the contract text has no risk or the risk degree is lower than the first Threshold, the clause text can be used as a positive sample of the clause type and stored in the sample database; if the clause text in the contract text has a negative sample with risk or the risk degree is higher than the second threshold, the clause text can be As a counter-example sample of the clause type, it is stored in the sample database. In this way, in step S12, during the comparison, if the similarity with the positive sample is higher, the risk level of the item is lower, and if the similarity with the negative sample is higher, the risk level of the item is higher. The higher the risk level.

Further, the terms of the target contract can also be saved to the sample database, so as to update the text classification model with new samples of the sample database.

Optionally, in step S12, in addition to outputting the risk level of each of the terms according to the comparison result, a risk report may also be generated according to the risk level of each of the terms and sent to the client for confirmation by the client, and , and after the client confirms, save the risk report.

In addition, the contract review method provided in this embodiment may further include: outputting the comparison result by using a text clause output model; the outputting the comparison result includes: displaying the difference between the target contract and the matched clause template ; show the favorable party of the target contract (the term is self-, opposite, or neutral); and/or mark the clause at risk and the standard template clause that matches the clause at risk.

That is to say, in this embodiment, the comparison result in step S12 can be presented in various forms, and the above-mentioned examples do not constitute a limitation to the present application. For example, the comparison result can also annotate the clauses with risk, that is, adding a remark column, annotating suggestions and reasons for modifying the clauses, or outputting a revised version according to the clause template and the risk level.

The embodiment of the present invention also provides a contract review device, including:

Text recognition module 11, used to recognize the text information of the target contract;

The comparison module 12 is configured to, based on the text information of the target contract, use a text classification model to identify the contract type of the target contract and the clause types of each clause of the target contract, and the contract according to the target contract Type and the clause type of each said clause, find a matching clause template in the sample database to compare the clause content, and output the risk level of each said clause according to the comparison result.

Wherein, the comparison module 12 finds a matching clause template in the sample database to compare clause contents, and outputs the risk level of each clause according to the comparison result. The contract type of the said clause and the clause template matching the clause type; compare the similarity of the word feature vector between each clause and the matched clause template; The risk level of the clause template, output the risk level of each said clause.

Further, the sample database includes a plurality of clause templates that match the contract type and clause type of each of the clauses, and the comparison module 12 finds the matching clause template in the sample database to compare the clause contents, and The step of outputting the risk level of each of the clauses according to the comparison result includes: acquiring a plurality of clause templates matching the contract type and clause type of each of the clauses from the sample database; A plurality of said clause templates are respectively compared for the similarity of word feature vectors to confirm the clause template with the greatest similarity with each of the clauses; The risk level of the clause template is output, and the risk level of each clause is output.

The clause template and the text classification model have been described in detail in the previous section, and will not be repeated here. In addition to outputting the risk level of each of the clauses according to the comparison result, optionally, the comparison module 12 is further configured to generate a risk report according to the risk level of each of the clauses and send it to the client for confirmation by the client.

In addition, corresponding to the contract review method provided in this embodiment, the contract review apparatus further includes an output module 13, and the output module 13 is configured to output the comparison result by using a text clause output model; The results include: showing the difference between the target contract and the matching clause template; showing the favorable party of the target contract; and/or, marking the clause at risk and the clause at risk. matching standard template terms, etc.

The contract review apparatus may further include: a storage module 14, configured to save the terms of the target contract to the sample database, so as to update the text classification model with new samples of the sample database. In addition, client-confirmed risk reports can be stored.

All in all, each module in the contract review device provided by this embodiment is respectively used to implement each step of the contract review method provided by this implementation. Therefore, for the specific description of the functions that each module can implement, please refer to the corresponding contract review method described above. The relevant description of the steps will not be repeated where repeated. In addition, the contract reviewing device can achieve the same technical effect as the above-mentioned contract reviewing method, which will not be repeated here.

It can be understood that, in the contract review device, the text recognition module 11 , the comparison module 12 , the output module 13 and the storage module 14 may be combined in one device, or any one of them may be implemented in one device. The module can be divided into multiple sub-modules, or, at least part of the functions of one or more modules of the text recognition module 11, the comparison module 12, the output module 13 and the storage module 14 can be combined with other modules. At least part of the functions of the modules are combined and implemented in one functional module. According to an embodiment of the present invention, in the contract review apparatus, at least one of the text recognition module 11 , the comparison module 12 , the output module 13 and the storage module 14 may be at least partially implemented as Hardware circuits, such as Field Programmable Gate Arrays (FPGA), Programmable Logic Arrays (PLA), System on Chip, System on Substrate, System on Package, Application Specific Integrated Circuit (ASIC), or circuits that can be integrated or packaged It can be implemented in any other reasonable manner, such as hardware or firmware, or in an appropriate combination of software, hardware and firmware.

It can be seen from the above description that the contract reviewing method of the embodiment of the present invention can be applied to the contract reviewing apparatus of the embodiment of the present invention. In addition, the contract review apparatus can be configured on an electronic device, wherein the electronic device can be a personal computer, a mobile terminal, etc., and the mobile terminal can be a mobile phone, a tablet computer, or other hardware devices with various operating systems. The electronic device includes a processor and a memory, and the memory is used to store a computer program; when the computer program is executed by the processor, the contract review method provided in this embodiment is implemented.

In the electronic device, the memory may include random access memory (Random Access Memory, RAM), or may include non-volatile memory (Non-Volatile Memory, NVM), such as at least one disk memory.

Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

The processor can be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; it can also be a digital signal processor (Digital Signal Processing, DSP), dedicated integrated Circuit (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

This embodiment further provides a readable storage medium, where a computer program is stored in the readable storage medium, and when the computer program is executed by a processor, the contract review method provided in this embodiment is implemented.

The readable storage medium can be a tangible device that can hold and store instructions for use by the instruction execution device, such as, but not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or the above. any suitable combination. More specific examples (non-exhaustive list) of readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or Flash memory), static random access memory (SRAM), portable compact disc read only memory (CD-ROM), digital versatile disc (DVD), memory sticks, floppy disks, mechanical coding devices, and any suitable combination of the foregoing. The computer programs described herein can be downloaded to various computing/processing devices from readable storage media, or to external computers or external storage devices over a network such as the Internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers. A network adapter card or network interface in each computing/processing device receives the computer program from the network and forwards the computer program for storage in a readable storage medium in the respective computing/processing device. The computer program for carrying out the operations of the present invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or any other program in one or more programming languages. Combining source or object code written in programming languages including object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer program may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server . Where a remote computer is involved, the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, through the Internet using an Internet service provider) connect). In some embodiments, electronic circuits, such as programmable logic circuits, field programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), that can execute computer programmable logic circuits, are personalized by utilizing state information from a computer program. Program instructions are read to implement various aspects of the present invention.

To sum up, the contract review method, device and readable storage medium provided by the present invention first identify the text information of the target contract, and then use the text classification model to identify the target contract based on the text information of the target contract. According to the contract type of the target contract and the clause type of each clause of the target contract, and according to the contract type of the target contract and the clause type of each clause, find a matching clause template in the sample database to compare the clause content, and according to The comparison result outputs the risk level of each of the mentioned items. That is, first through the automatic identification of the target contract text, and then use the text classification model to search the sample database for the clause template that matches the type of the clause of the target contract, perform automatic comparison and analysis, and output the risk points in the contract text. In this way, the efficiency of contract review is improved by automatically reviewing and proofreading the contract.

The above description is only a description of the preferred embodiments of the present invention, and is not intended to limit the scope of the present invention. Any changes and modifications made by those of ordinary skill in the field of the present invention based on the above disclosure all belong to the protection scope of the claims.

Claims

A contract review method, comprising:

Textual information identifying the target contract;

Based on the text information of the target contract, the text classification model is used to identify the contract type of the target contract and the clause type of each clause of the target contract, and the contract type of the target contract and the clause type of each clause of the target contract According to the clause type, find the matching clause template in the sample database to compare the clause content, and output the risk level of each clause according to the comparison result.
The contract review method according to claim 1, wherein the step of finding a matching clause template in a sample database to compare clause contents, and outputting the risk level of each clause according to the comparison result comprises:

Obtain from the sample database a clause template matching the contract type and clause type of each of the clauses;

Compare the similarity of word feature vectors to each of the clauses and the matched clause templates;

According to the obtained similarity and the matched risk level of the clause template, the risk level of each clause is output.
The contract review method according to claim 1, wherein the sample database includes a plurality of clause templates matching the contract type and clause type of each of the clauses, and the matching clause is found in the sample database The template compares the content of the clauses, and outputs the risk level of each clause according to the comparison result, including:

obtaining, from the sample database, a plurality of clause templates matching the contract type and clause type of each of the clauses;

Performing a comparison of the similarity of word feature vectors on each of the clauses and a plurality of matched clause templates to confirm the clause template with the greatest similarity with each of the clauses;

According to the maximum similarity obtained by comparison and the risk level of the item template with the maximum similarity, the risk level of each item is output.
The contract review method according to claim 2 or 3, wherein the clause templates in the sample database include: positive samples with no risk or with a risk degree lower than a first threshold, and/or, with risk or Counter-examples whose risk is higher than the second threshold;

If the similarity between the clause and the positive example is higher, the risk level of the clause is lower, and if the similarity between the clause and the negative example is higher, the risk level of the clause is higher.
The contract review method according to claim 1, wherein the text classification model is obtained by training contract samples in advance.
The contract review method according to claim 5, wherein the step of training contract samples to obtain the text classification model comprises:

Marking the text paragraphs of each contract sample in the sample database to mark the contract type, clause type and risk level of each text paragraph;

The text paragraphs marked with contract types, clause types and risk levels are trained by using a neural network model to obtain the text classification model.
The contract review method according to claim 6, wherein the contract review method further comprises: saving the terms of the target contract to the sample database, so as to update the text with a new sample of the sample database classification model.
The contract review method according to claim 1, wherein the contract review method further comprises:

outputting the alignment result using a textual term output model;

Wherein, the outputted comparison results include:

display the differences between the target contract and the matching clause template;

show the favourable parties of said target contract; and/or

Label the terms at risk and the standard template terms that match the terms at risk.
The contract review method according to claim 1, wherein the contract review method further comprises:

According to the risk level of each said clause, a risk report is generated and sent to the client for confirmation by the client, and,

After confirmation by the client, the risk report is saved.
A contract review device, characterized in that it includes:

Text recognition module, used to recognize the text information of the target contract;

The comparison module is used to identify the contract type of the target contract and the clause types of each clause of the target contract based on the text information of the target contract, using a text classification model, and the contract type according to the target contract Find a matching clause template in the sample database to compare the clause content with the clause type of each clause, and output the risk level of each clause according to the comparison result.
The contract review apparatus according to claim 10, wherein the comparison module finds a matching clause template in a sample database to compare clause contents, and outputs the risk level of each clause according to the comparison result include:

Obtain from the sample database a clause template matching the contract type and clause type of each of the clauses;

Compare the similarity of word feature vectors to each of the clauses and the matched clause templates;

According to the obtained similarity and the matched risk level of the clause template, the risk level of each clause is output.
The contract review apparatus according to claim 10, wherein the sample database includes a plurality of clause templates that match the contract type and clause type of each of the clauses, and the comparison module finds the corresponding clauses in the sample database. The matched clause template compares clause content, and outputs the risk level of each clause according to the comparison result, including:

obtaining, from the sample database, a plurality of clause templates matching the contract type and clause type of each of the clauses;

Performing a comparison of the similarity of word feature vectors on each of the clauses and a plurality of matched clause templates to confirm the clause template with the greatest similarity with each of the clauses;

According to the maximum similarity obtained by comparison and the risk level of the item template with the maximum similarity, the risk level of each item is output.
The contract review apparatus according to claim 11 or 12, wherein the clause templates in the sample database include: positive samples with no risk or a risk degree lower than a first threshold, and/or with a risk or Counter-example samples whose risk is higher than the second threshold;

If the similarity between the clause and the positive example is higher, the risk level of the clause is lower, and if the similarity between the clause and the negative example is higher, the risk level of the clause is higher.
The contract review apparatus according to claim 13, wherein the text classification model is obtained by pre-training contract samples.
The contract review apparatus according to claim 14, wherein the step of training contract samples to obtain the text classification model comprises:

Marking the text paragraphs of each contract sample in the sample database to mark the contract type, clause type and risk level of each text paragraph;

The text paragraphs marked with contract types, clause types and risk levels are trained by using a neural network model to obtain the text classification model.
The contract review device according to claim 15, characterized in that, the contract review device further comprises: a storage module, configured to save the terms of the target contract to the sample database, so as to utilize the new data of the sample database. The samples update the text classification model.
The contract review device according to claim 10, wherein the contract review device further comprises an output module, and the output module is configured to output the comparison result by using a text clause output model;

Wherein, the outputted comparison results include:

display the differences between the target contract and the matching clause template;

show the favourable parties of said target contract; and/or

Label the terms at risk and the standard template terms that match the terms at risk.
The contract review apparatus according to claim 10, wherein the comparison module is further configured to generate a risk report according to the risk level of each of the clauses and send it to the client for confirmation by the client.
A readable storage medium, wherein a computer program is stored in the readable storage medium, and when the computer program is executed by a processor, the contract review method according to any one of claims 1 to 9 is implemented .