JP2022169992A - Information processing apparatus, information processing method, terminal program, server program, and contract correction support system - Google Patents

Abstract

To provide an information processing apparatus configured to improve the accuracy of correction in accordance with actual correction results of contracts, an information processing method, a terminal program, a server program, and a contract correction support system.SOLUTION: An information processing apparatus includes: a receiving unit which receives a contract and a designation of words included in the contract, from a user terminal used by a user; a generation unit which generates a revision for the designated words using a learning model trained with learning data including a first dataset and configured to receive words in the contract as input data and output corrected words obtained by correcting the words as correct-answer data; and a display unit which displays the revision on the user terminal.SELECTED DRAWING: Figure 3

Description

The present invention relates to an information processing device, an information processing method, a terminal program, a server program, and a contract revision support system.

Conventionally, a technique for assisting revision of a contract is known. For example, Patent Literature 1 discloses a system for correcting (proofreading) a contract in accordance with rules governing relationships between clauses used in the contract.

Japanese Patent Application Laid-Open No. 2020-160971

Here, there is a need to improve the accuracy of contract revisions as the number of contract revisions increases. However, the technique disclosed in Patent Literature 1 has a problem that it is difficult to improve the accuracy of the revision of the contract even if the processing results increase because of the rule-based revision processing.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide an information processing apparatus, an information processing method, a terminal program, a server program, and a contract revision support system that can improve the accuracy of revisions in accordance with the track record of revisions to a contract. do.

An information processing apparatus according to an aspect of the present invention includes a receiving unit that receives a contract and specification of wording included in the contract from a user terminal used by a user, and a receiving unit that receives the wording of the contract as input data. Using a learning model learned from learning data including a first data set in which corrected corrected wording is correct data, a generation unit generates a correction plan for a specified wording, and a user terminal displays the correction plan. a display unit;

In an information processing method according to one aspect of the present invention, a computer accepts a contract and specification of wording included in the contract from a user terminal used by a user, and uses the wording of the contract as input data. Using a learning model learned from learning data including a first data set in which the corrected corrected wording is correct data, a correction plan for the designated wording is generated and the correction plan is displayed on the user terminal.

A terminal program according to one aspect of the present invention provides a user terminal used by a user with a terminal reception function for accepting a contract and specification of wordings included in the contract from the user, and inputting the wordings of the contract as input data. A terminal storage function that stores a learning model trained using learning data including a first data set that uses the corrected wording as correct data, and a specified wording is input to the learning model, and A generation function for generating a revision proposal and a terminal display function for displaying the revision proposal are realized.

A server program according to an aspect of the present invention provides a server device communicably connected to a user terminal used by a user, and a reception function for receiving, from the user terminal, a contract and specification of language included in the contract. , a memory function for storing a learning model trained using learning data including a first data set in which the wording of the contract is input data and the modified wording obtained by correcting this wording is correct data; and learning the specified wording. A generation function for inputting to the model to generate a revision proposal for the wording and a display function for displaying the revision proposal on the user terminal are realized.

A contract revision support system according to one aspect of the present invention is a contract revision support system including a user terminal used by a user and a server device communicably connected to the user terminal, wherein the user terminal comprises: A terminal receiving unit for receiving a contract and specification of wordings included in the contract from the user, and when the terminal receiving unit receives specification of the contract and wordings from the user, the specified wordings are transmitted to the server device. a terminal transmission unit, wherein the server device includes a reception unit that receives the wording transmitted from the user terminal; A model storage unit that stores a learning model learned by learning data including a set, a generation unit that inputs the received wording to the learning model and generates a correction proposal for the wording, and transmits the correction proposal to the user terminal. a transmitting unit, and the user terminal comprises a terminal receiving unit for receiving the revision proposal, and a terminal display unit for displaying the revision proposal.

According to the present invention, it is possible to provide an information processing apparatus, an information processing method, a terminal program, a server program, and a contract revision support system that can improve the accuracy of revision in accordance with the track record of revision of a contract.

1 is a diagram for explaining a system configuration example of a contract revision support system according to this embodiment; FIG. BRIEF DESCRIPTION OF THE DRAWINGS It is a figure for demonstrating an example of the outline|summary of the contract correction assistance system which concerns on this embodiment. It is a figure showing an example of functional composition of a server device concerning this embodiment. It is a figure for demonstrating an example of the calculation method of the degree of certainty of the contract revision support system which concerns on this embodiment. It is a figure showing an example of functional composition of a user terminal concerning this embodiment. It is a figure which shows the example of a screen of the contract correction support system which concerns on this embodiment. It is a figure which shows the example of a screen of the contract correction support system which concerns on this embodiment. It is a figure which shows the example of a screen of the contract correction support system which concerns on this embodiment. It is a figure which shows the operation example of the contract correction support system which concerns on this embodiment. It is a figure which shows the operation example of the contract correction support system which concerns on this embodiment. It is a figure which shows an example of the hardware constitutions of the server apparatus and user terminal which concern on this embodiment. It is a figure which shows an example of a structure of the learning model of the contract revision assistance system which concerns on this embodiment, and learning data.

A preferred embodiment of the present invention (hereinafter referred to as "this embodiment") will be described with reference to the accompanying drawings. It should be noted that, in each figure, the same reference numerals have the same or similar configurations.

In this embodiment, by using the contract revision support system 1 according to this embodiment, a user such as a person in charge of the legal department of a company can An example of making corrections to a received contract will be explained. However, the present invention is not intended to be limited to this example. This “contract received from a business partner, etc.” is, for example, a document created by a business partner correcting a contract template (hereinafter also referred to as “contract template”). Also, the contract may be, for example, a document in which a sales representative of the company corrects the contract template to create a draft contract, and the client corrects the draft. The contract revision support system 1 can also be used for contracts other than contracts created by modifying contract templates in this way, for example. A plurality of contract templates may exist, and for example, they may exist for at least one of the user type and the contract type. Here, the "user type" is used to distinguish users according to specific types, such as an organization (for example, a company, an association, etc.), a department, etc., to which the user belongs. Here, the "type of contract" is used to distinguish contracts according to specific types, such as business consignment agreements, non-disclosure agreements (NDA), sales contracts, and the like. Also, a contract template may exist, for example, for each organization to which the user belongs and for each business partner of the organization.

<1. System configuration>
A system configuration example of a contract revision support system 1 according to the present embodiment will be described with reference to FIG.

The contract revision support system 1 is a system that uses natural language processing technology to provide a user with a service that supports the revision of a contract (hereinafter also referred to as "revision support service"). As shown in FIG. 1, the contract revision support system 1 includes a server device 100 used by a person who provides a revision support service, and a user terminal 200 used by a user. The server device 100 and the user terminal 200 are connected via a network N so as to be able to communicate with each other. In this embodiment, the terms “contract” and “contract file” are used synonymously. The contract is saved in the file format of various document creation software. Common examples of document creation software include Adobe Acrobat (PDF) from Adobe and Word (registered trademark) from Microsoft.

The network N is composed of a wireless network and a wired network. Examples of the network N include a mobile phone network, a PHS (Personal Handy-phone System) network, a wireless LAN (Local Area Network), 3G (3rd Generation), LTE (Long Term Evolution), 4G (4th Generation), 5G (5th Generation), WiMax (registered trademark), infrared communication, Bluetooth (registered trademark), wired LAN, telephone line, power line network, and IEEE1394-compliant network.

The server device 100 is an information processing device for providing a correction support service. By executing a predetermined program (server program), the server device 100 implements a server function that provides various functions for amending the contract according to various requests and various specifications from the user.

A user terminal 200 is an information processing device used by a user. The user terminal 200 is, for example, a terminal such as a smart phone or a laptop. The user terminal 200 can install and operate predetermined software (for example, document creation software, etc.) and can communicate with the server device 100 . By executing a predetermined program, the user terminal 200 cooperates with the server device 100 to transmit and receive information (hereinafter also referred to as “display information”) regarding a screen for correcting the contract, and display this screen. or receive a specified input from the user. In this embodiment, an example will be described in which the predetermined program is a terminal-side program (hereinafter also referred to as a "terminal program") dedicated to the contract revision support system 1. FIG. Furthermore, in this embodiment, an example in which the terminal program is an add-in of document creation software installed in the user terminal 200 will be described.

<2. Overview>
An outline of the contract revision support system 1 will be described with reference to FIG.

(1) As shown in FIG. 2, in the contract revision support system 1, first, the user displays (opens) the contract received from the business partner or the like with the document creation software to which the terminal program is added. The user designates wordings to be corrected for the wordings included in the contract displayed on the document creation software.

(2) Next, the user designates a correction proposal display button in a state in which the wording to be corrected is designated. This "revision proposal display button" is a button for requesting a revision proposal for the wording of the designated contract. The details of the revision proposal display button will be described later.

(3) Next, when the revision proposal display button is specified as in (2) above, the contract revision support system 1 uses the learning model to generate a revision proposal for the wording specified in (1) above. do. Then, the contract revision support system 1 causes the user terminal 200 to display the generated revision proposal. Here, the “learning model” is a model learned by learning data (teacher data) including the first data set. Here, the "first data set" is a data set in which the wording of the contract is used as input data, and corrected wording obtained by correcting this wording is used as correct data. Details of the learning model will be described later.

(4) Next, when the user specifies that the proposed amendment displayed in (3) above should be reflected in the contract, the contract amendment support system 1 applies this proposed amendment to the contract displayed in the document creation software. To reflect.

Based on the above configuration, the contract revision support system 1 uses a learning model learned from learning data consisting of a contract wording and a revised wording that is a set of revised wording to generate a contract revision proposal. to users. This learning model increases the amount of learning data and improves its performance as the number of revisions to contracts increases. Therefore, the contract revision support system 1 can improve the accuracy of revision according to the track record of revision of the contract.

<3. Functional configuration>
<3-1. Functional Configuration of Server Device>
A functional configuration of the server device 100 according to the present embodiment will be described with reference to FIG. As shown in FIG. 3, the server device 100 includes a control section 110, a communication section 120, and a storage section .

[Control section]
Control unit 110 includes reception unit 111 , learning unit 112 , generation unit 113 , and display unit 114 . The control unit 110 may also include a document reflecting unit 115, for example.

[Reception Department]
The receiving unit 111 receives various requests and various designations from the user terminal 200 . The receiving unit 111 receives, for example, a contract and specification of wording included in the contract. For example, when the user designates the wording of the contract displayed on the document creation software and requests a revision proposal, the reception unit 111 receives the request from the terminal reception unit 211 of the user terminal 200, which will be described later, as this displayed request. Receive a written contract and specified language. The "displayed contract and specified wording" may be, specifically, attribute information of the displayed contract and a character string of the specified wording.

"Contract attribute information" includes, for example, identification information for identifying each contract (e.g., contract name, contract file name or ID, etc.), contract type, information on the contractor, etc. . In addition, this "information on the contractor (hereinafter also referred to as "contractor information")" is, for example, the attribute information of the parties ("A", "B", etc.) described in the contract, and identifies the parties. including identification information for the purpose, business content, business scale, business form, or capital status of the parties.

For example, the receiving unit 111 may receive, from the user terminal 200, a revised wording in which the wording of the contract is revised. For example, when the user modifies the wording of the contract displayed in the document creation software, the receiving unit 111 receives the corrected wording of the contract from the terminal receiving unit 211 of the user terminal 200. - 特許庁

For example, the reception unit 111 may receive from the user terminal 200 a request to adopt a revision proposal displayed by the display unit 114, which will be described later. For example, when the user designates a revision proposal reflection button a22 displayed on a revision proposal display screen A2 shown in FIG. You may receive an offer to adopt the proposed amendments displayed in the

For example, the receiving unit 111 may receive, from the user terminal 200, a specification of the wording of the contract and a request that the wording does not need to be modified.

For example, the receiving unit 111 may receive designation of a wording corresponding to the data set to be reflected in the learning data, from among the following wordings (a) to (d) displayed by the display unit 114 . For example, the receiving unit 111 may receive an instruction from the user terminal 200 as to whether or not to reflect the second data set corresponding to the corrected wording of (a) below displayed by the display unit 114 in the learning data. Further, the receiving unit 111 may receive, for example, from the user terminal 200 an instruction as to whether or not to reflect the fourth data set corresponding to the correction proposal (b) below displayed by the display unit 114 .

[Study Department]
The learning unit 112 causes the learning model to learn using the learning data. A learning model is a model learned by a machine learning method. The learning model is typically a model learned by a deep learning method, but is not limited to this. The learning model may be, for example, a model using other algorithms such as SVM (Support Vector Machine) or decision tree.

The learning model may be, for example, a recurrent neural network (RNN) as a model using a deep learning technique. Also, the learning model may be, for example, an LSTM (Long Short-Term Memory) capable of considering long-term time series among RNNs. Also, the learning model may be, for example, a seq2seq (sequence-to-sequence) model using RNN for the encoder and decoder, and may further include an attention mechanism.

The learning data may include, for example, at least one of the following second to sixth data sets in addition to the first data set.
・Second data set: A data set in which the wording of the contract before correction corresponding to the revised wording corrected by the user is used as input data, and this revised wording is used as correct data ・Third data set: Input data and correct data are used as input data A data set that uses the same contract template wording (in this case, the target contract is a document created by modifying the contract template)
・Fourth data set: Data set that uses the wording of the contract corresponding to the revision proposal adopted by the user as input data and the wording of this revision proposal as correct data Data set with the wording of the contract as input data and correct data ・6th data set: The wording of the contract corresponding to the displayed revision proposal is used as input data, and the displayed revision proposal is modified by the user is the correct data set

For example, in the revision of a contract, there are cases in which a state in which there is no revision from the contract template can be correct. According to the above configuration, by reflecting the third data set in which the input data and the correct data are the same contract template wording in the learning data, the learning model is trained with the data set that reproduces such a case. be able to.

For example, during learning, the learning unit 112 may evaluate the accuracy of the correction proposal output from the learning model. The learning unit 112 may, for example, calculate the BLEU score of the correction plan based on the correction plan output from the learning model and the correct data.

[Data reflection part]
Learning unit 112 includes data reflecting unit 1121 . The data reflecting unit 1121 reflects each of the first to sixth data sets in the learning data manually by the user or automatically (when a predetermined condition is satisfied, etc.).

The data reflecting section 1121 may include, for example, a first data reflecting section 1121a. When the user himself/herself corrects the wording of the contract, the first data reflection unit 1121a takes the wording of the contract before correction corresponding to the corrected wording corrected by the user as input data, and regards the corrected wording as correct data. The second data set is reflected in the learning data.

For example, the first data reflection unit 1121a reflects, in the learning data, a second data set in which the wording of the contract is input data and the corrected wording by the user is correct data, as follows.
・Input data: The text of the contract “If a lawsuit arises regarding this contract, the Nagoya District Court shall be the exclusive jurisdictional court of first instance.”
・Correct data: Amended wording in which the user modified the wording of the above contract ``If the need for litigation regarding this contract arises, the Tokyo District Court shall be the exclusive agreement jurisdictional court of the first instance.'' User modified the exclusive agreement jurisdictional court from "Nagoya District Court" to "Tokyo District Court")

Based on the above configuration, when the user corrects the wording of the contract, the first data reflection unit 1121a can reflect the content of the correction as learning data. Therefore, the first data reflection unit 1121a can improve the accuracy of the learning model as the number of revisions of the contract by the user increases. In addition, the first data reflection unit 1121a can efficiently accumulate the learning data, and thus can efficiently improve the accuracy of the learning model.

The first data reflection unit 1121a may reflect the second data set in the learning data, for example, based on the result of the user's instruction as to whether or not to reflect the second data set in the learning data. For example, when the receiving unit 111 receives a user's instruction to reflect the second data set in the learning data, the first data reflecting unit 1121a reflects the second data set in the learning data. On the other hand, the first data reflecting unit 1121a does not reflect the second data set in the learning data, for example, when the receiving unit 111 receives a user's instruction not to reflect the second data set in the learning data.

Based on the above configuration, the first data reflection unit 1121a can reflect the user's intentions on the learning data and, by extension, on the learning model. For example, when a user makes an exceptional amendment to a special contract, the user may not want to apply this amendment to a general contract. In such a case, the first data reflection unit 1121a can prevent the learning model from learning the exceptional correction according to the user's instruction.

The data reflecting section 1121 may include, for example, a second data reflecting section 1121b. The second data reflection unit 1121b uses the text of the contract corresponding to the revision proposal adopted by the user as input data, and reflects the fourth data set in which the text of the adopted revision proposal is correct data in the learning data. do.

With the above configuration, when the wording of the contract is corrected by the correction proposal generated by the generation unit 113, the second data reflection unit 1121b can reflect the contents of the correction proposal as learning data. Therefore, the second data reflection unit 1121b can improve the accuracy of the learning model as the number of contract revisions based on revision proposals increases. In addition, the second data reflection unit 1121b can efficiently accumulate the learning data, and thus can efficiently improve the accuracy of the learning model.

The second data reflection unit 1121b reflects the fourth data set in the learning data, for example, based on the result of the user's instruction as to whether or not to reflect the fourth data set in the learning data received by the reception unit 111. good too.

Based on the above configuration, the second data reflection unit 1121b can reflect the user's intentions on the learning data and, by extension, on the learning model. For example, when the generation unit 113 generates an exceptional revision proposal for a special contract or the like, it may not be desired to apply this exceptional revision proposal to a general contract. Also, the revision proposal generated by the generation unit 113 may not be appropriate. In such cases, the second data reflection unit 1121b can prevent the learning model from learning this exceptional correction proposal or inappropriate correction proposal in accordance with the user's instruction.

The data reflecting section 1121 may include, for example, a third data reflecting section 1121c. The third data reflection unit 1121c reflects, in the learning data, the fifth data set in which the wording of the contract for which no correction is required is set as the input data and the correct data.

For example, as a result of a user who is a member of the legal department reviewing a contract received from a business partner, there is a case in which a state in which there is no correction to the wording of this contract may be correct. Based on the above configuration, the third data reflection unit 1121c reflects the fifth data set in which the input data and the correct data are the same contract wording in the learning data, thereby reproducing such a case. The set can be fed back to the training data.

The data reflecting section 1121 may include, for example, a fourth data reflecting section 1121d. The fourth data reflection unit 1121d uses the wording of the contract corresponding to the revision proposal as input data, and reflects, in the learning data, a sixth data set in which the wording of the revision proposal is corrected as correct data.

Based on the above configuration, the fourth data reflecting unit 1121d can feed back the learning data in a form in which errors in the revision proposal are corrected. For example, if the learning model is a model using a neural network, the fourth data reflection unit 1121d can utilize the sixth data set for back propagation (error propagation method) when learning this learning model. .
[Generation part]
The generation unit 113 uses the learning model learned by the learning unit 112 to generate a revision proposal for the text of the contract designated by the user. For example, the generation unit 113 first performs natural language processing such as morphological analysis on the phrase specified by the user, and divides the phrase into units such as words and phrases. The generation unit 113 may, for example, refer to the dictionary storage unit 131 that stores a word dictionary, and divide the phrase into units such as words according to the word dictionary. The generating unit 113 divides the specified wording into predetermined units such as words, inputs the result to the learning model (input layer), and causes the learning model to output a correction proposal for the specified wording. good.

[Calculation part]
The generation unit 113 may include the calculation unit 1131, for example. The calculation unit 1131 uses the learning model to calculate the certainty of the revision plan. The “certainty degree of the revision proposal” is a degree indicating the certainty of the revision proposal. For example, the calculation unit 1131 may calculate a statistical value obtained by statistically processing the certainty of each word included in the wording of the revision proposal as the certainty of the revision proposal. A statistical value is, for example, an average value, a median value, a mode value, or the like. The degree of certainty of each word may be, for example, the distribution of the generation probability (appearance probability) of each word at the word position. The confidence of each word may be, for example, the co-occurrence probability of the i-th word and the j-th word. Further, the certainty factor of the revision proposal may be, for example, a statistic value of the BLEU score of the revision proposal for each contract, each clause, or each clause, which is calculated by the learning unit 112 during learning.

Here, with reference to FIG. 4, an example of the calculation method of the certainty factor of the revision proposal according to this embodiment will be described. In this example, a seq2seq model is used as a learning model, and a generation probability (in units of %) is calculated as the degree of certainty of each word included in a correction plan.

As shown in FIG. 4, the calculation unit 1131 calculates the certainty of each word when outputting the correction proposal. For example, when the wording of a contract is input to the learning model as input data, the calculation unit 1131 calculates the distribution of word generation probabilities at the position of the first word, and finds the word “ko”, which is the word with the maximum probability (90%). output. Next, the calculation unit 1131 inputs this output “ko” to the learning model (hidden layer), calculates the distribution of word generation probabilities at the next word position, and finds the word with the maximum probability (93%) is output. Next, the calculation unit 1131 inputs this output “wa” to the learning model, further calculates the distribution of word generation probabilities at the next word position, and finds the word with the maximum probability (80%) “ output the following. The calculation unit 1131 can thus calculate the distribution of word generation probabilities at each word position using the learning model, and output the word with the maximum probability together with its generation probability.

[Display part]
Returning to FIG. 3, the description continues. The display unit 114 causes the user terminal 200 to display the correction proposal generated by the generation unit 113 . The display unit 114 generates, for example, display information for displaying the revision proposal display screen A2 shown in FIG. The display unit 114 transmits the generated display information to the terminal display unit 213 of the user terminal 200, and causes the user terminal 200 to display the revision proposal display screen A2.

Based on the above configuration, the display unit 114 presents to the user a contract revision proposal based on a learning model learned with learning data that includes a contract wording and a revised wording that is a set of the contract wording. be able to. This learning model increases the amount of learning data and improves its performance as the number of revisions to contracts increases. Therefore, the display unit 114 can present to the user a revised proposal with improved accuracy according to the track record of revisions to the contract.

The display unit 114 displays, for example, on the user terminal 200, (a) the revised wording of the contract wording corrected by the user, (b) the wording of the proposed revision adopted by the user, and (c) or (d) a revised wording in which the user has revised the wording of the revision proposal. In addition, the display unit 114 displays, for example, on the user terminal 200 any one of (a) to (d) as a candidate list of data sets to be reflected in the learning data (hereinafter also simply referred to as a “reflection candidate list”). A list containing the wording may be displayed.

The display unit 114 may determine the mode of displaying the revision proposal based on the degree of certainty of the revision proposal calculated by the calculation unit 1131, for example. This “mode of display of revision proposals” includes, for example, presence or absence of display of revision proposals, presence or absence of highlighting of each sentence and each word (for example, bold font, underline, etc.), each sentence and each word This includes setting the character color or background color of words, whether or not to display the degree of confidence, and the like.

For example, the display unit 114 may cause the user terminal 200 to display the statistical value of the certainty of the words included in the revision proposal calculated by the calculation unit 1131 together with the correction proposal. Further, the display unit 114 sets, for example, a first threshold for the certainty of the revision proposal, and if the certainty of the revision proposal is smaller than the first threshold, the display mode of the revision proposal is such that the revision proposal is not displayed. may be determined. For example, if the display unit 114 does not display a correction proposal for the wording specified by the user in this way, the display unit 114 displays a message to that effect instead (for example, "There is no recommended correction proposal") on the user terminal. 200 may be displayed. In addition, the display unit 114 sets a second threshold value for each degree of certainty of each word included in the correction proposal, and if the degree of certainty of the word is smaller than the second threshold, the word is regarded as a word to be confirmed and the user is alerted. In order to do so, the mode of display of this word may be determined to be highlighted.

Based on the above configuration, the display unit 114 can present to the user the certainty of the revision proposal and words requiring confirmation based on the certainty. For this reason, the display unit 114 provides the user with information such as the likelihood of the revision proposal and words to be confirmed, etc., as materials for determining whether or not to reflect the revision proposal and whether or not the revision proposal should be further revised. can be provided.

[Document reflection part]
The document reflecting unit 115 reflects the revision proposal adopted by the user in the contract accepted by the accepting unit 111 . For example, when the user adopts the revision proposal displayed on the user terminal 200, the document reflection unit 115 sends the terminal document reflection unit 212 of the user terminal 200 or the document creation software to the contract displayed by the document creation software. to reflect this proposed amendment. Upon receiving this instruction from the server device 100, the terminal document reflection unit 212 or the document creation software reflects this revision proposal in the displayed contract. According to such a configuration, the document reflecting unit 115 can reflect the revised proposal, which has been improved in accuracy according to the track record of revisions, in the contract.

[Communication]
The communication unit 120 transmits/receives various information including display information to/from the user terminal 200 or the like via the network N. FIG. For example, the communication unit 120 receives information about the contract and information input on the screen from the user terminal 200 via the network N, and transmits display information for displaying various screens to the user terminal 200. do. The communication unit 120 has a transmission unit 121 and a reception unit 122 . The transmission unit 121 transmits the correction proposal to the user terminal 200, for example. The receiving unit 122 receives, for example, the wording transmitted from the user terminal 200, for example, the wording included in the contract specified by the user.

[Memory part]
The storage unit 130 stores learning models, learning data, contract templates, display information, knowledge information, and the like. Further, the storage unit 130 may store, for example, the user information and the contract template in association with each other. The storage unit 130 may store each information using a database management system (DBMS), or may store each information using a file system. When using a DBMS, a table may be provided for each piece of information described above, and each piece of information may be managed by associating the tables.

"User information" is information about a user. The user information may include, for example, identification information for identifying the user (for example, the user's name, ID, etc.), or the type of the user. For example, if the user is a contractor, all or part of the contractor information may be used as the user information.

The storage unit 130 may include a dictionary storage unit 131, for example. The dictionary storage unit 131 stores a word dictionary in which one or more words are registered. Also, in this word dictionary, for example, a single word combining a plurality of words may be registered. This single word may be registered, for example, for each of at least one of the user type and the contract type. For example, the term “target company” is generally divided into multiple words such as “target” and “company”. It may be registered as a single word “company”. Further, in this word dictionary, a phrase combining a plurality of words used in a contract may be registered as a single word for each company as a type of user.

According to the above configuration, it is possible to prevent a single word consisting of a plurality of words that should not be split in the word splitting process from being split. Therefore, it is possible to support the construction of a highly accurate learning model that is more realistic.

The storage unit 130 may include a model storage unit 132, for example. The model storage unit 132 stores learning models. The storage unit 130 may also include a learning data storage unit 133, for example. The learning data storage unit 133 stores learning data.

<3-2. Functional Configuration of User Terminal>
A functional configuration of the user terminal 200 according to the present embodiment will be described with reference to FIG. As shown in FIG. 5 , the user terminal 200 includes a terminal control section 210 , a terminal communication section 220 and a terminal storage section 230 .

[Terminal control part]
The terminal control unit 210 includes a terminal reception unit 211 , a terminal document reflection unit 212 and a terminal display unit 213 . Also, the terminal control unit 210 may include a determination unit 214, for example.

[Terminal reception part]
The terminal accepting unit 211 accepts a contract and specification of language included in the contract from the user. When the user designates the wording of the contract displayed on the document creation software and requests a revision proposal, the terminal reception part 211 receives the displayed contract and the designated wording as the request. The terminal accepting unit 211 may accept, for example, the character string at the cursor position specified by the user as the specified wording. Note that the unit of wording to be specified may be one or more wordings arbitrarily designated by the user, or may be a sentence, clause, article, or the like including the designated wording.

For example, the terminal reception unit 211 may receive, from the user, designation of the wording included in the contract together with the fact that the wording does not need to be corrected.

For example, the terminal reception unit 211 may receive, from the user, a revised wording in which the wording of the contract is revised. The terminal reception unit 211 may, for example, extract a correction part (difference) from the contract template or the previous version of the contract, and accept the extracted correction part as a corrected wording by the user. This "previous version" may be, for example, one previous version, or the latest version saved by a different user (for example, a user of a business partner, etc.) in the document creation software among the previous versions may be

As a method of extracting the wording corresponding to the wording specified by the user from the wordings of the contract template stored in the terminal storage unit 230, the terminal receiving unit 211 extracts the specified wording and the wording of the contract template. may be calculated. The terminal display unit 213 may extract the wording of the contract template that maximizes the calculated similarity, and extract the difference between the extracted wording and the specified wording as the corrected part. It should be noted that the unit for extracting this wording may be for each article or for each clause.

The terminal reception unit 211 may receive knowledge information from the user, for example. Here, "knowledge information" is information indicating knowledge about the contract. This "knowledge about contracts" is the accumulation of users' knowledge about contracts. It may be a summary of The knowledge information may be registered, for example, by at least one of the types of contracts and the clauses.

[Terminal document reflection part]
The terminal document reflecting unit 212 reflects the revision proposal adopted by the user in the contract specified by the user. For example, when the terminal receiving unit 211 receives a request for reflection of a revision proposal from the user, the terminal document reflecting unit 212 applies this correction to the wording specified by the user among the wordings of the contract displayed by the document creation software. reflect the proposal. Specifically, the terminal document reflecting unit 212 deletes the designated wording from the contract and instead writes the wording of the proposed amendment to the contract. As another example, when the terminal reception unit 211 receives a reflection request for a revision proposal from the user, the terminal document reflection unit 212 deletes the designated wording and instead writes the wording of the revision proposal into the contract. You may instruct the document preparation software to do so.

[Terminal display]
When the terminal reception unit 211 receives a correction proposal from the server device 100, the terminal display unit 213 displays the correction proposal on the user terminal 200. FIG. Further, for example, when the user designates the wording of the contract, the terminal display unit 213 may display the knowledge information corresponding to the designated wording. As a method of extracting the knowledge information corresponding to the wording specified by the user from the knowledge information stored in the terminal storage unit 230, the terminal display unit 213, for example, extracts the specified wording and each clause of the contract template. A degree of similarity with the wording may be calculated. The terminal display unit 213 may extract the knowledge information of the article having the maximum calculated similarity and display it on the user terminal 200 . For example, if the language specified by the user is included in Article 22 of the contract, the terminal display unit 213 may display knowledge information corresponding to Article 22.

[Determination part]
The determination unit 214 determines whether or not the wording specified by the user among the wordings included in the contract is corrected. The determination unit 214 refers to, for example, the terminal storage unit 230 that stores the contract template, and determines whether the contract wording specified by the user matches the contract template wording corresponding to this wording. judge. For example, the determining unit 214 may determine that there is no correction when determining that these wordings match, and determine that there is correction when determining that they do not match.

When the judging unit 214 refers to the wording corresponding to the wording specified by the user from among the wordings of the contract template stored in the terminal storage unit 230, for example, similar to the extraction method by the terminal receiving unit 211 described above, , based on the degree of similarity between the designated wording and the wording of the contract template, the wording of the corresponding contract template may be extracted and referred to.

[Terminal communication part]
The terminal communication unit 220 transmits/receives various information including display information to/from the server device 100 and the like via the network N. FIG.

The terminal communication unit 220 has a terminal transmission unit 221 . For example, when the terminal reception unit 211 receives the contract and the specification of the terms included in the contract, the terminal transmission unit 221 transmits the specified terms to the server device 100 via the network N. In addition, the terminal transmission unit 221 may transmit information related to the target contract (for example, part of the contract attribute information, etc.) to the server device 100 together with the specified wording. Further, for example, even if the terminal reception unit 211 receives the designation of the wording included in the contract, the terminal transmitting unit 221 does not have to transmit the wording designated as not requiring correction to the server device 100. . Note that, in this case, the terminal transmission unit 221 may transmit to the server device 100 only the fact that correction is unnecessary.

If the user designates that the text of the contract does not need to be modified, there is no need for the server device 100 to generate a proposed modification. In such a case, based on the above configuration, the terminal transmission unit 221 can reduce the load of communication processing and the bandwidth usage rate of the network by not transmitting the character string of the specified wording.

For example, even when the terminal reception unit 211 receives the designation of the wording included in the contract, the terminal transmission unit 221 receives the designation when the determination unit 214 determines that the designated wording is not corrected (this designation If it is determined that the specified wording matches the wording of the contract template), the designated wording may not be transmitted to the server device 100 . In this case, the terminal transmission unit 221 may transmit to the server device 100 only the fact that these words match.

If the wording of the contract and the wording of the contract template match, it can be considered that the user or the contract revision support system 1 or the like has not made any corrections. In such a case, based on the above configuration, the terminal transmission unit 221 does not transmit a character string of words that can be regarded as uncorrected, thereby reducing the load of communication processing and the bandwidth usage rate of the network. be able to.

The terminal communication unit 220 has a terminal reception unit 222 . The terminal reception unit 222 receives correction proposals, display information, and the like from the server device 100 .

[Terminal memory]
The terminal storage unit 230 stores contracts, contract templates, knowledge information, second, fourth to sixth data sets, and the like.

<4. Screen example>
Screen examples of the contract revision support system 1 will be described with reference to FIGS.

FIG. 6 is a diagram showing an example of a support menu screen displaying a menu of support functions of the contract correction support system 1. As shown in FIG. As shown in FIG. 6, the support menu screen A1 is displayed on the sidebar of the editing screen by the terminal display unit 213 when the user designates the menu screen display button a10 on the add-in tab of the editing screen of the document creation software. be. The support menu screen A1 includes a correction support area a11 (indicated as "Toranomaki" in FIG. 6 and the like) and a data reflection area a12 (indicated as "Feedback" in FIG. 6 and the like), which are switched by tabs. In this example, it is assumed that the tab of the correction support area a11 is designated.

In the correction support area a11, the terminal display unit 213 displays a correction proposal display button a111, a correction unnecessary button a112, a template display button a113, a knowledge display button a114, and a support information display area a115.

The revision proposal display button a111 is input means for displaying a revision proposal on the user terminal 200. FIG. For example, when the user designates the wording of the contract and the revision proposal display button a111, the terminal reception unit 211 of the user terminal 200 receives the request for the revision proposal of the designated wording, and the reception unit 111 of the server device 100 Send this request to When the reception unit 111 of the server device 100 receives this request, the generation unit 113 generates a correction proposal for the specified wording and transmits it to the user terminal 200 . When the terminal reception unit 211 of the user terminal 200 receives this transmitted revision proposal, the terminal display unit 213 displays this revision proposal on the revision proposal display screen A2.

The template display button a113 (indicated as "model Article 22, paragraph 1" in FIG. 6, etc.) is used to display each clause displayed in the document creation software and the wording of the contract template corresponding to each clause. is an input means for For example, when the template display button a113 is designated by the user, the terminal reception unit 211 of the user terminal 200 receives a template display request, and the terminal display unit 213 stores the request in the terminal storage unit 230 or the storage unit 130 of the server device 100. Acquire the applicable contract template (in this example, the template of Article 22, Paragraph 1). The terminal display unit 213 displays the wording of the acquired contract template in the support information display area a115. The support information display area a115 is an area for displaying a contract template and knowledge information as information for supporting correction of the contract.

The knowledge display button a114 (indicated as "Toranomaki" in FIG. 6, etc.) is an input means for displaying knowledge information corresponding to each clause and each clause displayed in the document creation software. For example, when the knowledge display button a114 is designated by the user, the terminal reception unit 211 of the user terminal 200 receives this knowledge information display request, and the terminal display unit 213 receives the terminal storage unit 230 or the storage unit 130 of the server device 100. acquires the knowledge information stored in and displays it in the support information display area a115.

FIG. 7 is a diagram showing an example of a revision proposal display screen of the contract revision support system 1. As shown in FIG. As shown in FIG. 7, the correction proposal display screen A2 is displayed as a modal window by the terminal display unit 213 when the menu screen display button a10 is designated by the user. The revision proposal display screen A2 includes a pre-correction wording display area a20, a correction wording display area a21, and a revision proposal reflection button a22.

In the pre-correction wording display area a20, the terminal display unit 213 displays the contract wording specified by the user to request the revision proposal. In addition, in the correction wording display area a21, the terminal display unit 213 displays the wording of the generated correction proposal.

The revision proposal reflection button a22 is an input means for reflecting the revision proposal in the contract when the user adopts the revision proposal. For example, when the revision proposal reflection button a22 is specified by the user, the terminal reception unit 211 of the user terminal 200 receives the reflection request of this revision proposal, and the terminal document reflection unit 212 receives the contract document displayed by the document creation software. to reflect this amendment.

FIG. 8 is a diagram showing an example of a data reflection area of the contract correction support system 1. As shown in FIG. As shown in FIG. 8, the data reflection area a12 is displayed by the terminal display unit 213 when the tab is specified. The data reflection area a12 includes a reflection request button a121 and a reflection candidate list display area a122.

The reflection request button a121 is input means for reflecting the data set specified by the user in the learning data. For example, when the reflection request button a121 is designated by the user, the terminal reception unit 211 of the user terminal 200 receives a reflection request for the data set designated in the reflection candidate list display area a122, It is transmitted to the reflection unit 1121 . The data reflection unit 1121 of the server device 100 receives the transmitted data set and reflects it in the learning data stored in the learning data storage unit 133 .

The reflection candidate list display area a122 is an area for displaying a list of the second, fourth to sixth data sets that are candidates for reflection to the learning data. In other words, the reflection candidate list display area a122 is an area for displaying a list of contract wordings corresponding to the second, fourth to sixth data sets that are reflection candidates, and correction wordings by the user or the contract correction support system 1. be. The reflection candidate list display area a122 includes, for example, a first reflection candidate a122a, a second reflection candidate a122b, and a third reflection candidate a122c when there are three data set candidates to be reflected in the learning data. In the first reflection candidate a122a, for example, if the correction wording displayed in the first reflection candidate a122a is the correction wording corresponding to the second data set, the terminal display unit 213 displays this correction wording in the "before correction" The wording of the contract before amendment corresponding to is displayed, and this amended wording is displayed in "after amendment". When the check box of the first reflection candidate a122a is specified by the user and the reflection request button a121 is specified by the user, second data in which the wording of the contract before correction is input data and the correction wording is correct data The set is transmitted to the server device 100 by the terminal display unit 213 and reflected in the learning data.

<5. Operation example>
An operation example of the contract revision support system 1 according to the present embodiment will be described with reference to FIGS. FIG. 9 is a flow chart showing the flow of processing when learning a learning model, and a flow chart showing the flow of processing for displaying a correction proposal using the learned learning model. Note that the order of processing in the flow diagrams and sequence diagrams shown below is an example, and may be changed as appropriate.

As shown in FIG. 9A, the learning unit 112 of the server device 100 learns a learning model using learning data stored in the learning data storage unit 133 (S10). The learning unit 112 stores the learned learning model in the model storage unit 132 (S11).

As shown in FIG. 9B, the receiving unit 111 of the server device 100 receives the contract and the specification of the terms included in the contract from the user terminal 200 (S20). The generation unit 113 generates a correction proposal for the wording designated by the user terminal 200 using the learned learning model (S21). The display unit 114 causes the user terminal 200 to display the generated revision proposal (S22).

FIG. 10 shows the document creation software of the user terminal 200, the user terminal 200, and the user terminal 200 in the processing from receiving a request for a revision proposal for the wording of the contract from the user to reflecting the data set corresponding to the revision of the wording in the learning data. 2 is a sequence diagram showing the terminal program of , and the processing flow and interaction of the server apparatus 100. FIG.

As shown in FIG. 10, the terminal reception unit 211 of the terminal program 200b of the user terminal 200 (hereinafter also simply referred to as "terminal program 200b") receives a correction proposal display request from the user (S30). When receiving this display request, the terminal reception unit 211 acquires the text of the contract designated by the user (specifically, the character string of the text) by executing the document creation software 200a of the user terminal 200 (hereinafter simply referred to as (S31). Upon receiving this acquisition request, the document creation software 200a responds to the terminal program 200b with the terms of the contract specified by the user (S32). The terminal accepting unit 211 of the terminal program 200b accepts the contract wording and this contract that have been responded (S33).

The judging unit 214 of the terminal program 200b judges whether or not there is a correction in the wording of the accepted contract (S34). Specifically, the determination unit 214 determines whether or not the wording of the accepted contract matches the wording of the contract template corresponding to this wording. If the determination unit 214 determines that these wordings match, it determines that there is no correction, and if it determines that these words do not match, it determines that there is correction.

The user terminal 200 and the server device 100 execute the processing in the area indicated by the composite fragment opt1 (Option1) when the result of the determination by the determination unit 214 is that there is correction. Specifically, the terminal reception unit 211 of the terminal program 200b requests the server device 100 to make a correction proposal for the wording designated by the user (S35). When the reception unit 111 of the server device 100 receives the request for the correction proposal, the generation unit 113 of the server device 100 uses the learning model to generate a correction proposal for the specified wording (S36).

The transmission unit 121 of the server device 100 transmits the generated revision proposal (S37). When the terminal reception unit 211 of the user terminal 200 receives this correction proposal, the terminal display unit 213 displays this correction proposal on the user terminal 200 (S38).

The user terminal 200 executes the processing within the area indicated by the composite fragment opt2 (Option2) when the displayed revision proposal is adopted by the user. Specifically, the terminal reception unit 211 of the terminal program 200b receives from the user that the revision proposal is to be adopted (S39). The terminal document reflecting unit 212 of the terminal program 200b instructs the document creation software 200a to reflect this proposed revision in the displayed contract (S40). Upon receiving this reflection instruction, the document creation software 200a reflects the proposed amendment to the specified wording in the contract (S41).

The document creation software 200a receives from the user revised wording in which the wording of the contract is revised (S42). The language of this Agreement shall include the language of the proposed amendment as reflected above. The document creation software 200a links this corrected wording to the terminal program 200b (S43). The terminal accepting unit 211 of the terminal program 200b accepts the corrected wording by the linked user.

The terminal reception unit 211 of the terminal program 200b receives a display request for the reflection candidate list from the user (S45). When the terminal reception unit 211 receives this display request, the terminal display unit 213 displays, on the user terminal 200, a reflection candidate list including the wording of the adopted correction proposal and the correction wordings by the user (S46).

The terminal reception unit 211 of the terminal program 200b receives from the user a specification of a dataset to be reflected in the learning data, among the datasets corresponding to the phrases included in the displayed reflection candidate list (S47). The terminal transmission unit 221 of the terminal program 200b transmits the designated data set to the server apparatus 100, and requests reflection of this data set in the learning data (S48). When the reception unit 111 receives the reflection request and the dataset, the data reflection unit 1121 of the server device 100 reflects the dataset in the learning data (S49).

<6. Hardware configuration>
With reference to FIG. 11, an example of a hardware configuration in which the server device 100 and the user terminal 200 described above are implemented by a computer 800 will be described. The function of each device can also be implemented by being divided into a plurality of devices.

As shown in FIG. 11, a computer 800 includes a processor 801, a memory 803, a storage device 805, an input I/F section 807, a data I/F section 809, a communication I/F section 811, and a display device. 813.

Processor 801 controls various processes in computer 800 by executing programs stored in memory 803 . For example, the functional units included in the control unit 110 of the server device 100 and the terminal control unit 210 of the user terminal 200 execute a program temporarily stored in the memory 803 (for example, a terminal program in the case of the user terminal 200) by the processor 801. It can be realized by executing

The memory 803 is a storage medium such as a RAM (Random Access Memory). The memory 803 temporarily stores program codes of programs executed by the processor 801 and data necessary for executing the programs.

The storage device 805 is a non-volatile storage medium such as a hard disk drive (HDD) or flash memory. The storage device 805 stores an operating system, various models and various programs for realizing the above configurations. In addition, the storage device 805 can also store a table for registering learning data, knowledge information, user information, etc., and a DB for managing the table. Such programs and data are referred to by the processor 801 by being loaded into the memory 803 as necessary.

An input I/F unit 807 is a device for receiving input from the user. Specific examples of the input I/F unit 807 include a keyboard, mouse, touch panel, various sensors, wearable devices, and the like. The input I/F unit 807 may be connected to the computer 800 via an interface such as USB (Universal Serial Bus).

A data I/F unit 809 is a device for inputting data from outside the computer 800 . A specific example of the data I/F unit 809 is a drive device for reading data stored in various storage media. Data I/F unit 809 may be provided outside computer 800 . In that case, the data I/F unit 809 is connected to the computer 800 via an interface such as USB.

The communication I/F unit 811 is a device for performing data communication with a device external to the computer 800 via the Internet N by wire or wirelessly. Communication I/F unit 811 may be provided outside computer 800 . In that case, the communication I/F unit 811 is connected to the computer 800 via an interface such as USB.

The display device 813 is a device for displaying various information. Specific examples of the display device 813 include a liquid crystal display, an organic EL (Electro-Luminescence) display, and a wearable device display. The display device 813 may be provided outside the computer 800 . In that case, the display device 813 is connected to the computer 800 via, for example, a display cable. Further, when a touch panel is adopted as the input I/F section 807 , the display device 813 can be integrated with the input I/F section 807 .

It should be noted that the above-described embodiment is an example for explaining the present invention, and is not intended to limit the present invention only to the embodiment. Also, the present invention can be modified in various ways without departing from the gist thereof. Furthermore, those skilled in the art can adopt embodiments in which each element described above is replaced with equivalents, and such embodiments are also included in the scope of the present invention.

[Modification]
Although the present invention has been described based on the above embodiments, the following cases are also included in the present invention.

[Modification 1]
At least part of each configuration in the server device 100 according to the above embodiment may be provided in the user terminal 200 . For example, the learning unit 112 and the generation unit 113 of the server device 100 are implemented in the user terminal 200, and at least one of the dictionary storage unit 131, the model storage unit 132, and the learning data storage unit 133 of the server device 100 is installed in the terminal storage unit 230. May be implemented. Further, for example, all or part of the reception unit 111 of the server device 100 is connected to the terminal reception unit 211 of the user terminal 200, all or part of the document reflection unit 115 is connected to the terminal document reflection unit 212, and all or part of the display unit 114 is connected to the terminal reception unit 211 of the user terminal 200. A part may be mounted on the terminal display unit 213 .

For example, the user terminal 200 includes a terminal reception unit 211 that receives a contract and specification of wording contained in the contract from the user, and a corrected wording obtained by using the wording of the contract as input data and correcting the wording as correct data. A terminal storage unit 230 that stores a learning model learned using learning data including the first data set that (not shown), and a terminal display unit 213 for displaying the correction proposal.

According to the above configuration, the correction support service can be provided to the user by the user terminal 200 alone without using the server device 100 . Further, according to the above configuration, it is possible to provide the correction support service to the user even when the user terminal 200 is in an offline environment.

[Modification 2]
In the above embodiment, an example in which the model storage unit 132 that stores the learning model is provided in the storage unit 130 of the server device 100 has been described, but the location where the learning model is stored is not limited to this. The learning model may be stored, for example, in the storage unit of the external device. The generation unit 113 of the server device 100 may, for example, instruct the API provided by the external device to use the function of the learning model to generate a correction proposal using the learning model. good. The generation unit 113 may acquire a revision proposal from this external device as a response to this instruction.

[Modification 3]
Although not shown in the above embodiment, as shown in FIG. 12, a plurality of learning models may exist, and for example, may be constructed for each type of contract or each type of user. Furthermore, a learning model may be constructed for each organization to which the user belongs and for each business partner of the organization, for example. In addition to learning models, learning data may also be stored in DBs for each type of contract or for each type of user. For example, the terminal program 200b of the user terminal 200 requests a contract, a revision proposal including the specification of the terms of the contract, and a request to reflect the data set in the learning data. authentication information (for example, user ID, etc.) may be transmitted to the server device 100 . The server apparatus 100 may acquire the type of user by extracting corresponding user information from the user information stored in the storage unit 130 using the user authentication information as a search key. The generation unit 113 of the server device 100 may identify the learning model to which the correction plan is requested based on the type of the contract received as the specification of the contract and the type of the acquired user. Further, the data reflection unit 1121 of the server device 100 may specify the learning data DB to which the data set is to be reflected, based on the type of the received contract or the type of the acquired user.

[Modification 4]
In the above embodiment, an example was described in which the terminal reception unit 211 receives a request for a revision proposal when the user designates the wording included in the contract and the user designates the revision proposal display button a11. is not limited to this. For example, when the user designates the wording included in the contract, that is, the terminal reception unit 211 may receive the request for the revision proposal only by the user's designation of the wording (for example, designation by the cursor position).

[Modification 5]
In the above embodiment, the document creation function is realized by document creation software, and the correction support service is provided by adding a terminal program to this document creation software. It is not limited to this. For example, the terminal program may have all or part of the document creation function instead of the document creation software.

[Modification 6]
Although not shown in the above embodiment, when the reception unit 111 receives from the user terminal 200 the designated wording and the fact that correction is not required, the display unit 114 displays the designated wording of the contract as It may be displayed so as to be identifiable from the wording. The display unit 114 may, for example, change the mode of display of the text of the contract designated as not requiring correction. For example, the display unit 114 sets the character color or background color of the character string of the text of the contract designated as not requiring modification relative to the standard color (for example, black for the character color). The color may be changed to an inconspicuous color (for example, if the character color is gray, etc.).

REFERENCE SIGNS LIST 1 contract revision support system 100 server device 110 control unit 111 reception unit 112 learning unit 113 generation unit 114 display unit 115 document reflection unit 120 communication unit 130 Storage unit 200 User terminal 210 Terminal control unit 211 Terminal reception unit 212 Terminal document reflection unit 213 Terminal display unit 214 Determination unit 220 Terminal communication unit 230 Terminal storage unit , 800... Computer, 801... Processor, 803... Memory, 805... Storage device, 807... Input I/F section, 809... Data I/F section, 811... Communication I/F section, 813... Display device.

Claims (16)

a reception unit that receives a contract and specification of language contained in the contract from a user terminal used by a user;
Generation for generating a revision proposal for the specified wording using a learning model trained by learning data including a first data set in which the wording of the contract is input data and the revised wording obtained by correcting this wording is correct data Department and
A display unit that causes the user terminal to display the revision proposal,
Information processing equipment. The reception unit receives, from the user terminal, a revised wording obtained by modifying the wording of the contract,
Further comprising a first data reflection unit that reflects, in the learning data, a second data set in which the wording of the contract before correction corresponding to the corrected wording is input data and the corrected wording is correct data,
The information processing device according to claim 1 . The contract is a document created by modifying a contract template,
The learning data further includes a third data set in which the input data and the correct data are the same contract template wording,
The information processing apparatus according to claim 1 or 2. The reception unit receives, from the user terminal, a purport to adopt the displayed revision proposal,
a second data reflection unit for reflecting, in the learning data, a fourth data set in which the wording of the contract corresponding to the adopted revision proposal is input data and the wording of the revision proposal is correct data;
A document reflection unit that reflects the adopted amendment proposal in the contract,
The information processing apparatus according to any one of claims 1 to 3. The receiving unit receives, from the user terminal, designation of the wording of the contract and a request that the wording does not need to be modified,
Further comprising a third data reflection unit that reflects a fifth data set having the wording designated as not requiring correction as input data and correct data in the learning data,
The information processing apparatus according to any one of claims 1 to 4. The reception unit receives, from the user terminal, a correction wording obtained by correcting the wording of the displayed correction proposal,
Further comprising a fourth data reflection unit that reflects, in the learning data, a sixth data set in which the wording of the contract corresponding to the revision proposal is input data and the correction wording is correct data,
The information processing apparatus according to any one of claims 1 to 5. The display unit causes the user terminal to display corrected wording obtained by correcting the wording of the contract,
The receiving unit receives an instruction as to whether or not to reflect the second data set corresponding to the displayed corrected wording in the learning data from the user terminal,
The first data reflection unit reflects the second data set in the learning data based on the result of the instruction.
The information processing apparatus according to claim 2. The display unit causes the user terminal to display the wording of the adopted revision proposal,
The reception unit receives an instruction as to whether or not to reflect the fourth data set corresponding to the displayed revision proposal from the user terminal,
The second data reflection unit reflects the fourth data set in the learning data based on the result of the instruction.
The information processing apparatus according to claim 4. The information processing device further includes a calculation unit that calculates the confidence level of the correction plan using the learning model,
The display unit determines a mode of display of the revision proposal, including whether or not to display the revision proposal, based on the degree of certainty.
The information processing apparatus according to any one of claims 1 to 8. further comprising a learning unit that refers to a dictionary storage unit that stores a word dictionary in which one or more words are registered, and uses the learning data to learn the learning model;
The word dictionary is a single word registered for each of at least one of the type of the user and the type of the contract, and is a single word that is a combination of a plurality of words.
The information processing apparatus according to any one of claims 1 to 9. the computer
Receiving a contract and specification of language contained in the contract from the user terminal used by the user,
Using a learning model trained by learning data including a first data set in which the wording of the contract is input data and the corrected wording is correct data, generating a revision proposal for the specified wording,
causing the user terminal to display the revision proposal;
Information processing methods. on the user terminal used by the user,
a terminal reception function that receives from the user a contract and specification of language contained in the contract;
a terminal storage function for storing a learning model trained using learning data including a first data set in which a contract wording is input data and corrected wordings obtained by correcting this wording are correct data;
a generation function for inputting the specified wording into the learning model and generating a revision proposal for the wording;
realizing a terminal display function that displays the revision proposal;
terminal program. A server device communicably connected to a user terminal used by a user,
a reception function that receives a contract and specification of language contained in the contract from the user terminal;
A memory function that stores a learning model trained using learning data including a first data set that uses the wording of a contract as input data and correct data that is a corrected wording that is corrected from this wording;
a generation function for inputting the specified wording into the learning model and generating a revision proposal for the wording;
a display function for displaying the revision proposal on the user terminal;
server program. A contract revision support system including a user terminal used by a user and a server device communicably connected to the user terminal,
The user terminal is
a terminal reception unit that receives from the user a contract and specification of language contained in the contract;
a terminal transmission unit that transmits the specified wording to the server device when the terminal reception unit receives the contract and the designation of the wording,
The server device
a receiving unit that receives the transmitted text from the user terminal;
a model storage unit that stores a learning model learned by learning data including a first data set that uses the wording of the contract as input data and correct wording that corrects the wording;
a generation unit that inputs the received wording to the learning model and generates a revision proposal for the wording;
a transmitting unit that transmits the revision proposal to the user terminal,
The user terminal comprises a terminal receiving unit that receives the revision proposal, and a terminal display unit that displays the revision proposal.
Contract revision support system. The terminal reception unit receives from the user a designation of the wording and a designation that the wording does not need to be modified;
The terminal transmission unit does not transmit the wording designated as not requiring correction to the server device even when the terminal reception unit receives the designation of the wording.
A contract revision support system according to claim 14. The user terminal has a determination unit that refers to a terminal storage unit that stores a template of the contract and determines whether or not the specified wording matches the wording of the template corresponding to the wording. further prepared,
Even when the terminal reception unit receives the specification of the wording, the terminal transmission unit, if the determination unit determines that the specified wording matches the wording of the template, Do not send the specified wording to the server device;
A contract revision support system according to claim 15.

Images