Information processing device, information processing method, and program

Abstract

To support the development of user opinions. [Solution] The information processing device 2 includes a sentence acquisition unit 100 that acquires a first sentence relating to the user's opinion, an evaluation information acquisition unit 102 that acquires first evaluation information relating to the evaluation of the persuasiveness of the first sentence, and an output unit 104 that outputs information prompting the user to input a second sentence relating to their opinion based on the first evaluation information.

Description

[[Technical Field]]

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and a program. [[Background Art]]

[0002] Conventionally, techniques for evaluating the logicality of opinions provided by users are known (for example, Patent Document 1). [[Prior Art Documents]] [[Patent Documents]]

[0003] [[Patent Document 1]] Japanese Unexamined Patent Publication No. 2017-219989 [[Summary of the Invention]] [[Problem to be Solved by the Invention]]

[0004] However, conventional techniques have not been able to sufficiently support users in constructing their opinions.

[0005] The present disclosure provides an information processing apparatus, an information processing method, and a program capable of supporting users in constructing their opinions. [[Means for Solving the Problem]]

[0006] An information processing apparatus according to an aspect of the present disclosure includes: a sentence acquisition unit that acquires a first sentence relating to a user's opinion; an evaluation information acquisition unit that acquires first evaluation information relating to evaluation of the persuasiveness of the first sentence; and an output unit that outputs information prompting input of a second sentence relating to the user's opinion based on the first evaluation information.

[0007] An information processing method according to another aspect of the present disclosure is executed by a computer, the method including: acquiring a first sentence relating to a user's opinion; acquiring first evaluation information relating to evaluation of the persuasiveness of the first sentence; and outputting information prompting input of a second sentence relating to the user's opinion based on the first evaluation information.

[0008] A program relating to another aspect of this disclosure causes a computer to perform the following actions: obtain a first sentence relating to the user's opinion; obtain first evaluation information relating to an assessment of the persuasiveness of the first sentence; and output information prompting the user to input a second sentence relating to the user's opinion based on the first evaluation information. [Effects of the Invention]

[0009] This disclosure can help users form their own opinions. [Brief explanation of the drawing]

[0010] [Figure 1] This is a diagram illustrating the overview of System 1 according to this embodiment. [Figure 2] This diagram illustrates an example of the functional configuration of System 1 according to this embodiment. [Figure 3] This is a diagram illustrating an example of the operation of System 1 according to this embodiment. [Figure 4] This figure illustrates an example of the display screen of the terminal device 3 according to this embodiment. [Figure 5] This figure illustrates an example of the display screen of the terminal device 3 according to this embodiment. [Figure 6] This figure illustrates an example of the display screen of the terminal device 3 according to this embodiment. [Figure 7] This figure illustrates an example of the display screen of the terminal device 3 according to this embodiment. [Figure 8] This figure illustrates an example of the display screen of the terminal device 3 according to this embodiment. [Figure 9] This diagram illustrates an example of the hardware configuration of System 1 according to this embodiment. [Modes for carrying out the invention]

[0011] 1. Overview One of the exemplary objectives of System 1 according to this embodiment (hereinafter simply referred to as "System 1") is to support the user in constructing their opinion. In this embodiment, supporting the user in constructing their opinion includes the user learning and / or training in methods for constructing an opinion.

[0012] An opinion may, for example, consist of claims, reasons, and specific examples. In this embodiment, the claims, reasons, and specific examples included in the opinion may be collectively referred to as the "content of the opinion."

[0013] A claim is the core of an opinion and can also be called a conclusion. Reasons can be said to be logical explanations or evidence to support the claim. Specific examples can be said to be actual cases or situations that support the claim or reason. For example, an example of an opinion on the topic of "Should we eat breakfast every day or not?" is: "You should eat breakfast every day because regular meals promote metabolism and help maintain a healthy weight. Specifically, there are studies showing that people who habitually eat breakfast have a lower risk of obesity than those who do not." In this example, the part "You should eat breakfast every day" is an example of a claim. The part "Because regular meals promote metabolism and help maintain a healthy weight" is an example of a reason. And the part "There are studies showing that people who habitually eat breakfast have a lower risk of obesity than those who do not" is an example of a specific example. In this embodiment, "constructing an opinion" may mean creating an opinion on a certain topic that includes such claims, reasons, and specific examples.

[0014] It is not always easy to construct a well-developed opinion. For example, with respect to a certain theme, a situation can be assumed where a person constructing an opinion can intuitively come up with a claim, but cannot quickly come up with reasons and / or specific examples. In particular, such a situation is assumed to be particularly likely to occur when a person constructing an opinion must do so in a language different from their first language (for example, when the first language of the person constructing the opinion is Japanese and they must construct the opinion in English).

[0015] Debate is one type of training for constructing opinions. However, conventionally, conducting training for constructing opinions through debate has not always been easy, since there are cases where an opponent to debate with the person constructing the opinion and / or an instructor to evaluate and provide feedback on the content of the debate are required. Furthermore, since debate involves countering the opponent's opinion, the difficulty of constructing an opinion may be higher compared to cases where it is sufficient to simply state one's own opinion.

[0016] According to the system 1, the exemplary problems as described above can be solved. That is, the exemplary effects achieved by the system 1 can be at least one of: (1) assisting a user in constructing an opinion including a claim, reasons, and specific examples; (2) assisting in constructing an opinion in a language different from the user's first language; and (3) assisting in constructing an opinion assuming a debate scenario.

[0017] An overview of the operation of the system 1 will be described with reference to Figure 1. The system 1 includes a terminal device 3 used by a user, an information processing device 2 that executes at least part of processing for supporting the user in constructing an opinion, and an LCM server device 4 that hosts a Large Language Model (LLM).

[0018] First, the terminal device 3 accepts input of a first text relating to the user's opinion from the user, and transmits the first text to the information processing device 2 (S1). Note that the first text is an example of a first sentence.

[0019] Next, the information processing device 2 generates an instruction to cause evaluation of the persuasiveness of the first text, and transmits this instruction to the LLM server device 4 (S2). The LLM server device 4 evaluates the persuasiveness of the first text in accordance with the instruction based on the LLM hosted by itself, and transmits a response including the result to the information processing device 2 (S3).

[0020] This response includes evaluation information relating to the evaluation of the persuasiveness of the first text. The evaluation information may include information relating to the logicality, clarity, and the degree of sufficiency of specific examples in the first text. When the evaluations of the logicality, clarity, and the degree of sufficiency of specific examples of the first text are high, it can be said that the first text includes an assertion, a reason, and a specific example.

[0021] Next, based on this evaluation information, the information processing device 2 generates information prompting input of a second text relating to the user's opinion, and transmits the information to the terminal device 3 (S4). The second text is an example of a second sentence. In response to this, the terminal device 3 can accept input of the second text from the user.

[0022] The information prompting input of the second text is generated based on the evaluation information relating to the evaluation of the persuasiveness of the first text. Accordingly, by referring to the information prompting input of the second text, the user can reinforce the persuasiveness of the opinion stated in the first text. Thereby, the system 1 can support the construction of an opinion including an assertion, a reason, and a specific example by the user (see (1) of the above exemplary effect).

[0023] Note that in one embodiment, the terminal device 3 accepts input of the first text and the second text that are in a language other than the user's first language. Thereby, the system 1 can support the construction of an opinion in a language different from the first language (see (2) of the above exemplary effect).

[0024] In one embodiment, the information prompting the input of the second text may include opposing opinions to the first text. With this configuration, System 1 can assist in constructing opinions in anticipation of a debate (see (3) of the exemplary effects above).

[0025] As will be explained in more detail later, System 1 can repeatedly perform the same processing as steps S1 to S4 described above for the first text, but for the nth text (where n is a natural number from 1 to N-1). With this configuration, the user is prompted to input the (n+1th)th text based on the evaluation of the persuasiveness of the nth text, thus repeatedly reinforcing the persuasiveness of the user's opinion.

[0026] In one embodiment, the logical validity of an opinion may be evaluated based on whether the reasons adequately support the claim, and whether the claim and reasons are adequately supported by specific examples. In another embodiment, the clarity of an opinion may be evaluated based on whether the claim, reasons, and specific examples are consistent, and whether the content is concise and contains few redundant expressions.

[0027] In this embodiment, if a word prefixed with "1st" to "Nth" is not particularly distinguished from "1st" to "Nth," or if they are referred to collectively, the prefix may be omitted. For example, if the first sentence and the second sentence are not particularly distinguished, or if they are referred to collectively, they may simply be referred to as "sentence."

[0028] 2. Functional Configuration Referring to Figure 2, the functional configuration of System 1 in this embodiment will be described. System 1 includes an information processing device 2, a terminal device 3, an LLM server device 4, and a communication network 5. The information processing device 2, the terminal device 3, and the LLM server device 4 are configured to communicate with each other via the communication network 5.

[0029] 2.1 Information Processing Device 2 The information processing device 2 performs at least part of the processing that supports the user in constructing their opinion. In one embodiment, the information processing device 2 is a server device when the terminal device 3 is a client device. In one embodiment, the information processing device 2 is a cloud server device. The information processing device 2 may be a device that includes, for example, one or more virtual or physical web server devices and one or more virtual or physical database server devices.

[0030] The information processing device 2 comprises a control unit 10, a storage unit 12, a network interface unit 14, and a bus 16. The control unit 10, the storage unit 12, and the network interface unit 14 are electrically connected via the bus 16.

[0031] 2.1.1 Control Unit 10 The control unit 10 functions as a text acquisition unit 100, an evaluation information acquisition unit 102, an output unit 104, a generation unit 106, a determination unit 108, and a specification unit 110 by executing various programs stored in the storage unit 12, which will be described later. The following describes the various configurations that the information processing device 2 can realize based on these functions.

[0032] The sentence acquisition unit 100 acquires the first sentence regarding the user's opinion. The evaluation information acquisition unit 102 acquires first evaluation information regarding the evaluation of the persuasiveness of the first sentence. The output unit 104 outputs information prompting the user to input a second sentence regarding their opinion, based on the first evaluation information. The information prompting the user to input a second sentence may include at least a part of the first evaluation information.

[0033] In this embodiment, "sentence" may be a group of one or more words that have a complete meaning (which can also be called a "sentence" in the narrow sense), but is not limited to this. In this embodiment, "sentence" may be, for example, a word, a group of words, a sentence in the narrow sense, a paragraph, or a text.

[0034] In one embodiment, the sentence may be text and / or audio. In one example, the sentence acquisition unit 100 acquires text entered by the user into the terminal device 3 as a sentence. In another example, the sentence acquisition unit 100 acquires audio entered by the user via the microphone or the like of the terminal device 3 as a sentence. In this embodiment, the acquisition of a sentence by the sentence acquisition unit 100 may include acquiring data corresponding to that sentence (for example, text data and audio data).

[0035] In one embodiment, the text relating to the user's opinion may include, for example, the content that the user might argue in a debate, the content of an email the user creates, and the content that the user might discuss in an interview during job hunting.

[0036] In one embodiment, the evaluation information acquisition unit 102 may acquire evaluation information based on natural language processing techniques that can be appropriately selected by a person skilled in the art.

[0037] In one example, the evaluation information acquisition unit 102 can acquire evaluation information based on a machine learning model (hereinafter referred to as the "persuasiveness evaluation model") which is constructed by inputting training data in which an index indicating the persuasiveness of a sentence is labeled (annotated) to the values ​​obtained by vectorizing a sentence using doc2vec and / or word2vec, etc. The evaluation information acquisition unit 102 can acquire an index indicating the persuasiveness of a sentence (an example of evaluation information) by inputting the values ​​obtained by vectorizing the sentence to be evaluated into the persuasiveness evaluation model.

[0038] In another example, the evaluation information acquisition unit 102 may acquire evaluation information based on a machine learning model (hereinafter, this machine learning model will be referred to as the "multiple evaluation criteria model," and together with the "persuasiveness evaluation model" described above, it will be referred to as the "evaluation model") which is constructed by inputting learning data into which values ​​obtained by vectorizing sentences using, for example, doc2vec and / or word2vec, are labeled (annotated) with an index indicating the logic of the sentence, an index indicating the clarity of the sentence, and an index indicating the degree of completeness of the specific examples of the sentence. The evaluation information acquisition unit 102 can acquire an index indicating the logic of the sentence, an index indicating the clarity of the sentence, and an index indicating the degree of completeness of the specific examples of the sentence (these three indicators are an example of evaluation information) by inputting the values ​​obtained by vectorizing the sentence to be evaluated into the multiple evaluation criteria model (an example of evaluating a sentence using one or more evaluation criteria).

[0039] In one embodiment, the evaluation information acquisition unit 102 acquires evaluation information by inputting an instruction (hereinafter referred to as "evaluation instruction") to the LLM to evaluate the persuasiveness of the sentence. Inputting an evaluation instruction to the LLM includes sending a request (in one example, an HTTP request) containing the evaluation instruction to the LLM server device 4. Acquiring evaluation information also includes receiving a response (in one example, an HTTP response) corresponding to the request from the LLM server device 4.

[0040] In one embodiment, the evaluation instruction may include an instruction to evaluate at least one of the following: the logicality of the sentence, the clarity of the sentence, and the completeness of the specific examples. The evaluation instruction may also include an instruction to select an evaluation criterion from among the logicality of the sentence, the clarity of the sentence, and the completeness of the specific examples that is evaluated relatively lower compared to the other evaluation criteria.

[0041] In one embodiment, the output unit 104 outputting information prompting the input of a second sentence based on evaluation information includes inputting an instruction to generate information prompting the input of a second sentence based on evaluation information (hereinafter referred to as a "comment generation instruction") to the LLM, and outputting the response obtained thereby.

[0042] In one embodiment, the first evaluation information is determined by evaluating the first sentence using one or more evaluation criteria, and the information prompting input of the second sentence includes information prompting input regarding an evaluation criterion that is poorly rated compared to other evaluation criteria. In one embodiment, the one or more evaluation criteria include at least one of logicality, clarity, and the degree of completeness of specific examples.

[0043] Among one or more evaluation criteria, a criterion that receives a low rating compared to others may be a bottleneck in improving the persuasiveness of an opinion. Therefore, prompting input regarding the low-rated criterion can more efficiently support users in constructing their opinions.

[0044] In one embodiment, among one or more evaluation criteria, an evaluation criterion that receives a lower evaluation compared to other evaluation criteria can be identified based on a multi-evaluation criterion model. For example, by inputting the vectorized values ​​of the sentence to be evaluated into a multi-evaluation criterion model, and comparing the relative magnitudes of an index indicating the logicality of the sentence, an index indicating the clarity of the sentence, and an index indicating the richness of the specific examples of the sentence, an evaluation criterion that receives a lower evaluation compared to other evaluation criteria among logicality, clarity, and richness of specific examples (an example of one or more evaluation criteria) can be identified. For example, if the evaluation level corresponds to the magnitude of the index, and when the vectorized values ​​of the sentence to be evaluated are input into a multi-evaluation criterion model, and the index indicating the clarity of the sentence is "100", the index indicating the logicality of the sentence is "50", and the index indicating the richness of the specific examples of the sentence is "30", then the evaluation criterion with a relatively low evaluation can be identified as the richness of specific examples, which has a relatively small index.

[0045] For example, if, in the first sentence, the evaluation of the degree of detail in the example is lower than that of logic and clarity (an example of other evaluation criteria) among logic, clarity, and the degree of detail in the example, the information prompting input in the second sentence may include information prompting input of an example. In this case, the degree of detail in the example corresponds to "an example of an evaluation criterion that is evaluated lower than that of other evaluation criteria."

[0046] In another example, if, in the first sentence, the degree of logical coherence is rated lower than the degree of clarity and the degree of specific examples (an example of other evaluation criteria), the information prompting input for the second sentence may include information pointing out a logical leap in the content of the opinion in the first sentence. In this case, logical coherence corresponds to an example of "an evaluation criterion that is rated lower than other evaluation criteria among one or more evaluation criteria."

[0047] In one embodiment, the sentence acquisition unit 100 further acquires a second sentence, and the evaluation information acquisition unit 102 further acquires second evaluation information regarding the evaluation of the persuasiveness of the second sentence. In one embodiment, the output unit 104 further outputs information prompting the user to input a third sentence regarding their opinion, based on the second evaluation information.

[0048] In one embodiment, the sentence acquisition unit 100 further acquires a second sentence, and the output unit 104 further outputs information prompting the input of a summary sentence that summarizes the first and second sentences.

[0049] Prompting users to input a concluding sentence can serve as training for them in constructing persuasive opinions. For example, if the first sentence has sufficient specific examples but is relatively lacking in logic, and the second sentence then reinforces the logic, prompting the user to input a concluding sentence for both the first and second sentences can serve as training for constructing an opinion that is sufficiently logical and includes sufficient specific examples.

[0050] In one embodiment, the text acquisition unit 100 further acquires a summary text, the information processing device 2 further includes a generation unit 106 that generates a revised text reflecting modifications and / or additions to the summary text, and the output unit 104 further outputs the revised text. With this configuration, the user can specifically identify areas that need further improvement in the summary text.

[0051] In one embodiment, the generation unit 106 generates a revised text based on a natural language processing algorithm that can be arbitrarily selected by those skilled in the art.

[0052] In one embodiment, the generation unit 106 generates a revised text by inputting instructions (hereinafter referred to as "revision instructions") to the LLM for modifying and / or adding to the summary text. The generation of a revised text by the generation unit 106 by inputting revision instructions to the LLM includes sending a request including the revision instructions to the LLM server device 4 and receiving a response from the LLM server device 4.

[0053] In one embodiment, the correction instructions may include instructions regarding the relationship between the summary sentence and the corrected sentence. Instructions regarding the relationship between the summary sentence and the corrected sentence may include, for example, instructions not to correct the summary sentence too much, instructions not to use words in the corrected sentence that are not used in the summary sentence, instructions to estimate the skill level and / or experience of the user who entered the summary sentence and generate a corrected sentence that corresponds to that skill level and / or experience, and instructions to generate a corrected sentence that maintains the logical flow of the summary sentence as much as possible. If the corrected sentence is too different from the summary sentence, the user may find it difficult to refer to the corrected sentence.

[0054] In one embodiment, the information prompting the user to input the second sentence includes opposing viewpoints to the first sentence. With this configuration, the user inputs the second sentence taking into account the opposing viewpoints to the first sentence. This helps the user construct an opinion that incorporates more multifaceted perspectives.

[0055] In one embodiment, the sentence acquisition unit 100 further acquires a fourth sentence input by the user, and the information processing device 2 further includes a determination unit 108 that determines whether the fourth sentence satisfies predetermined conditions for a hint request. If the fourth sentence satisfies the predetermined conditions, the output unit 104 further outputs hints for creating a persuasive sentence based on the fourth sentence and / or the first sentence. If the fourth sentence does not satisfy the predetermined conditions, the evaluation information acquisition unit 102 further acquires other evaluation information regarding the evaluation of the persuasiveness of the fourth sentence.

[0056] For example, if the sentence entered by the user in response to information prompting the input of a second sentence (an example of a fourth sentence) contains the text "I don't know," the determination unit 108 may determine that the sentence satisfies predetermined conditions for requesting a hint. In this case, the output unit 104 may output hints for creating a persuasive sentence based on the sentence (an example of a fourth sentence) and / or sentences entered by the user up to that point (an example of a first sentence).

[0057] In another example, if the sentence entered by the user in response to information prompting the input of a second sentence (an example of a fourth sentence) includes text indicating the reason for the opinion, the determination unit 108 may determine that the sentence does not satisfy the predetermined conditions for requesting a hint. In this case, the evaluation information acquisition unit 102 may acquire evaluation information corresponding to the sentence (an example of other evaluation information).

[0058] In one embodiment, the determination unit 108's determination of whether a sentence satisfies predetermined conditions relating to a hint request includes determining whether at least a part of the sentence is consistent with one or more words relating to a hint request (e.g., "I don't know" and "hint," etc.). In another embodiment, the determination unit 108's determination of whether a sentence satisfies predetermined conditions relating to a hint request includes determining whether at least a part of the sentence has predetermined characteristics relating to the necessity of a hint (e.g., whether there are errors in vocabulary and / or grammar, etc.).

[0059] In one embodiment, the determination unit 108's determination of whether a sentence satisfies predetermined conditions relating to a hint request includes inputting an instruction (hereinafter referred to as a "determination instruction") to the LLM that causes the LLM to determine whether the sentence satisfies the predetermined conditions, and obtaining the response.

[0060] In one embodiment, the output unit 104 outputting hints for creating a persuasive sentence includes outputting a standard hint corresponding to the fourth sentence and / or the first sentence. The correspondence between the fourth sentence and / or the first sentence and the standard hint can be stored in the storage unit 12 in advance as setting information, for example. In another embodiment, the output unit 104 outputting hints for creating a persuasive sentence includes outputting hints corresponding to predetermined characteristics regarding the necessity of hints. For example, if the sentence contains errors in vocabulary and / or grammar, the output unit 104 may output a hint such as, "So you were trying to say ○○."

[0061] In one embodiment, the output unit 104 outputting hints for creating a persuasive sentence includes inputting an instruction to generate hints based on the fourth sentence and / or the first sentence (hereinafter referred to as the "hint generation instruction") to the LLM and outputting the response obtained thereby.

[0062] In one embodiment, the information processing device 2 further includes a specification unit 110 that identifies evaluation criteria for enhancing the persuasiveness of a sentence from one or more evaluation criteria related to sentence evaluation, and the hints for creating a persuasive sentence output by the output unit 104 are determined based on the evaluation criteria identified by the specification unit 110.

[0063] For example, if the identification unit 110 identifies that, with respect to a particular sentence, the logicalness of the sentence is relatively insufficient compared to the other evaluation criteria (one or more evaluation criteria), the output unit 104 may output hints to reinforce that logicalness.

[0064] In one embodiment, the identification unit 110 identifies evaluation criteria for enhancing the persuasiveness of a sentence from one or more evaluation criteria based on a multiple evaluation criteria model. For example, the identification unit 110 can input vectorized values ​​of the sentence to be evaluated into a multiple evaluation criteria model and identify evaluation criteria for enhancing the persuasiveness of a sentence based on the relative magnitudes of indicators such as logicality, clarity, and the degree of completeness of specific examples obtained therefrom.

[0065] In one embodiment, the identification unit 110 can identify an evaluation criterion for enhancing the persuasiveness of a sentence by inputting an instruction (hereinafter referred to as "identification instruction") to the LLM that causes the LLM to identify an evaluation criterion for enhancing the persuasiveness of a sentence from among one or more evaluation criteria.

[0066] 2.1.2 Storage section 12 The storage unit 12 stores various information necessary for the operation of the information processing device 2. In one embodiment, the storage unit 12 stores the program to be executed by the control unit 10.

[0067] 2.1.3 Network Interface Unit 14 The network interface unit 14 enables communication with other devices via the communication network 5.

[0068] 2.2 Terminal device 3 Terminal device 3 is a communication device used by the user. Examples of terminal devices 3 include smartphones, personal computers, tablet devices, and wearable devices. Terminal device 3 includes an input interface, an output interface, and a communication interface.

[0069] The input interface is an interface that allows terminal device 3 to receive input from the user. The input interface may be a touch panel, microphone, camera, keyboard, mouse, etc.

[0070] The output interface is an interface for transmitting information to the user through images, sound, etc. The output interface includes a display (which may also function as a touch panel) and speakers.

[0071] The communication interface is an interface for enabling communication with other devices via the communication network 5. The communication interface may be a wireless communication interface or a wired communication interface.

[0072] Terminal device 3 may access services provided by information processing device 2, for example, via a web browser, or it may access such services by installing dedicated software.

[0073] Terminal device 3 may accept the user's specification of the language of the text to be entered. In this case, the user may specify a language different from their first language.

[0074] 2.3 LLM Server Device 4 The LLM server device 4 is a server device that hosts the LLM on its own device. The LLM server device 4 receives requests such as HTTP requests via an API (Application Programming Interface), executes processing according to the request, and returns a response such as an HTTP response. The LLM may be, for example, a deep learning model with hundreds of millions of parameters and trained on hundreds of gigabytes or more of natural language training data. An example of an LLM is gpt-4o.

[0075] 2.4 Communication Network 5 Communication network 5 enables communication between each device included in system 1. Communication network 5 enables communication between devices based, for example, on the TCP / IP protocol.

[0076] 3 operations Refer to Figures 3-8 to explain an example of the operation of System 1. Note that in the following example, the sentences are assumed to be text.

[0077] 3.1 Sequence Figure 3 is a sequence diagram showing an example of the operation of System 1.

[0078] First, the information processing device 2 transmits information about the theme on which the user will formulate their opinion to the terminal device 3 (S100). The information about the theme may be, for example, pre-stored in the storage unit 12, or it may be information that the information processing device 2 has received in advance from the LLM server device 4.

[0079] Next, System 1 repeatedly performs the process described in steps S102 to S114 below for each of the natural numbers n=1 to N-1.

[0080] First, terminal device 3 receives the nth text input from the user (S102). Next, terminal device 3 transmits the nth text to information processing device 2 (S104). Receiving the nth text is one example of obtaining a sentence relating to the user's opinion.

[0081] The information processing device 2 generates the nth instruction, which is an instruction corresponding to the nth text, based on the nth text received (S106). The nth instruction may include, for example, (1) to (11) below. The nth response is the response obtained by inputting the nth instruction to the LLM server device. (1) The nth text (2) Text 1 to Text (n-1) (3) 1st instruction ~ n-1st instruction (4) First response to the (n-1)th response (5) Information prompting input of the second text ~ Information prompting input of the nth text (6) General instructions Example: "The nth text, which is the subject of persuasiveness evaluation, as well as the first to (n-1)th texts, the first instruction to (n-1)th instruction, the first response to (n-1)th response, and the information prompting input of the second text to the nth text, are as described above. Based on this, please process according to the following instructions." (7) Judgment instructions Example: "Determine whether the nth text requires a hint." (8) Hint generation instructions Example: "If text n is determined to be requesting a hint, generate a hint to improve one of the following aspects of the user's opinion: its logical coherence, clarity, or the completeness of its specific examples, based on the content of texts 1 through (n-1). If text n is determined not to be requesting a hint, proceed according to the following instructions." (9) Evaluation Instructions Example: "Evaluate the logic, clarity, and richness of specific examples in Text No. n using a scale of 1 to 10. Then, evaluate the overall logic, clarity, and richness of specific examples across Texts No. 1 through No. n using a scale of 1 to 10." (10)Specific instructions Example: "Identify the evaluation criterion with the lowest metric among the logical consistency, clarity, and level of detail in texts 1 through n." (11) Comment generation instructions Example: "Generate and output comments to improve the evaluation criterion with the lowest metric across texts 1 through n."

[0082] Sending instructions including the above evaluation instructions to the LLM server device 4 is an example of determining evaluation information by evaluating a sentence using one or more evaluation criteria.

[0083] The information processing device 2 transmits the generated nth instruction to the LLM server device 4 (S108). Next, the information processing device 2 receives the nth response generated by the LLM server device 4 based on the nth instruction (S110). Receiving the nth response is one example of obtaining evaluation information regarding the evaluation of the persuasiveness of the sentence.

[0084] Next, the information processing device 2 generates information prompting the input of the (n+1)th text based on the received nth response (S112). The information prompting the input of the (n+1)th text may include a comment generated based on the comment generation instruction of the nth instruction.

[0085] Next, the information processing device 2 transmits information prompting the input of the (n+1)th text to the terminal device 3 (S114). Generating and transmitting information prompting the input of the (n+1)th text based on the nth response is an example of outputting information prompting the input of a sentence about the user's opinion based on evaluation information.

[0086] As System 1 repeatedly executes steps S102 to S114 for each of n=1 to N-1, the user inputs the first text to the (N-1)th text into Terminal Device 3, and Information Processing Device 2 transmits information prompting the input of the second text to information prompting the input of the Nth text to Terminal Device 3.

[0087] Subsequently, terminal device 3 receives input of the Nth text from the user (S116). Next, terminal device 3 transmits the Nth text to information processing device 2 (S118).

[0088] In response, the information processing device 2 generates information prompting the input of a summary text that combines the first to Nth texts (S120). The information processing device 2 transmits the information prompting the input of the summary text to the terminal device 3 (S122).

[0089] In response, terminal device 3 accepts the input of the summary text (S124). Next, terminal device 3 transmits the summary text to information processing device 2 (S126).

[0090] The information processing device 2 generates instructions corresponding to the received summary text (S128). The instructions corresponding to the summary text may include, for example, the following (1) to (9). (1) Summary Text (2) Texts 1 to N (3) 1st instruction ~ N-1st instruction (4) 1st response to N-1st response (5) Information prompting input of the second text ~ Information prompting input of the Nth text (6) General instructions Example: "The summary text, which is the subject of persuasiveness evaluation, as well as the first to nth texts, first to n-th instruction, first response to n-th response, and information prompting input of the second text to the nth text, related to past interactions with the user are as described above. Based on these, please process according to the following instructions." (7) Evaluation Instructions Example: "Please evaluate the logical consistency, clarity, and completeness of the examples in the summary text using a scale of 1 to 10." (8) Correction instructions Example: "Please generate a revised version that strengthens the persuasiveness of the summary text. When doing so, please be careful not to over-revise the summary text, as we want the revised version to be easily accessible to users." (9) Comment generation instructions Example: "Generate and output comments to reinforce the persuasiveness of the concluding comments, and / or comments regarding the intent of the revised text. If there is any content included in the concluding comments that is not included in the first to nth texts, please point it out."

[0091] The information processing device 2 sends an instruction corresponding to the generated summary text to the LLM server device 4 (S130). Next, the information processing device 2 receives the response generated by the LLM server device 4 based on the instruction corresponding to the summary text (S132). The information processing device 2 sends the summary evaluation information, including the response, to the terminal device 3 (S134).

[0092] 3.2 Display screen example Refer to Figures 4-8 to describe examples of display screens in terminal device 3. Note that, in the following, the language in which terminal device 3 accepts input and / or displays information may be a language specified by the user, at least for some display screens. The language specified by the user may be different from the user's first language (for example, English if the user's first language is Japanese).

[0093] Furthermore, the following describes an example where N in Figure 3 is "6" (i.e., an example where the user is asked to input the first to sixth texts and a summary sentence for them).

[0094] The example screen in Figure 4 shows text fields t100 to t108, an input field d100, and a voice input button d102. The input field d100 is a display element for the user to input text, for example, via the keyboard provided by the terminal device 3 (including a virtual keyboard displayed on the screen). The voice input button d102 is a display element for the user to input voice, for example, via the microphone provided by the terminal device 3, and input the transcribed text into the input field d100.

[0095] Text t100 contains information related to the topic and may be displayed on terminal device 3 immediately after step S100 in Figure 3. Text t100 reads, "Do you think a two-day weekend is enough? Let us know your opinion."

[0096] Text t102 may be the text initially entered by the user. Text t102 may be displayed on terminal device 3 immediately after step S104 in n=1 in Figure 3. Text t102 displays "No, I don't think so." This may correspond to the assertion among the assertion, reason, and specific example of the user's opinion.

[0097] Text t104 may be displayed based on the response corresponding to text t102. Text t104 may be displayed in terminal device 3 immediately after step S114 in n=1 in Figure 3. Text t104 displays, "That's an interesting opinion! Could you tell me why you think a two-day weekend isn't enough?" This is an example of information that prompts the input of the following text.

[0098] Text t106 may be text entered by the user after text t102. Text t106 may be displayed on terminal device 3 immediately after step S104 in n=2 in Figure 3. Text t106 displays, "Because I think we need to improve the balance between school time and free time." This may correspond to the reason among the user's assertion of opinion, reason, and specific example.

[0099] Text t108 may be displayed based on the response corresponding to text t106. Text t106 may be displayed on terminal device 3 immediately after step S114 in n=2 in Figure 3. Text t108 displays, "That's a compelling argument. Do you have any specific examples that illustrate the need to improve the balance between school time and free time?" This is an example of information that prompts the input of the next text.

[0100] The example screen in Figure 5 shows a continuation of the example screen in Figure 4. In addition to the input field d100 and voice input button d102 described above, the example screen in Figure 5 also displays text t200~t206.

[0101] Text t200 may be text entered by the user after text t106 in Figure 4. Text t200 may be displayed on terminal device 3 immediately after step S104 in n=3 in Figure 3. Text t200 displays, "Having free time allows me to use it for hobbies, which improves my study efficiency." This may correspond to a specific example among the user's opinion, statement, and reason. Text t200 may also have an aspect as a reason for the user's opinion.

[0102] Text t202 may be displayed based on the response corresponding to text t200. Text t202 may be displayed on terminal device 3 immediately after step S114 in n=3 in Figure 3. Text t202 displays, "Could you tell me the relationship between having time for hobbies and improved study efficiency?" This is an example of information that prompts the input of the next text, and also an example of information that prompts input regarding one or more evaluation criteria, specifically an evaluation criterion that is rated lower than other evaluation criteria (in this example, logical thinking).

[0103] Text t204 may be text entered by the user after text t200. Text t204 may be displayed on terminal device 3 immediately after step S104 in n=4 in Figure 3. Text t204 displays "Please give me a hint." This is an example of a sentence that satisfies predetermined conditions regarding a hint request.

[0104] Text t206 may be displayed based on the response corresponding to text t204. Text t206 may be displayed in terminal device 3 immediately after step S114 in n=4 in Figure 3. Text t204 displays, "Of course. Let's think about how relaxing can help you concentrate on studying. Let's think about how having time for hobbies can help you recharge and stay motivated to study." This is an example of a hint for creating a persuasive sentence that is output based on a sentence already entered by the user.

[0105] The example screen in Figure 6 shows a continuation of the example screen in Figure 5. In addition to the input field d100 and voice input button d102 described above, the example screen in Figure 6 displays text t300~t306.

[0106] Text t300 may be text entered by the user after text t204 in Figure 5. Text t300 may be displayed on terminal device 3 immediately after step S104 in n=5 in Figure 3. Text t300 displays, "Being able to use free time for hobbies helps me relax, and as a result, I can concentrate on studying." This may correspond to a specific example among the user's opinion, statement, and reason. Text t300 may also have an aspect as a reason for the user's opinion.

[0107] Text t302 may be displayed based on the response corresponding to text t300. Text t302 may be displayed in terminal device 3 immediately after step S114 in n=5 in Figure 3. Text t302 displays, "Great logic! I disagree slightly. I think a two-day weekend would make it easier to maintain a routine. What do you think?" This is an example of information prompting the input of the next text, as well as an example of a counter-argument to the statement.

[0108] Text t304 may be text entered by the user after text t300. Text t304 may be displayed on terminal device 3 immediately after step S116 in Figure 3. Text t304 displays, "For example, hobbies that require a significant amount of time, such as camping or traveling, are not adequately handled with only two days off per week." This may correspond to the specific example among the user's assertion of opinion, reasons, and specific examples.

[0109] Text t306 displays the message, "Please summarize the opinions discussed so far." Text t306 may be displayed on terminal device 3 immediately after step S122 in Figure 3. This is an example of information prompting the user to enter a summary sentence that summarizes the multiple sentences already entered.

[0110] The example screen in Figure 7 shows a continuation of the example screen in Figure 6. In addition to the input field d100 and voice input button d102 described above, the example screen in Figure 7 also displays text t400.

[0111] Text t400 displays the following: "I believe a two-day weekend is insufficient because I think we need to improve the balance between school time and free time. This is especially true for people with hobbies that require a significant amount of time, such as camping or traveling, who cannot adequately balance these activities. Increasing free time allows for more time to be spent on hobbies and promotes relaxation. And relaxation, in turn, improves study efficiency." Text t400 may be displayed on terminal device 3 immediately after step S126 in Figure 3. This is an example of a concluding statement.

[0112] The example screen in Figure 8 shows a continuation of the example screen in Figure 7. In the example screen in Figure 7, the text t500~t506 is displayed.

[0113] Text t500 displays the following: "That's a great point! Your reasons and examples strongly support your argument. Below are some examples that further strengthen the persuasiveness of your argument. Please refer to them." This is an example of evaluation information corresponding to a concluding sentence.

[0114] Text t502 displays the following: "I believe a two-day weekend is insufficient because we need to improve the balance between school time and free time. In particular, people with hobbies that require a significant amount of time, such as camping or traveling, cannot secure enough time for these activities on weekends alone. For example, when going camping, the actual activity time is limited when you include preparation and cleanup. By having more free time, you can spend more time on such hobbies and get sufficient relaxation. As a result, you can approach studying in a refreshed state, which is thought to lead to improved learning efficiency." This is an example of a revised sentence.

[0115] Text t504 states, "An example of not having enough free time when going camping has been added." Text t506 states, "The causal relationship between relaxation and improved study efficiency has been clarified." These are examples of information regarding the differences between the summary sentence and the revised sentence.

[0116] Although not shown in the diagram, terminal device 3 may also display further examples of expressions not used in the revised text. This configuration allows for efficient support of the user's opinion formation, even if, for example, the intent of the revised text differs from what the user intended, or if the revised text does not include expressions that the user can and / or wants to use.

[0117] 4 Hardware Configuration Referring to Figure 9, an example of a hardware configuration when the devices included in System 1 described above are implemented by computer 70 will be explained. Note that the functions of each device can also be implemented by dividing them among multiple devices.

[0118] As shown in Figure 9, the computer 70 includes a processor 700, a storage device 702, an input interface 704, a data interface 706, a communication interface 708, and a display device 710.

[0119] The processor 700 controls various processes in the computer 70 by executing programs stored in the storage device 702. For example, each functional unit of the control unit 10 of the information processing device 2 can be realized by the processor 700 executing programs stored in the storage device 702.

[0120] The storage device 702 is a storage medium such as RAM (Random Access Memory). RAM temporarily stores the program code of the program executed by the processor 700, as well as data required during program execution.

[0121] The storage device 702 can also be a non-volatile storage medium such as a hard disk drive (HDD) or flash memory. The storage device 702 stores the operating system and various programs for realizing the above configurations. The storage medium storing these various programs may be a non-transitory computer-readable medium. In addition, the storage device 702 can also store tables for registering various information and a database for managing those tables. Such programs and data are loaded into the storage device 702 as needed and accessed by the processor 700.

[0122] The input interface 704 is a device for receiving input from the user. Specific examples of the input interface 704 include cameras, buttons, microphones, keyboards, mice, touch panels, various sensors, and wearable devices. The input interface 704 may be connected to the computer 70 via an interface such as USB (Universal Serial Bus).

[0123] The data interface 706 is a device for inputting data from outside the computer 70. Specific examples of the data interface 706 include drive devices for reading data stored on various storage media. The data interface 706 may also be located outside the computer 70. In that case, the data interface 706 would be connected to the computer 70 via an interface such as USB.

[0124] The communication interface 708 is a device for performing data communication with external devices of the computer 70 via a communication network 5, either wired or wirelessly. The communication interface 708 may also be located outside the computer 70. In that case, the communication interface 708 would be connected to the computer 70 via an interface such as USB.

[0125] The display device 710 is a device for displaying various types of information. Specific examples of the display device 710 include liquid crystal displays, organic EL (Electro-Luminescence) displays, and displays for wearable devices. The display device 710 may be located outside the computer 70. In that case, the display device 710 is connected to the computer 70 via, for example, a display cable. Furthermore, if a touch panel is used as the input I / F 704, the display device 710 can be configured as an integrated unit with the input I / F 704.

[0126] Furthermore, the components of the device included in the System 1 described above are such that a program stored in the storage device 702 is executed by the processor 700, thereby realizing a defined process in cooperation with other hardware. In other words, these components are conceived as both software or firmware, and as corresponding hardware, and in both concepts, they are also described and interpreted as "function," "means," "part," "processing circuit," "unit," or "module," etc.

[0127] 5 Variations The embodiments described above are provided to facilitate understanding of this disclosure and are not intended to limit it. The configurations that the embodiments may have are not limited to those exemplified and can be modified as appropriate. Furthermore, configurations shown in different embodiments can be partially substituted or combined.

[0128] 5.1 First Variation In "3.2 Example of Display Screen" of the above embodiment, an example was described in which N in Figure 3 is "6" (i.e., an example in which the user is prompted to input the first text to the sixth text and a summary sentence thereafter). In connection with this, the information processing device 2 can control the processing according to the above embodiment so that the user inputs sentences other than the summary sentence a predetermined number of times.

[0129] In one embodiment, after acquiring a sentence from the terminal device 3, the information processing device 2 identifies the number of times the user has entered sentences up to that point, including the sentence in question. If it is determined that the number of sentences has reached a predetermined number, it outputs information prompting the user to enter a summary sentence (see S116 to S122 in Figure 3). If it is determined that the number of sentences has not reached a predetermined number, it acquires evaluation information corresponding to the most recently acquired sentence and outputs information prompting the user to enter the next sentence based on that evaluation information (see S102 to S114 in Figure 3).

[0130] In this case, the information processing device 2 may output information prompting the user to input a summary statement, regardless of the content of the sentence already entered by the user (for example, whether it is logical, clear, etc.), if the number of times the user has entered a sentence has reached a predetermined number.

[0131] 5.2 Second Variation In one embodiment, the evaluation information acquisition unit 102 may acquire second evaluation information based on at least a portion of the first sentence, the first evaluation information, and the information prompting the input of the second sentence, in addition to the second sentence. That is, the evaluation information acquisition unit 102 may acquire second evaluation information based on the process leading up to the acquisition of the second sentence. In the above embodiment, one example of this configuration is that the nth instruction transmitted to the LLM server device 4 may include the first text to the (n-1)th text and the first response to the (n-1)th response.

[0132] 5.3 Third Variation In the above embodiment, an example was described in which the information processing device 2 inputs instructions to the LLM server device 4, but the LLM may be hosted by the information processing device 2. That is, the information processing device 2 may execute the processing according to the above embodiment using a local LLM. Note that the LLM may be replaced by, for example, a large-scale deep learning model such as a Transformer that has been tuned to suit the above embodiment.

[0133] 5.4 Fourth Variation In Figure 3 of the above embodiment, an example was described in which an nth instruction, including a general instruction, a determination instruction, a hint generation instruction, an evaluation instruction, a specific instruction, and a comment generation instruction, is sent to the LLM server device 4 to execute a series of processes, but the example is not limited to this. At least a part of the processes to be executed by the LLM server device 4 may be executed by the information processing device 2 instead. In one example, the information processing device 2 may determine whether the nth text contains a predetermined word or the like related to the hint request, and if it determines that it does not contain it, it may send an nth instruction, including a general instruction, an evaluation instruction, a specific instruction, and a comment generation instruction, to the LLM server device 4, and if it determines that it does contain it, it may send an nth instruction, including a general instruction and a hint generation instruction, to the LLM server device 4. In other words, the instruction corresponding to the determination instruction may be executed by the information processing device 2 instead of the LLM server device 4.

[0134] 5.5 Fifth Variation With reference to Figures 4-8 of the above embodiment, the example of the display screen described above may be such that at least a portion of the information displayed by the terminal device 3 is displayed in a predetermined language, and at least another portion of the information displayed by the terminal device 3 is displayed in a language other than the predetermined language. In this case, the predetermined language is the user's first language, and the other language may be a language other than the user's first language. The terminal device 3 may accept the user's specification of the predetermined language and / or other languages. The terminal device 3 may estimate or determine the predetermined language and / or other languages ​​based, for example, its own location information.

[0135] For example, if the user specifies Japanese as the first language and English as the other language, the terminal device 3 may display the sentences entered by the user (e.g., texts t102 and t106 in Figure 4) and the information prompting the user to enter the next sentence (e.g., texts t104 and t108 in Figure 4) in English, while the feedback for the summary sentence (e.g., texts t500, t504, and t506 in Figure 8) may be displayed in Japanese.

[0136] 5.6 Sixth Variation The above embodiment describes an example in which the user inputs a summary sentence separately from the first to nth texts, but is not limited to this. The information processing device 2 may treat the first to nth texts input by the user as a summary sentence. The information processing device 2 may generate a summary sentence based on the first to nth texts (for example, by combining these texts) and control the terminal device 3 so that a revised sentence based on the generated summary sentence is displayed. The information processing device 2 may prompt the user to input a summary sentence separately from the first to nth texts if certain conditions regarding the progress of opinion formation are met, and may treat the first to nth texts as a summary sentence if those conditions are not met. The predetermined conditions may be that the user's opinion is not formed within a predetermined number of sentence inputs (for example, at least one of the user's assertion, reason, and specific example is significantly missing, and the user has requested hints multiple times).

[0137] 5.7 Seventh Variation In the above embodiment, an example was described in which the output unit 104 outputs information prompting the user to input a second sentence regarding their opinion based on the first evaluation information. However, the information output by the output unit 104 is not limited to this. In one embodiment, the output unit 104 outputs a second sentence that reflects modifications and / or additions to the first sentence to enhance its persuasiveness, based on the first sentence and the first evaluation information. That is, while the above embodiment of the information processing device 2 mainly described an example in which the user is prompted to input a second sentence that is more persuasive than the first sentence, the information processing device 2 in other embodiments may present the user with a second sentence that is more persuasive than the first sentence, for example, as an answer corresponding to the first sentence. With this configuration, the user can grasp a more persuasive sentence based on the first sentence they have input. This makes it possible to support the user in constructing their opinion more efficiently.

[0138] The information processing device 2 may send an instruction to the LLM server device 4 to generate a sentence that reflects the modifications and / or additions made to the first sentence to enhance its persuasiveness, and may present the sentence to the user based on the response. Furthermore, the various configurations described in the above embodiment can be applied to determine the second sentence that the information processing device 2 presents to the user.

[0139] The information processing device 2 may further output a sentence that reflects the modifications and / or additions made to the first sentence to enhance its persuasiveness, along with information prompting the input of the second sentence according to the above embodiment.

[0140] 6. Supplement The wording in this embodiment may be understood as follows, to the extent that it does not cause any contradiction.

[0141] In this embodiment, "performing a predetermined process based on predetermined information" may mean performing the predetermined process based on at least a portion of the predetermined information, performing the predetermined process based on at least the predetermined information, or performing the predetermined process probabilistically based on the predetermined information. In other words, "performing a predetermined process based on predetermined information" is not limited to performing the predetermined process based solely on the predetermined information.

[0142] In this embodiment, "executing another process based on a predetermined process" may mean any of the following: executing the other process after the predetermined process has been executed; executing the predetermined process and the other process consecutively; executing the other process based on information determined by the predetermined process; executing the other process on the condition that the predetermined process has been executed; or executing the other process by means of the predetermined process. Furthermore, "executing another process by a predetermined process" may be understood in the same way as "executing another process based on a predetermined process."

[0143] In this embodiment, "the predetermined information includes other information" may mean either that at least a portion of the predetermined information is the other information, or that the other information can be obtained based on the predetermined information.

[0144] In this embodiment, "a predetermined process includes other processes" may mean either that at least a part of the predetermined process is the other process (i.e., the other process is performed in the process of obtaining the result of the predetermined process), or that one aspect of the predetermined process is the other process.

[0145] In this embodiment, "a predetermined object and another object correspond" may mean that there is a one-to-one relationship between the predetermined object and the other object, that the other object is included in a predetermined set identified based on the predetermined object, or that the other object can be identified based on the predetermined object. Furthermore, "a predetermined object and another object corresponding" is not limited to being managed, for example, in a database. Also, "a predetermined object and another object being associated" may be understood in the same way as "a predetermined object and another object corresponding."

[0146] In this embodiment, "acquiring information" includes making the information processable by the control unit 10. "Acquiring information" may include, for example, receiving the information from another device, obtaining the information through predetermined processing, and reading the information from the storage unit 12.

[0147] In this embodiment, "generating information" may mean either making the information obtained by a predetermined process processable in the control unit 10, or storing the information obtained by the predetermined process in the storage unit 12.

[0148] In this embodiment, "determining information" may mean either selecting at least one piece of information from one or more pieces of information, or generating new information.

[0149] In this embodiment, "outputting information" may mean either transmitting the information to another device or outputting the information as audio or video.

[0150] 7. Example Configuration This disclosure includes the following technologies:

[0151] [Note 1] Information processing device 2 comprises: a sentence acquisition unit 100 that acquires a first sentence regarding the user's opinion; an evaluation information acquisition unit 102 that acquires first evaluation information regarding the evaluation of the persuasiveness of the first sentence; and an output unit 104 that outputs information prompting the user to input a second sentence regarding the user's opinion based on the first evaluation information.

[0152] [Note 2] The first evaluation information is determined by evaluating the first sentence using one or more evaluation criteria, and the information prompting input for the second sentence includes information prompting input regarding an evaluation criterion that is lower in evaluation compared to other evaluation criteria, as described in Appendix 1, for the information processing device 2.

[0153] [Note 3] The information processing device 2 described in Appendix 2 includes at least one of the following evaluation criteria: logic, clarity, and the degree of completeness of specific examples.

[0154] [Note 4] The information processing device 2 according to any one of the appendices 1 to 3, wherein the sentence acquisition unit 100 further acquires a second sentence, and the evaluation information acquisition unit 102 further acquires second evaluation information regarding the evaluation of the persuasiveness of the second sentence.

[0155] [Note 5] The output unit 104 further outputs information prompting the user to input a third sentence regarding their opinion, based on the second evaluation information, as described in Appendix 4, for the information processing device 2.

[0156] [Note 6] The information processing device 2 according to any one of the appendices 1 to 5, wherein the sentence acquisition unit 100 further acquires a second sentence, and the output unit 104 further outputs information prompting the input of a concluding sentence that summarizes the first and second sentences.

[0157] [Note 7] The information processing device 2 described in Appendix 6 further comprises a text acquisition unit 100 which acquires a summary text and a generation unit 106 which generates a revised text that reflects the modifications and / or additions to the summary text, and an output unit 104 which further outputs the revised text.

[0158] [Note 8] The information prompting input for the second sentence is provided by the information processing device 2, which includes any opposing opinions to the first sentence, as described in one of the appendices 1 to 7.

[0159] [Note 9] The information processing device 2 described in any one of the appendices 1 to 8, wherein the evaluation information acquisition unit 102 acquires first evaluation information by inputting an instruction to a large-scale language model to evaluate the persuasiveness of the first sentence.

[0160] [Note 10] The information processing device 2 according to any one of the appendices 1 to 9, further comprising: a sentence acquisition unit 100 which further acquires a fourth sentence input by the user and a determination unit 108 which determines whether the fourth sentence satisfies predetermined conditions for a hint request; if the fourth sentence satisfies the predetermined conditions, an output unit 104 further outputs hints for creating a persuasive sentence based on the fourth sentence and / or the first sentence; if the fourth sentence does not satisfy the predetermined conditions, an evaluation information acquisition unit 102 further acquires other evaluation information regarding the evaluation of the persuasiveness of the fourth sentence.

[0161] [Note 11] The information processing device 2 described in Appendix 10 further comprises a specification unit 110 that identifies evaluation criteria for enhancing the persuasiveness of the first sentence from one or more evaluation criteria for evaluating the first sentence, and hints for creating a persuasive sentence are determined based on the evaluation criteria identified by the specification unit 110.

[0162] [Note 12] Information processing device 2 comprises: a sentence acquisition unit 100 that acquires a first sentence relating to a user's opinion; an evaluation information acquisition unit 102 that acquires first evaluation information relating to the persuasiveness of the first sentence; and an output unit 104 that outputs a second sentence that reflects modifications and / or additions to the first sentence that enhance its persuasiveness, based on the first sentence and the first evaluation information.

[0163] [Note 13] An information processing method in which a computer 70 performs the following actions: obtains a first sentence regarding the user's opinion; obtains first evaluation information regarding the persuasiveness of the first sentence; and outputs information prompting the user to input a second sentence regarding the user's opinion based on the first evaluation information.

[0164] [Note 14] A program that causes computer 70 to perform the following actions: obtain the first sentence regarding the user's opinion; obtain first evaluation information regarding the persuasiveness of the first sentence; and output information prompting the user to input a second sentence regarding their opinion based on the first evaluation information. [Explanation of symbols]

[0165] 1...System, 2...Information processing device, 3...Terminal device, 4...LLM server device, 10...Control unit, 70...Computer, 100...Text acquisition unit, 102...Evaluation information acquisition unit, 104...Output unit, 106...Generation unit, 108...Determination unit, 110...Specification unit

Claims

1. A sentence retrieval unit that retrieves the first sentence regarding the user's opinion, An evaluation information acquisition unit that acquires first evaluation information regarding the evaluation of the persuasiveness of the first sentence, Based on the first evaluation information, an output unit outputs information prompting the user to input a second sentence regarding their opinion, An information processing device equipped with the following features.

2. The first evaluation information is determined by evaluating the first sentence using one or more evaluation criteria. The information prompting input for the second sentence includes information prompting input regarding the evaluation criterion among the one or more evaluation criteria that has a lower evaluation compared to the other evaluation criteria. The information processing apparatus according to claim 1.

3. The information processing apparatus according to claim 2, wherein the one or more evaluation criteria include at least one of logic, clarity, and the degree of completeness of specific examples.

4. The aforementioned sentence acquisition unit further acquires the second sentence, The information processing apparatus according to claim 1, wherein the evaluation information acquisition unit further acquires second evaluation information relating to the evaluation of the persuasiveness of the second sentence.

5. The information processing apparatus according to claim 4, wherein the output unit further outputs information prompting the input of a third sentence regarding the user's opinion based on the second evaluation information.

6. The aforementioned sentence acquisition unit further acquires the second sentence, The information processing apparatus according to claim 1, wherein the output unit further outputs information prompting the input of a concluding sentence that summarizes the first sentence and the second sentence.

7. The aforementioned sentence acquisition unit further acquires the summary sentence, The system further comprises a generation unit that generates a revised text that reflects the revisions and / or additions to the aforementioned summary text, The information processing apparatus according to claim 6, wherein the output unit further outputs the corrected text.

8. The information processing apparatus according to claim 1, wherein the information prompting input for the second sentence includes objections to the first sentence.

9. The information processing apparatus according to claim 1, wherein the evaluation information acquisition unit acquires the first evaluation information by inputting an instruction to a large-scale language model to evaluate the persuasiveness of the first sentence.

10. The sentence acquisition unit further acquires the fourth sentence entered by the user, The fourth sentence further comprises a determination unit that determines whether or not it satisfies predetermined conditions relating to the request for a hint, If the fourth sentence satisfies the predetermined conditions, the output unit further outputs hints for creating a persuasive sentence based on the fourth sentence and / or the first sentence. If the fourth sentence does not satisfy the predetermined conditions, the evaluation information acquisition unit further acquires other evaluation information relating to the evaluation of the persuasiveness of the fourth sentence, as described in claim 1.

11. The system further comprises a specific unit that identifies evaluation criteria for enhancing the persuasiveness of the first sentence from one or more evaluation criteria relating to the evaluation of the first sentence, The information processing apparatus according to claim 10, wherein the hints for creating the aforementioned persuasive sentence are determined based on evaluation criteria identified by the specific unit.

12. A sentence retrieval unit that retrieves the first sentence regarding the user's opinion, An evaluation information acquisition unit that acquires first evaluation information regarding the evaluation of the persuasiveness of the first sentence, An output unit that outputs a second sentence in which revisions and / or additions to the first sentence have been made to enhance the persuasiveness of the first sentence, based on the first sentence and the first evaluation information; An information processing device equipped with the following features.

13. Computers To obtain the first sentence regarding the user's opinion, To obtain first evaluation information regarding the persuasiveness of the first sentence, Based on the first evaluation information, information prompting the user to input a second sentence regarding their opinion is output. An information processing method that performs [this action].

14. On the computer, To obtain the first sentence regarding the user's opinion, To obtain first evaluation information regarding the persuasiveness of the first sentence, Based on the first evaluation information, information prompting the user to input a second sentence regarding their opinion is output. A program that executes something.

Images