Document creation device, document creation method, information processing program, and information processing system

Abstract

To provide a document creation device, a document creation method, an information processing program, and an information processing system that deliver personalized response texts tailored to the individual situations and emotions of contributors.SOLUTION: A document creation device 1 comprises: a data acquisition unit 131 that obtains posts submitted by contributors and related information provided by contributors; a prompt transmission unit 132 that creates a prompt containing questions for generating features of the post and the related information along with reasons for their generation and sends the same to a language model server 3; a response reception unit 133 that receives response texts from the language model server 3; and a response transmission unit 134 that transmits the response texts to an information terminal 2.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a text creation device, a text creation method, an information processing program, and an information processing system for creating text. [Background technology]

[0002] Conventionally, a system is known that extracts information about a user's emotions from messages sent by the user, generates and stores an emotion quotient specified based on the extracted information, and collects the user's emotions (see, for example, Patent Document 1).

[0003] Furthermore, for example, a user who tends to normally become angry and a user who normally does not tend to become angry will have different emotional states when using the term "angry." Taking into account that different users may have different emotional states even when using the same word, a system is known that can provide information appropriate to the user's state (see, for example, Patent Document 2). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-33569 [Patent Document 2] Japanese Patent Publication No. 2022-155034 Summary of the Invention [Problem to be solved by the invention]

[0005] Conventional systems acquire multiple posts by a user, extract characteristic words contained in the posts, and identify changes in the user's status. For example, FIG. 3 of Patent Document 2 discloses a text creation device equipped with a posted message acquisition unit, an extraction unit, a change identification unit, and a response message creation unit. This device identifies changes in the status of multiple posts posted by a user within a specified period of time, and creates a response message based on those changes.

[0006] Conventional systems have the ability to extract characteristic words from posted text, but their methods rely on simple extraction of characteristic words, making it difficult to reflect the poster's intentions and emotions in detail, taking into account the complex context, metaphors, nuances, etc. of the text.

[0007] Furthermore, in conventional systems, responses are created based on predefined templates, which means that responses are uniform and it is difficult to provide personalized responses that reflect the poster's individual situation or feelings.

[0008] Therefore, the present invention has been made in consideration of these points, and aims to provide a writing device, writing method, information processing program, and information processing system that can provide a response sentence that is personalized to the poster who provided the posted message and related information, based on the poster's individual situation and emotions. [Means for solving the problem]

[0009] The text creation device of the present invention includes a data acquisition unit that acquires a message posted by a poster and related information provided by the poster; a prompt sending unit that sends a prompt including a question that generates characteristics related to the message and the related information and a reason for generating the characteristics; a response message receiving unit that receives a response message generated by the language model server; and a response message sending unit that sends the response message to an information terminal.

[0010] The data acquisition unit includes in the posted text, for example, details of actions and feelings, such as a diary, and the related information provided by the poster includes data indicating the poster's values ​​and goals, as well as health data such as the poster's number of steps and sleep time.

[0011] The prompt sent by the prompt sending unit to the language model server includes a question for generating an outline of the posted message and a reason for generating the message.

[0012] The prompt sent by the prompt sending unit to the language model server includes a question for generating representative characteristic words contained in the posted text and the reasons for generating the characteristic words.

[0013] The prompt sent by the prompt sending unit to the language model server includes a question for generating a word that expresses the emotion of the posted text and a reason for generating the word.

[0014] The prompt sent by the prompt sending unit to the language model server includes a question for generating an emotion score indicating the degree of positivity or negativity of the posted message and a reason for generating the emotion score.

[0015] The prompt sent by the prompt sending unit to the language model server includes a behavior score that evaluates the extent to which the poster's behavior was based on data indicating their values ​​and goals, from the posted text acquired by the data acquisition unit and related information provided by the poster, and a question that generates a reason for the generation of the behavior score.

[0016] The prompt sent by the prompt sending unit to the language model server includes a question that instructs not to respond negatively to the posted text, does not use negative conjunctions, and causes the user to generate specific praise content.

[0017] The prompts sent by the prompt sending unit to the language model server include questions that tailor responses based on speaking patterns specified by the poster.

[0018] The text creation method of the present invention includes the steps of: acquiring a message posted by a poster and related information provided by the poster; transmitting a prompt including a question for generating characteristics related to the message and the related information and a reason for generating the characteristics; receiving a response to the prompt calculated by the language model server from the language model server; and transmitting the response to an information terminal used by the poster, all of which are executed by a computer.

[0019] The information processing program of the present invention is characterized in that it causes a computer to execute the steps of: acquiring a message posted by a poster and related information provided by the poster; transmitting a prompt including a question for generating characteristics related to the message and the related information and reasons for generating the characteristics; receiving a response message to the prompt calculated by the language model server from the language model server; and transmitting the response message to an information terminal used by the poster.

[0020] The information processing system of the present invention comprises a writing device, an information terminal used by a poster, and a language model server. The writing device has a data acquisition unit that acquires a message posted by the poster and related information provided by the poster, a prompt sending unit that sends a prompt to the language model server, the prompt including a question that generates characteristics related to the message and the related information and reasons for their generation, a response message receiving unit that receives from the language model server a response message to the prompt calculated by the language model server, and a response message sending unit that sends the response message to the information terminal used by the poster. The information terminal has a data transmission unit that transmits the message entered by the poster and the related information, a response message receiving unit that receives the response message from the writing device and stores it in a memory unit after the data transmission unit has sent the message, and a display unit that retrieves the response message from the memory unit and displays it on a display unit. [Effects of the Invention]

[0021] According to the present invention, it is possible to provide a poster who has provided a posted message and related information with a response message that is personalized to the poster's individual situation and feelings. [Brief explanation of the drawings] [Figure 1] FIG. 1 is a diagram illustrating an overview of an information processing system. [Figure 2] FIG. 10 is a diagram showing an example of a message including a response sentence created by the text creation device. [Figure 3]FIG. 1 is a diagram illustrating a configuration of a text creation device. [Figure 4] FIG. 2 is a diagram illustrating a configuration of an information terminal. [Figure 5] FIG. 10 is a flowchart showing a process flow of the writing device for creating a response sentence. DETAILED DESCRIPTION OF THE INVENTION [Outline of Information Processing System S] 1 is a diagram showing an overview of an information processing system S. The information processing system S includes a text creation device 1, an information terminal 2, and a language model server 3. The information processing system S is a system for providing a personalized response message to a poster U who has posted a message and provided related information.

[0022] The writing creation device 1 is a computer that acquires a posted message posted by a poster U and related information provided by the poster U, transmits a prompt including a question for generating characteristics related to the posted message and the related information and the reason for their generation to a language model server 3, and acquires a response message to the prompt calculated by the language model server 3. The information terminal 2 is an information terminal used by the poster U, and is, for example, a smartphone, a tablet, or a personal computer. As shown in FIG. 1 , the information terminal 2 transmits the posted message and the related information to the writing creation device 1 via a network N. After receiving the posted message and the related information, the writing device 1 generates a prompt including a question for generating characteristics related to the posted message and the related information and the reason for their generation, and transmits the prompt to the language model server 3 via the network N. After receiving the response message from the language model server 3, the writing device 1 transmits the response message to the information terminal 2.

[0023] The content of the message posted by poster U is arbitrary, but is expected to include, for example, the poster U's actions and feelings, like a diary. The related information provided by poster U is expected to include data indicating the poster U's values ​​and goals, as well as health data such as the poster U's steps and sleep time. The prompts sent to the language model server are expected to include questions for generating an outline of the message and the reason for its creation, questions for generating representative characteristic words contained in the message and the reason for their creation, questions for generating words expressing the sentiment of the message and the reason for their creation, questions for generating an emotion score indicating the positivity or negativity of the sentiment of the message and the reason for their creation, questions for generating, from the message and the related information, an action score evaluating the extent to which poster U's actions were based on the data indicating their values ​​and goals, and the reason for their creation, questions for instructing the poster U not to respond negatively to the message and to generate specific praise without using negative conjunctions, and questions for adjusting the response based on the speech pattern specified by poster U.

[0024] Fig. 2 is a diagram showing an example of a message including a response sentence created by the writing creation device 1. The writing creation device 1 receives a plurality of response sentences corresponding to a plurality of pre-set items from the language model server 3. In Fig. 2, seven items are shown as examples of the plurality of response sentences: "summary," "characteristic words," "emotion," "emotion score," "behavioral evaluation score based on values," "behavioral evaluation score based on goals," and "reason for each output."

[0025] The "summary" is a response sentence generated by sending to the language model server 3 a message posted by poster U and a prompt including a question for generating a summary of the message and the reason for its creation.

[0026] The "characteristic words" are response sentences generated by sending a prompt to the language model server 3, which includes a message posted by poster U and a question that asks for the generation of representative characteristic words contained in the message and the reason for their generation.

[0027] "Emotion" is a response sentence generated by sending a message posted by poster U and a prompt including a question to generate words expressing the emotion of the message and the reason for generating the words to language model server 3.

[0028] The "emotion score" is a response sentence generated by sending a prompt to the language model server 3, which includes a message posted by poster U and a question to generate an emotion score indicating the positivity or negativity of the emotion of the message, as well as the reason for generating the emotion score.

[0029] The "values-based behavioral evaluation score" and "goal-based behavioral evaluation score" are behavioral scores that evaluate the extent to which poster U's behavior was based on data that indicates poster U's values ​​and goals, based on the poster U's post and related information provided by poster U, and are response sentences generated by sending a prompt containing a question to generate the reason for the score to language model server 3.

[0030] The "reason for each output" is a response sentence indicating the reason for generation obtained as a result of asking the language model server 3 to generate the characteristics such as "summary," "characteristic words," "emotion," "emotion score," "behavioral evaluation score based on values," and "behavioral evaluation score based on goals."

[0031] [Configuration of Text Creation Device 1] 3 is a diagram showing the configuration of the writing device 1. The writing device 1 has a communication unit 11, a storage unit 12, and a control unit 13.

[0032] The communication unit 11 is a communication interface for transmitting and receiving data to and from the information terminal 2 or the language model server 3 via the network N. The communication unit 11 inputs the posted message received from the information terminal 2 and the related information provided by the poster U to the data acquisition unit 131. The communication unit 11 also transmits the prompt created by the prompt transmission unit 132 to the language model server 3. The communication unit 11 also receives a response message from the language model server 3 at the response message reception unit 133, inputs the response message to the response message transmission unit 134, and transmits it to the information terminal 2.

[0033] The storage unit 12 has storage media such as a read-only memory (ROM), a random access memory (RAM), and a solid-state drive (SSD). The storage unit 12 stores programs that the control unit 13 executes.

[0034] The control unit 13 includes, for example, a CPU (Central Processing Unit). The control unit 13 executes the programs stored in the storage unit 12, thereby functioning as a data acquisition unit 131, a prompt transmission unit 132, a response message reception unit 133, and a response message transmission unit 134.

[0035] The data acquisition unit 131 acquires the message posted by the poster U and the related information provided by the poster U via the communication unit 11.

[0036] The prompt sending unit 132 creates prompts including questions to generate an outline of the posted message and the reason for its creation, from the message posted by the poster U and related information provided by the poster U acquired by the data acquisition unit 131, questions to generate representative characteristic words contained in the posted message and the reason for their creation, questions to generate words that express the emotion of the posted message and the reason for their creation, questions to generate an emotion score that indicates the positivity or negativity of the emotion of the posted message and the reason for its creation, questions to generate an action score that evaluates the extent to which the poster U acted based on data that indicates the poster U's values ​​and goals and the reason for its creation, from the posted message and related information provided by the poster U, questions to instruct the poster U not to respond negatively to the posted message and to generate specific praise without using negative conjunctions, and questions to adjust the response based on the speaking pattern specified by the poster U, and sends these prompts to the language model server 3 via the communication unit 11.

[0037] The response sentence receiving unit 133 acquires the response sentence generated by the language model server 3 via the communication unit 11 .

[0038] The response message sending unit 134 sends the response message acquired by the response message receiving unit 133 to the information terminal 2 via the communication unit 11.

[0039] Here, "language model server 3" includes AI (Artificial Intelligence) that generates sentences using prompts as input, such as ChatGPT, Gemini, Notion AI, Catchy, Jasper, ELYZA Pencil, SAKUBUN, Magic Write, AI Verist, PlayAI, YouChat, Chatsonic, Socratic, Transcope, Rytr, Cohesive, Perplexity, Writesonic, Forefront Chat, HuggingChat, Azure Open AI Service, LLaMA, ChatPro, UnionGAI, Metareal AI, HyperCLOVA, and other sentence generation AIs, or programs, services, or software that use these.

[0040] Furthermore, the term "sentence generation AI" includes any language model or large language model (LLM), such as GPT (Generative Pre-trained Transformer)-1, GPT-2, GPT-3, GPT-3.5, GPT-4, Gemini 1.0 Pro, Gemini 1.5 Pro, Gemini 1.5 Flash, ELMo (Embeddings from Language Models), LaMDA (Large language Model Meta AI), PaLM (Pathways Language Model), PaLM2, NEMO LLM, Claude, Alpaca 7B, Vicuna 13B, OpenFlamingo, Transformer, or language models or large language models that have been modified, improved, transferred, or additionally trained.

[0041] [Configuration of Information Terminal 2] 4 is a diagram showing the configuration of the information terminal 2. The information terminal 2 has a communication unit 21, a display unit 22, an operation unit 23, a storage unit 24, and a control unit 25.

[0042] The communication unit 21 has a communication interface for transmitting and receiving data to and from the writing device 1 via the network N. The communication unit 21 has, for example, a wireless communication controller, and transmits the posted message input from the data transmission unit 251 and related information provided by the poster U to the writing device 1, and receives a response message from the writing device 1. The communication unit 21 inputs the received response message to the response message receiving unit 252 and stores it in the memory unit 24.

[0043] The display unit 22 is a display that acquires and displays a response sentence from the storage unit 24 .

[0044] The operation unit 23 is an operation device that accepts operations by the poster U, and is, for example, a touch panel that is provided over the display unit 22. The operation content of the poster U is notified to the data transmission unit 251.

[0045] The storage unit 24 includes a ROM and a RAM, and stores the programs executed by the control unit 22. The storage unit 24 also saves the response message received as the response message 252.

[0046] The control unit 25 includes, for example, a CPU, and functions as a data transmission unit 251 and a response message reception unit 252 by executing a program stored in the storage unit 24 .

[0047] The data transmission unit 251 transmits the posted message input by the poster U via the operation unit 23 and related information acquired by the data transmission unit 251 from the memory unit to the writing device 1 via the communication unit 21. After the data transmission unit 251 transmits the posted message, the response message reception unit 252 receives the response message from the writing device 1 and stores the received response message in the memory unit 24.

[0048] [Flowchart for creating a response] Fig. 5 is a flowchart showing the flow of processing for creating a response message by the text creation device 1. The flowchart shown in Fig. 5 starts from the point in time when the poster U provides a posted message and related information.

[0049] When the data acquisition unit 131 acquires data (S11), the data acquisition unit 131 inputs the data to the prompt transmission unit 132, which generates a prompt and transmits it to the language model server 3 (S12). The response sentence reception unit 133 receives the response sentence from the language model server 3 (S13), formats this response sentence in the format shown in Fig. 2, and transmits it to the information terminal 2 via the communication unit 1 (S14).

[0050] [Effects of Text Creation Device 1] As described above, the writing creation device 1 includes a data acquisition unit 131 that acquires a posted message and related information provided by the poster U, and a prompt sending unit 132 that sends a prompt including a question that generates the characteristics of the posted message and the reason for its creation to the language model server 3. By configuring the writing creation device 1 in this way, the writing creation device 1 can provide the poster U, who has provided the posted message and related information, with a response message that is personalized to the individual situation and feelings of the poster U.

[0051] The present invention has been described above using embodiments, but the technical scope of the present invention is not limited to the scope described in the above embodiments, and various modifications and changes are possible within the scope of the gist of the present invention. For example, all or part of the device can be configured by functionally or physically distributing or combining any unit. Furthermore, new embodiments resulting from any combination of multiple embodiments are also included in the embodiments of the present invention. The disclosure of new embodiments resulting from the combination will also have the effects of the original embodiments. [Explanation of symbols]

[0052] 1. Writing device 2. Information terminal 3 Language Model Server 11 Communications Department 12 Storage section 13 Control Unit 21 Communications Department 22 Display section 23 Control section 24 Memory section 25 Control Unit 131 Data Acquisition Unit 132 Prompt sending section 133 Response message receiving unit 134 Response message sending unit 251 Data Transmission Unit 252 Response Message Reception Unit

Claims

1. a data acquisition unit that acquires posts posted by posters and related information provided by the posters; a prompt sending unit that sends a prompt to a language model server, the prompt including a question that causes the language model server to generate features related to the posted text and the related information and a reason for generating the features; a response sentence receiving unit that receives, from the language model server, a response sentence to the prompt calculated by the language model server; a response message sending unit that sends the response message to an information terminal used by the poster; A writing device having the above.

2. The data acquisition unit is characterized in that the posted text includes, for example, details of actions and feelings, such as a diary, and the related information includes data indicating the poster's values ​​and goals, and health data, such as the number of steps taken and the amount of sleep. The document creation device according to claim 1 .

3. the prompt sent by the prompt sending unit to the language model server includes a question for generating an outline of the posted message and a reason for generating the outline.

3. The document creation device according to claim 1 or 2.

4. the prompt sent by the prompt sending unit to the language model server includes a question for generating representative feature words included in the posted text and reasons for generating the representative feature words. The document creation device according to any one of claims 1 to 3.

5. the prompt sent by the prompt sending unit to the language model server includes a question for generating a word expressing the emotion of the posted message and a reason for generating the word. The document creation device according to any one of claims 1 to 4.

6. the prompt sent by the prompt sending unit to the language model server includes a question for generating an emotion score indicating a positive or negative emotion of the posted message and a reason for generating the emotion score. The document creation device according to any one of claims 1 to 5.

7. The prompt transmitted to the language model server by the prompt transmission unit includes a behavior score that evaluates the extent to which the poster has acted based on data indicating the poster's values ​​and objectives, from the posted message and the related information, and a question that generates a reason for generating the behavior score. The document creation device according to any one of claims 1 to 6.

8. the prompt sent by the prompt sending unit to the language model server includes a question that instructs the user not to make a negative response to the posted message, does not use negative conjunctions, and causes the user to generate specific praise content. The document creation device according to any one of claims 1 to 7.

9. the prompt sent to the language model server by the prompt sending unit includes a question for adjusting a response sentence based on a speaking pattern specified by the poster. The document creation device according to any one of claims 1 to 8.

10. The computer executes A step of acquiring a message posted by a poster and related information provided by the poster; sending a prompt to a language model server, the prompt including a question for generating features related to the posted text and the related information and a reason for generating the features; receiving from the language model server a response sentence to the prompt computed by the language model server; transmitting the response text to an information terminal used by the poster; A method of writing sentences.

11. On the computer, A step of acquiring a message posted by a poster and related information provided by the poster; sending a prompt to a language model server, the prompt including a question for generating features related to the posted text and the related information and a reason for generating the features; receiving from the language model server a response sentence to the prompt computed by the language model server; transmitting the response text to an information terminal used by the poster; An information processing program characterized by causing the program to execute the above.

12. The system includes a writing device, an information terminal used by a contributor, and a language model server, The writing device includes: a data acquisition unit that acquires posts posted by posters and related information provided by the posters; a prompt sending unit that sends a prompt to a language model server, the prompt including a question that causes the language model server to generate features related to the posted text and the related information and a reason for generating the features; a response sentence receiving unit that receives, from the language model server, a response sentence to the prompt calculated by the language model server; a response message sending unit that sends the response message to an information terminal used by the poster; and The information terminal a data transmission unit that transmits the posted message and the related information; a response message receiving unit that receives the response message from the writing device after the data sending unit has sent the message and stores it in a storage unit; and a display unit that retrieves the response message from the storage unit and displays it on a display unit. An information processing system having the above.

Images