Information processing device, information processing method, and program

Abstract

This invention provides an information processing device, an information processing method, and a program that can improve the efficiency of obtaining answers that satisfy the user's objectives from a generative model. [Solution] One aspect of the information processing device of the present disclosure includes: a receiving unit that receives input data entered by a user; a generating unit that generates a prompt including the input data; an acquiring unit that inputs the prompt multiple times into a generation model to obtain multiple answers and / or generates multiple different prompts, inputs the multiple prompts into the generation model to obtain multiple answers; and an output unit that outputs the multiple answers.

Description

Technical Field

[0001] This disclosure relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] Conventionally, conversational AI services using generative artificial intelligence (Generative AI) have been known. For example, Patent Document 1 discloses a technique for providing an answer while referring to an appropriate document in response to a user's question using such generative AI.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in conventional conversational AI services, usually only a single answer can be obtained in response to an instruction from a user. For this reason, when the user does not obtain the desired answer or when the user aims to obtain a plurality of answers from different viewpoints, it is necessary to re-instruct the generation model (generative AI), adjust the instruction content, or change the generation model to be answered. Therefore, there is room for improving the efficiency until an answer that satisfies the user's purpose is obtained from the generation model in conventional conversational AI services.

[0005] This disclosure has been made in view of the above circumstances, and an object thereof is to provide an information processing apparatus, an information processing method, and a program capable of improving the efficiency until an answer that satisfies the user's purpose is obtained from a generation model.

Means for Solving the Problems

[0006] One aspect of the information processing device of the present disclosure includes: a receiving unit that receives input data entered by a user; a generating unit that generates a prompt including the input data; an acquiring unit that inputs the prompt multiple times into a generation model to obtain multiple answers and / or generates multiple different prompts, inputs the multiple prompts into the generation model to obtain multiple answers; and an output unit that outputs the multiple answers.

[0007] One aspect of the information processing method of the present disclosure includes: a receiving step in which a receiving unit receives input data entered by a user; a generating step in which a generating unit generates a prompt including the input data; an acquiring step in which an acquiring unit inputs the prompt multiple times into a generation model to acquire multiple answers and / or generates multiple different prompts, inputs the multiple prompts into the generation model to acquire multiple answers; and an output step in which an output unit outputs the multiple answers.

[0008] One aspect of the program of this disclosure causes a computer to perform the following steps: a receiving step of receiving input data entered by a user; a generating step of generating a prompt that includes the input data; a retrieval step of inputting the prompt multiple times into a generating model to obtain multiple answers and / or generate multiple different prompts, inputting the multiple prompts into the generating model to obtain multiple answers; and an output step of outputting the multiple answers. [Effects of the Invention]

[0009] According to this disclosure, it is possible to provide an information processing device, an information processing method, and a program that can improve the efficiency of obtaining a response that satisfies the user's objectives from a generative model. [Brief explanation of the drawing]

[0010] [Figure 1] Figure 1 is a block diagram showing an example of the system configuration of the data generation system of this embodiment. [Figure 2]Figure 2 is a block diagram showing an example of the hardware configuration of the information processing device of this embodiment. [Figure 3] Figure 3 is a block diagram showing an example of the functional configuration of the information processing device according to this embodiment. [Figure 4] Figure 4 shows an example of the UI screen of the AI ​​service for generating documents according to this embodiment. [Figure 5] Figure 5 shows an example of input data entered into the input area of ​​this embodiment. [Figure 6] Figure 6 shows an example of a prompt template in this embodiment. [Figure 7] Figure 7 shows an example of a prompt template in this embodiment. [Figure 8] Figure 8 shows an example of a response (first draft) displayed in the output area of ​​this embodiment. [Figure 9] Figure 9 shows an example of the response (second draft) displayed in the output area of ​​this embodiment. [Figure 10] Figure 10 is a flowchart showing an example of the data generation process performed in the information processing device of this embodiment. [Figure 11] Figure 11 shows an example of a prompt template for Modification Example 1. [Figure 12] Figure 12 shows an example of a prompt template for Modification Example 1. [Figure 13] Figure 13 shows an example of a tabular response displayed in the output area of ​​Modification 1. [Figure 14] Figure 14 shows an example of a report-format response displayed in the output area of ​​Modification 1. [Figure 15] Figure 15 shows an example of the UI screen of the AI ​​service for generating data according to Modification Example 3. [Modes for carrying out the invention]

[0011] Hereinafter, embodiments of the present disclosure (hereinafter simply referred to as "the present embodiment") will be described in detail with reference to the drawings. Note that the present disclosure is not limited to the following embodiments. Also, the following embodiments and each modification can be combined as appropriate.

[0012] Hereinafter, a service that generates various materials using generative AI (hereinafter sometimes referred to as "the material generation AI service") will be taken as an example to explain the method of the present disclosure. Specifically, a method of generating a plurality of prompts from a user's input and inputting the generated plurality of prompts into a generation model to obtain a plurality of answers (materials) will be described. However, the present disclosure is not limited to this.

[0013] First, the configuration of the material generation system of the present embodiment will be described.

[0014] FIG. 1 is a block diagram showing an example of the system configuration of the material generation system 1 of the present embodiment. As shown in FIG. 1, the material generation system 1 includes an information processing apparatus 10 and a generation model 20.

[0015] The information processing apparatus 10 and the generation model 20 are connected via a network 2. The network 2 can be realized by, for example, at least any one of the Internet and a LAN (Local Area Network). The network 2 may be a wired network, a wireless network, or a mixture of a wired network and a wireless network.

[0016] The information processing apparatus 10 is a terminal device used by a user who uses the material generation AI service, and examples include, but are not limited to, a PC (Personal Computer), a tablet terminal, or a smartphone.

[0017] The generative model 20 is a natural language processing model, also known as generative AI (Generative Artificial Intelligence), and provides services using natural language processing. In this embodiment, the generative model 20 is described as an interactive AI service obtained by fine-tuning a large language model (LLM), but it is not limited to this.

[0018] Figure 2 is a block diagram showing an example of the hardware configuration of the information processing device 10 of this embodiment. As shown in Figure 2, the information processing device 10 comprises a control device 11, a main memory 12, an auxiliary storage device 13, a communication device 14, an input device 15, a display device 16, and various buses 19. The control device 11, main memory 12, auxiliary storage device 13, communication device 14, input device 15, and display device 16 are connected via various buses 19. Thus, the information processing device 10 of this embodiment has a general hardware configuration using a normal computer.

[0019] The control device 11 controls the overall operation of the information processing device 10. The control device 11 may be, for example, at least one of a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit), but is not limited to these. There may be one or more CPUs or GPUs, and they may be single-core or multi-core.

[0020] Examples of main memory 12 include, but are not limited to, ROM (Read Only Memory) and RAM (Random Access Memory). ROM stores various programs, such as programs for controlling the information processing device 10 and programs for realizing the data generation AI service. RAM is used as a workspace when the control device 11 performs various controls based on the programs stored in ROM.

[0021] The auxiliary storage device 13 stores the various programs described above and various data used to realize the data generation AI service. The various programs described above only need to be stored in at least one of the main storage device 12 and the auxiliary storage device 13. Examples of the auxiliary storage device 13 include, but are not limited to, at least one of existing storage devices capable of magnetic, electrical, or optical storage, such as an HDD (Hard Disk Drive), SSD (Solid State Drive), and DVD (Digital Versatile Disc). The auxiliary storage device 13 may be built into the information processing device 10 or externally connected to the information processing device 10 via an interface such as USB (Universal Serial Bus). Furthermore, the auxiliary storage device 13 may be a NAS (Network Attached Storage) connected via a network such as a LAN or WAN (Wide Area Network).

[0022] The communication device 14 is used to communicate with the generative model 20 and the like via the network 2. Examples of the communication device 14 include, but are not limited to, a communication device for a wired LAN or a wireless communication device for a wireless LAN.

[0023] The input device 15 is used by the user for various inputs, selections, and specifications, and serves as a user interface between the user and the system. Examples of input devices 15 include, but are not limited to, keyboards, mice, and touch panels. The input device 15 may be built into the information processing device 10 or connected externally to the information processing device 10 via an interface such as USB.

[0024] The display device 16 displays various screens and serves as a user interface between the user and the device. Examples of the display device 16 include, but are not limited to, liquid crystal displays, organic electro-luminescence (OLED) displays, and touch panel displays. The display device 16 may be an internal display built into the information processing device 10, or an external display connected to the information processing device 10 via a display interface such as HDMI®.

[0025] In addition to the above configuration, the information processing device 10 may also be equipped with hardwired circuits such as ICs (Integrated Circuits), ASICs (Application Specific Integrated Circuits), and FPGAs (Field-Programmable Gate Arrays) that are specific to the information processing device 10.

[0026] Figure 3 is a block diagram showing an example of the functional configuration of the information processing device 10 in this embodiment. As shown in Figure 3, the information processing device 10 includes a display control unit 101, a reception unit 103, a generation unit 105, an acquisition unit 107, and an output unit 109. The display control unit 101, reception unit 103, generation unit 105, acquisition unit 107, and output unit 109 can be realized, for example, by the control device 11, main memory 12, and communication device 14 described in Figure 2. The control device 11 reads the program for realizing the data generation AI service stored in the main memory 12 (ROM) or auxiliary memory 13 and expands it into the main memory 12 (RAM). The control device 11 realizes each of the above-mentioned functional units by executing various processes according to the expanded program. Here, the case in which each of the above-mentioned functional units is realized as software has been explained as an example, but at least a part of each of the above-mentioned functional units may be realized as hardware. In this case, the functional unit to be realized as hardware can be realized, for example, by the hardwired circuit described above. Furthermore, any of the above-mentioned functional units may be implemented through the cooperation of software and hardware.

[0027] Furthermore, the information processing device 10 of this embodiment does not necessarily require all of the above-described functional units to be part of its essential configuration; at least some of the functional units can be omitted.

[0028] In the following, this embodiment will be described using, as appropriate, a case in which multiple drafts of a news release are generated from user input information, but this embodiment is not limited to this.

[0029] The display control unit 101 displays the UI (User Interface) screen of the document generation AI service on the display device 16 in response to a user's operation to start the document generation AI service using an input device 15 or the like. Examples of operations to start the document generation AI service include, but are not limited to, launching the application that implements the document generation AI service or connecting to (accessing) the Web service that implements the document generation AI service.

[0030] Figure 4 shows an example of the UI screen 201 of the document generation AI service of this embodiment, and shows the initial state (initial screen) of the UI screen 201 displayed when the document generation AI service is started. Note that in the UI screen 201 shown in Figure 4, some of the elements that make up the screen have been appropriately omitted for explanatory purposes.

[0031] As shown in Figure 4, the UI screen 201 consists of an input area 211, a generate button 231, output areas 241 and 261, and operation tabs 242 and 262.

[0032] Input area 211 is a display area where input data (case information) entered by the user is displayed. Input area 211 includes a predetermined number of input items and input fields associated with each of those input items. In the example shown in Figure 4, the input item "Who and Who" belonging to the main item "The most important news to convey in this announcement" is associated with input field 212, the input item "What" is associated with input field 213, and the input item "When" is associated with input field 214. In each input field, the user enters content corresponding to the corresponding input item in order to generate a draft of the news release. Alternatively, the user may upload a file in CSV (Comma Separated Values) format or similar so that the content corresponding to the input item is entered into each input field.

[0033] The generate button 231 is a UI component (UI image) that causes the generation model 20 to generate multiple responses (multiple drafts of the news release) based on the input data displayed (entered) in the input area 211.

[0034] Output areas 241 and 261 are display areas where multiple responses (multiple drafts of the news release) generated by the generative model 20 based on the input data displayed in input area 211 are displayed, respectively. In the example shown in Figure 4, news release draft 1 generated by the generative model 20 is displayed in output area 241, and news release draft 2 generated by the generative model 20 is displayed in output area 261. Note that either output area 241 or 261 will be displayed on the UI screen 201 by switching tabs using operation tabs 242 and 262.

[0035] Operation tab 242 is a UI component (UI image) for displaying output area 241 (document draft 1) on UI screen 201 via tab switching. Operation tab 262 is a UI component (UI image) for displaying output area 261 (document draft 2) on UI screen 201 via tab switching. In the example shown in Figure 4, operation tab 242 is selected, so output area 241 is placed on output area 261 and output area 241 is displayed on UI screen 201. If operation tab 262 is selected, output area 261 is placed on output area 241 and output area 261 is displayed on UI screen 201.

[0036] The output area 241 includes a predetermined set of answer items and answer fields associated with each of those answer items. In the example shown in Figure 4, the answer item "Title / Subtitle" is associated with answer field 243, the answer item "Lead" is associated with answer field 244, and the answer item "Background" is associated with answer field 245. Each answer field displays and outputs the content corresponding to the answer item, which is generated by the generative model 20.

[0037] The reception unit 103 receives input data entered by the user. The input data consists of input sentences entered by the user in each of a predetermined set of input fields. Specifically, the reception unit 103 receives input sentences entered by the user in each input field included in the input area 211 on the UI screen 201 using an input device 15 or the like as input data, and the display control unit 101 displays the input sentences entered in each input field.

[0038] Figure 5 shows an example of input data entered into the input area 211 of this embodiment. In addition, the example shown in Figure 5 also shows the input items belonging to the major category "Product Overview / Features," which were not explained in Figure 4.

[0039] In the example shown in Figure 5, the input field "XX Corporation" is entered and displayed in input field 212 for the input field "Who and who...". The input field "What..." is entered and displayed in input field 213 for the input field "Image conversion service utilizing generation AI". The input field "When..." is entered and displayed in input field 214 for the input field "December 2024". Furthermore, the input field "Product / service name is...", which belongs to the major category "Product overview / features", is entered and displayed in input field 215 for the input field "XX Image Conversion AI". Furthermore, the input field "Target customer base..." is entered and displayed in input field 216 for the input field "Customers who handle images in their work". Furthermore, the input field "Product / service is..." is entered and displayed in input field 217 for the input field "I want to convert images and photos to high resolution at low cost". Furthermore, regarding the input field "For customers who will be implementing the system...", the input text "It can achieve twice the resolution of other companies' systems. Costs can be halved." is entered and displayed in input field 218.

[0040] The input fields used to generate the news release draft are not limited to the examples above. For example, under a major category such as "Product Overview / Features," you may further use at least one of the following input fields: "Features of this product / service," "Challenges and Needs of Consumers Using the Product / Service Through Customers," "Benefits for Consumers," and "Price of the Product / Service." Alternatively, you may create a major category such as "Background and Challenges of Product / Service Development," and under that category, at least one of the following input fields may be used: "Social Issues and Trends Underlying the Provision of the Product / Service," "Sources Providing the Basis for Social Issues and Trends," "Contribution of the Product / Service to Social Issues and Trends," "Internal Policies and Plans that Form the Basis for Communicating the Product / Service," and "Past Businesses, Products, and Services Related to this Product / Service." Furthermore, you may create a major category such as "Future Developments," and under that category, at least one of the following input fields may be used: "Sales Target or Number of Installation Target for the Product / Service," and "Future Feature Additions."

[0041] The generation unit 105 generates multiple prompts that include the input data received by the reception unit 103. In this embodiment, the generation unit 105 generates multiple prompts by including the input data in each of several templates, each with different content. For example, several templates with different content are stored in an auxiliary storage device 13 or the like, and when the user selects the generate button 231 (see Figure 4), the generation unit 105 generates multiple prompts by including the input data received by the reception unit 103 in each template. The templates stored in the auxiliary storage device 13 or the like may be added by the user. The generation unit 105 may also generate multiple prompts by including input items in the input data (input sentences). In other words, multiple prompts may be generated using input data that consists of sets of several input items and the input sentences for those input items.

[0042] Figures 6 and 7 show examples of prompt templates in this embodiment, and the template 310 shown in Figure 6 and the template 320 shown in Figure 7 have different contents.

[0043] In the template 310 shown in Figure 6, the instruction content indicated by reference numeral 311 corresponds to the task (generating a draft of a news release) requested from the generation model 20, the instruction content indicated by reference numeral 313 corresponds to the output format of the response requested from the generation model 20, reference numeral 315 indicates the part in which the input data received by the reception unit 103 is inserted, and the instruction content indicated by reference numeral 317 corresponds to the task execution rules (conditions for generating the main text).

[0044] In template 320 shown in Figure 7, the contents indicated by reference numerals 311, 313, and 315 are the same (common) as those in template 310 shown in Figure 6. On the other hand, the task execution rules (conditions for generating the main text) indicated by reference numeral 327 differ from those in template 310 shown in Figure 6 (the contents indicated by reference numeral 317). Specifically, the task execution rules indicated by reference numeral 327 include the underlined contents in addition to those indicated by reference numeral 317, making the task execution rules indicated by reference numeral 327 more detailed than those indicated by reference numeral 317.

[0045] When the generation unit 105 generates multiple prompts using templates 310 and 320, the output format of the responses requested from the generation model 20 will be the same (common). Therefore, each of the multiple prompts will instruct the generation model 20 to provide responses with multiple identical response items. On the other hand, the execution rules for the tasks requested from the generation model 20 (conditions for generating the main text) are different for each other. Therefore, even if the response items (background, features, price, future developments) are the same, the content of the responses generated by the generation model 20 is expected to be different for each other.

[0046] Furthermore, the multiple response items displayed in the output area 241 mentioned above are set to correspond to the response output format indicated by the symbol 313.

[0047] When the generation unit 105 generates multiple prompts, the acquisition unit 107 inputs these multiple prompts into the generation model 20 to obtain multiple responses. Specifically, the acquisition unit 107 inputs each of the multiple prompts generated by the generation unit 105 into the generation model 20. The generation model 20 generates a response for each input prompt and outputs it to the information processing device 10. The acquisition unit 107 then retrieves the response for each prompt from the generation model 20.

[0048] As mentioned above, since the multiple prompts generated by the generation unit 105 have different content from each other, it is expected that the responses obtained by the acquisition unit 107 will also have different content from each other. Specifically, although the response items (background, features, price, future developments) are common to each response obtained by the acquisition unit 107, it is expected that at least one of the response contents will be different from each other. In other words, each of the multiple responses obtained by the acquisition unit 107 has multiple response items that are identical to each other, and it is expected that the response content of at least one of the multiple response items will be different from the other responses.

[0049] The output unit 109 outputs the multiple responses obtained by the acquisition unit 107. Specifically, the output unit 109 displays the multiple responses obtained by the acquisition unit 107 in output areas 241 and 261, respectively. The output unit 109 may also output the responses (drafts) displayed in output areas 241 and 261 to a file based on user input. Furthermore, the user may be able to edit or evaluate the responses (drafts) displayed in output areas 241 and 261.

[0050] For example, suppose the first draft is the response generated by the generation model 20 based on the first prompt generated using the input data shown in Figure 5 and the template 310 shown in Figure 6. Since the first draft contains the response content according to the response output format indicated by reference numeral 313, the output unit 109 displays and outputs the response content for each response item included in the first draft in the response field of the corresponding response item in the output area 241.

[0051] For example, suppose the second draft is the response generated by the generation model 20 based on the second prompt generated using the input data shown in Figure 5 and the template 320 shown in Figure 7. Since the second draft contains the response content according to the response output format indicated by reference numeral 313, the output unit 109 displays and outputs the response content for each response item included in the second draft in the response field of the corresponding response item in the output area 261.

[0052] Figure 8 shows an example of a response (first draft) displayed in output area 241 of this embodiment, and Figure 9 shows an example of a response (second draft) displayed in output area 261 of this embodiment. In the examples shown in Figures 8 and 9, the response item "Features," which was omitted from explanation in Figure 4, is also shown.

[0053] In the output area 241 shown in Figure 8, the answer fields 243 for the "Title / Subtitle," 244 for the "Lead Sentence," 245 for the "Background," and 246 for the "Features" section display the answers to the first draft generated by the generative model 20. Similarly, in the output area 261 shown in Figure 9, the answer fields 263 for the "Title / Subtitle," 264 for the "Lead Sentence," 265 for the "Background," and 266 for the "Features" section display the answers to the second draft generated by the generative model 20.

[0054] Note that, unlike template 320 shown in Figure 7, template 310 in Figure 6 does not specify the format for company names or the style of sentence endings. Therefore, in the response items "Background" and "Features" included in output area 241 shown in Figure 8, "Our Company" is used as the company name, and a mix of polite and informal styles are used for sentence endings. On the other hand, in the response items "Background" and "Features" included in output area 261 shown in Figure 9, the company name is standardized to "XX Corporation," and the sentence endings are standardized to the polite style.

[0055] The display control unit 101 displays multiple responses output by the output unit 109 in a switchable manner. For example, when the user selects the operation tab 242 (see Figure 4), the display control unit 101 displays the output area 241 showing the first draft on the UI screen 201. Also, for example, when the user selects the operation tab 262 (see Figure 4), the display control unit 101 displays the output area 261 showing the second draft on the UI screen 201.

[0056] Next, the operation of the data generation system of this embodiment will be described.

[0057] Figure 10 is a flowchart showing an example of the data generation process performed by the information processing device 10 of this embodiment.

[0058] First, the display control unit 101 displays the UI screen of the document generation AI service on the display device 16 when the user initiates the document generation AI service (step S101).

[0059] Next, the reception unit 103 receives the input text entered by the user in each input field included in the input area 211 on the UI screen 201 as input data (step S103), and the display control unit 101 displays the input text entered in each input field.

[0060] Next, the generation unit 105 generates multiple prompts by including input data in each of several templates, each with different content (step S105).

[0061] Next, when the generation unit 105 generates multiple prompts, the acquisition unit 107 inputs these multiple prompts into the generation model 20 to obtain multiple responses (step S107).

[0062] Next, the output unit 109 displays each of the multiple responses acquired by the acquisition unit 107 in the output area (step S109).

[0063] Next, the display control unit 101 displays multiple output areas from which the answers have been displayed by the output unit 109 in a switchable manner (step S111).

[0064] As described above, in this embodiment, multiple prompts are generated from user input, and these generated prompts are input to a generation model to obtain and output multiple answers. Therefore, according to this embodiment, it is possible to obtain multiple answers from different perspectives from the generation model at once, and the user can reduce the number of interactions with the generation model, thereby improving the efficiency of obtaining answers that satisfy the user's purpose from the generation model.

[0065] This is particularly useful in improving the efficiency of obtaining a response that satisfies the user's objective from a generative model, especially in embodiments like this one, where input items are predetermined and the user cannot freely write prompts.

[0066] Furthermore, in this embodiment, multiple responses from different perspectives obtained from the generative model are displayed in a switchable manner, allowing the user to compare and review multiple responses, and to efficiently perform simultaneous processing and comparative analysis of multiple tasks.

[0067] Furthermore, as in this embodiment, by having the generation model generate and display multiple answers where the answer items are common but the content of the answers differs from one another, the user can, for example, compare and consider various forms of presentation and decide which form of presentation to use.

[0068] (Variation 1) In the above embodiment, an example was described in which the generation model 20 generates multiple answers that have common answer items (answer format) but different answer content. In Modification 1, an example is described in which the generation model 20 generates multiple answers that have different answer formats. In the following, as examples of multiple answers with different answer formats, we will use the chapter-divided text format, table format, and report format answers described in the above embodiment as examples, but we are not limited to these. In Modification 1, we will mainly describe the parts that differ from the above embodiment, and will omit the explanation of parts that are the same as the above embodiment.

[0069] Figures 11 and 12 show an example of a prompt template for Modification Example 1.

[0070] The template 330 shown in Figure 11 is a template used to generate the third prompt. The third prompt is a prompt to instruct the generation model 20 to provide a response in tabular format. In the template 330 shown in Figure 11, the instruction content indicated by reference numeral 331 corresponds to the task requested from the generation model 20 (generating a response in tabular format), the instruction content indicated by reference numeral 333 corresponds to the output format of the response requested from the generation model 20, and reference numeral 315 indicates the part into which the input data received by the reception unit 103 is inserted.

[0071] The template 340 shown in Figure 12 is a template used to generate the fourth prompt. The fourth prompt is a prompt to instruct the generation model 20 to respond in report format. In the template 340 shown in Figure 12, the instruction content indicated by reference numeral 341 corresponds to the task requested from the generation model 20 (generating a response in report format), and reference numeral 315 indicates the part in which the input data received by the reception unit 103 is inserted.

[0072] In Modification 1, the second prompt described in the above embodiment is used as a prompt to instruct the generation model 20 to respond in a chapter-divided text format. In other words, the template 320 shown in Figure 7 is used.

[0073] The generation unit 105 generates multiple prompts using templates 320, 330, and 340. Specifically, the generation unit 105 generates the second to fourth prompts by including the input data received by the reception unit 103 in each of the templates 320, 330, and 340. When the generation unit 105 generates multiple prompts using templates 320, 330, and 340, the output formats of the responses requested from the generation model 20 will be different from each other. Therefore, each of the multiple prompts instructs the generation model 20 to provide responses in different response formats, and the responses generated by the generation model 20 are expected to have different response formats.

[0074] The acquisition unit 107 inputs the second to fourth prompts generated by the generation unit 105 to the generation model 20, and obtains chapter-divided text-format answers, tabular answers, and report-format answers from the generation model 20. In the modified example 1, the generation model 20 is assumed to be multimodal, but is not limited to this.

[0075] The output unit 109 outputs the multiple responses obtained by the acquisition unit 107. For example, in the UI screen 201 described in Figure 4, instead of output area 241, output areas 271 and 281 are provided, and the multiple responses obtained by the acquisition unit 107 are displayed and output to output areas 261, 271, and 281, respectively.

[0076] Figure 13 shows an example of a tabular response displayed in output area 271 of Modification 1, and Figure 14 shows an example of a report-format response displayed in output area 281 of Modification 1. Note that the examples in Figures 13 and 14 show tabular and report-format responses, but the specific content of the responses is omitted. Furthermore, the chapter-based text-format response has already been explained in Figure 9.

[0077] According to Modification 1, various types of materials can be generated simultaneously from a single input source. For example, chapter-formatted text responses can be used for company newsletters, tabular responses for departmental meetings, and report-formatted responses for management reports. All of these materials can be generated simultaneously from a single input source.

[0078] (Modification 2) In the above modified example 1, both the third and fourth prompts instructed the generation model 20 to generate an answer from the input data received by the reception unit 103. However, it is also possible to instruct the generation model 20 to generate an answer from the chapter-divided text format answer generated by the generation model 20 based on the third prompt. In other words, in the template 330 shown in Figure 11 and the template 340 shown in Figure 12, the portion indicated by reference numeral 315 may be replaced with an answer generated by the generation model 20 in chapter-divided text format, rather than the input data received by the reception unit 103. In this way, the generation model 20 can efficiently generate multiple answers with different answer formats.

[0079] (Variation 3) In the above embodiment, an example was described in which multiple answers generated by the generation model 20 are displayed in a switchable format, but they may also be displayed in a format that allows for direct comparison. Figure 15 shows an example of the UI screen 401 of the document generation AI service of Modification 3. In the UI screen 201 of Figure 4, either output area 241 or 261 was displayed on the UI screen 201 by switching tabs using operation tabs 242 and 262, but as in the UI screen 401 of Figure 15, output areas 241 and 261 may be displayed side by side. In other words, the display control unit 101 of Modification 3 may display multiple answers displayed in output areas 241 and 261 in a way that allows for comparison.

[0080] According to Modification 3, since multiple responses generated by the generative model 20 can be directly compared, simultaneous processing and comparative analysis of multiple tasks can be performed more efficiently.

[0081] (Modification 4) In the above embodiment, an example of generating multiple answers using a single generative model 20 was described, but multiple answers may be generated using multiple generative models. That is, the acquisition unit 107 in Modification 4 may acquire multiple answers by inputting multiple prompts to multiple generative models. For example, a general-purpose generative model and a high-precision generative model may be used as the multiple generative models, or a general-purpose generative model and a generative model that has learned or been further trained with specialized knowledge may be used. Also, for example, in Modification 1, a general-purpose generative model and an image-specific generative model may be used.

[0082] Alternatively, for example, the generative model 20 may be configured to change its parameters for each prompt to generate an answer. This method achieves the same effect as generating multiple answers using multiple generative models.

[0083] (Variation 5) In the above embodiment, an example was described in which multiple prompts are generated from user input and the generated multiple prompts are input into a generation model to obtain multiple answers. However, it is also possible to generate a single prompt from user input and input the generated single prompt into the generation model multiple times to obtain multiple answers. In other words, in Modification 5, the generation unit 105 generates one prompt including input data, and the acquisition unit 107, when one prompt is generated, inputs the single prompt into the generation model 20 multiple times to obtain multiple answers. In this way, it is possible to check what kind of variation occurs in the answers generated by the generation model 20. It is also possible to input the generated single prompt into multiple generation models to obtain multiple answers.

[0084] Furthermore, the above embodiment may be combined with Modification 5. That is, multiple prompts may be generated from user input, and multiple answers may be obtained by inputting the generated multiple prompts into the generation model multiple times. Note that it is not necessary to input all of the generated multiple prompts into the generation model multiple times; at least one of the generated multiple prompts may be input into the generation model multiple times, and the other prompts may be input into the generation model once. In other words, the acquisition unit 107 may input prompts into the generation model multiple times to obtain multiple answers and / or generate multiple different prompts, and input these multiple prompts into the generation model to obtain multiple answers.

[0085] (Experimental variation 6) In the above embodiments and their variations, the example of implementing the generation model 20 as a cloud service was used for explanation, but the generation model 20 may also be included in the information processing device 10. In this case, the generation model 20 can be implemented, for example, as an AI chat service that fine-tunes a small language model (SLM).

[0086] (program) The programs executed by the information processing devices of the above embodiments and each of the above modifications are provided in installable or executable file format, stored on a computer-readable storage medium such as a CD-ROM, CD-R, memory card, DVD, or flexible disk (FD).

[0087] Furthermore, the programs executed by the information processing devices of the above embodiments and their respective modifications may be stored on a computer connected to a network such as the Internet and provided by allowing downloads via the network. Alternatively, the programs executed by the information processing devices of the above embodiments and their respective modifications may be provided or distributed via a network such as the Internet. Furthermore, the programs executed by the information processing devices of the above embodiments and their respective modifications may be pre-installed in ROM or the like and provided.

[0088] The programs executed in the information processing devices of the above embodiments and their respective modifications are configured as modules for realizing the above-described parts on a computer. In actual hardware, for example, the CPU reads the program from the HDD into RAM and executes it, thereby realizing the above-described parts on the computer.

[0089] The above embodiments and their respective modifications are merely examples of how this disclosure may be implemented, and they do not restrict the technical scope of this disclosure. Therefore, this disclosure can be implemented in various ways without departing from its essence or its main features. For example, the above embodiments and their respective modifications may be combined as appropriate on a component basis. Also, for example, some components may be removed from the total components in the above embodiments and their respective modifications.

[0090] This disclosure also includes the following aspects:

[0091] (1) A reception unit that receives input data entered by the user, A generation unit that generates one or more prompts including the aforementioned input data, When the aforementioned prompt 1 is generated, the acquisition unit inputs the prompt 1 multiple times into the generation model to obtain multiple answers, and when multiple prompts are generated, the acquisition unit inputs the multiple prompts into the generation model to obtain multiple answers. An output unit that outputs the aforementioned multiple answers, An information processing device equipped with the following features.

[0092] (2) The generation unit generates the multiple prompts by including the input data in each of the multiple templates which have different contents from each other. The information processing device described in (1) above.

[0093] (3) Each of the above multiple responses has multiple response items that are identical to each other, In each of the above multiple responses, the content of at least one of the above multiple response items differs from the other responses. The information processing device described in (2) above.

[0094] (4) Each of the multiple prompts instructs the generation model to provide an answer for the multiple response items. The information processing device described in (3) above.

[0095] The aforementioned multiple responses differ from each other in their response format. The information processing device described in (2) above.

[0096] Each of the aforementioned prompts instructs the generative model to provide a response in a different format from the others. The information processing device described in (5) above.

[0097] The aforementioned input data consists of input sentences entered by the user in each of a predetermined set of input fields. The information processing device described in (1) above.

[0098] The acquisition unit inputs the multiple prompts to multiple generative models to acquire the multiple responses. The information processing device described in (1) above.

[0099] The system further includes a display control unit that displays the multiple answers in a switchable or comparative manner. An information processing device as described in any one of the above items (1) to (8).

[0100] The reception department receives input data entered by the user in a reception step, The generation unit generates one or more prompts including the input data, The acquisition step includes: when the acquisition unit generates the first prompt, it inputs the first prompt multiple times into the generation model to obtain multiple answers; when multiple prompts are generated, it inputs the multiple prompts into the generation model to obtain multiple answers; The output unit performs an output step that outputs the plurality of answers, Information processing methods including

[0101] A reception step that accepts input data entered by the user, A generation step that generates one or more prompts including the aforementioned input data, If the aforementioned prompt 1 is generated, the acquisition step involves inputting the aforementioned prompt 1 multiple times into the generation model to obtain multiple answers, and if the aforementioned multiple prompts are generated, inputting the aforementioned multiple prompts into the generation model to obtain multiple answers. An output step that outputs the aforementioned multiple answers, A program that causes a computer to execute something. [Explanation of symbols]

[0102] 1. Document Generation System 2 Network 10 Information Processing Devices 20 Generative Models 101 Display Control Unit 103 Reception Department 105 Generation part 107 Acquisition Department 109 Output section

Claims

1. A reception unit that receives input data entered by the user, A generation unit that generates a prompt including the aforementioned input data, An acquisition unit that inputs a prompt multiple times into a generation model to obtain multiple answers and / or generates multiple different prompts, and inputs those multiple prompts into the generation model to obtain multiple answers, An output unit that outputs the aforementioned multiple answers, An information processing device equipped with the following features.

2. The generation unit generates the multiple prompts by including the input data in each of the multiple templates, each having different content. The information processing apparatus according to claim 1.

3. Each of the aforementioned multiple responses has multiple response items that are identical to each other. In each of the above multiple responses, the content of at least one of the above multiple response items differs from the other responses. The information processing apparatus according to claim 2.

4. Each of the aforementioned multiple prompts instructs the generation model to provide an answer for the aforementioned multiple response items. The information processing apparatus according to claim 3.

5. The aforementioned multiple responses differ from each other in their response format. The information processing apparatus according to claim 2.

6. Each of the aforementioned prompts instructs the generative model to provide a response in a different format from the others. The information processing apparatus according to claim 5.

7. The aforementioned input data consists of input sentences entered by the user in each of a predetermined set of input fields. The information processing apparatus according to claim 1.

8. The acquisition unit inputs the multiple prompts to multiple generative models to acquire the multiple responses. The information processing apparatus according to claim 1.

9. The system further includes a display control unit that displays the multiple answers in a switchable or comparative manner. The information processing apparatus according to any one of claims 1 to 8.

10. The reception department receives input data entered by the user in a reception step, The generation unit performs a generation step of generating a prompt that includes the input data, The acquisition step involves the acquisition unit inputting prompts multiple times into the generation model to obtain multiple answers and / or generating multiple different prompts, and inputting these multiple prompts into the generation model to obtain multiple answers. The output unit performs an output step that outputs the plurality of answers, Information processing methods including

11. A reception step that accepts input data entered by the user, A generation step that generates a prompt including the aforementioned input data, A retrieval step involves inputting prompts multiple times into a generation model to obtain multiple answers and / or generating multiple different prompts, and inputting those multiple prompts into the generation model to obtain multiple answers. An output step that outputs the aforementioned multiple answers, A program that causes a computer to execute something.

Images