Information processing device, control method for information processing device, and program

Abstract

This invention provides an information processing device, a control method for the information processing device, and a program that streamline the process of obtaining comprehensive answers using multiple language models, thereby improving user convenience. The information processing device 1 includes a response result acquisition unit 33 that sends a request prompt containing the user's request to a plurality of language models 3 (for example, language models 3-1 to 3-3) and acquires a response result for the request from each of the plurality of language models 3, and an integration processing unit 35 that generates an integration instruction prompt that instructs the language model 3 to integrate the response results of each of the plurality of language models 3, and the response result acquisition unit 33 acquires an integrated response result that reflects the response results of the plurality of language models 3 by sending the integration instruction prompt to the language model 3.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, a control method for an information processing apparatus, and a program.

Background Art

[0002] Conventionally, information processing technologies that use generative AI-based language models such as LLMs to perform data analysis and generate ideas have been known. For example, Patent Document 1 describes this type of technology.

[0003] Patent Document 1 relates to a data analysis apparatus that analyzes various data. The data analysis apparatus of Patent Document 1 generates a system prompt by assigning input information in the form of natural language by the user and analysis background information to a prompt template, and transmits text information including the input information and the system prompt to a pre-trained language model that can interpret natural language, and generates display data based on the response output from the pre-trained language model by the transmission.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Generative AI-based language models such as LLMs output different answers even when the same prompt is input depending on the learning data and parameters. For example, even when the same prompt is input to GPT-4o and Claude (registered trademark) respectively, the answer results are different between GPT-4o and Claude.

[0006] When users seek comprehensive ideas or solutions, utilizing multiple language models is a viable option. However, consolidating the results from multiple language models requires sending prompts to each language model provider, retrieving the results from each, and then organizing them—a time-consuming process.

[0007] This invention has been made in view of the above circumstances, and aims to provide an information processing device, a control method for the information processing device, and a program that can streamline the process of obtaining a comprehensive answer using multiple language models and improve user convenience. [Means for solving the problem]

[0008] To achieve the above objective, one aspect of the present invention is an information processing device comprising: a response result acquisition unit that transmits a request prompt containing the user's request to a plurality of language models and acquires a response result for the request from each of the plurality of language models; and an integration processing unit that generates an integration instruction prompt that instructs the language models to integrate the response results of each of the plurality of language models, wherein the response result acquisition unit acquires an integrated response result that reflects the response results of the plurality of language models by transmitting the integration instruction prompt to the language models.

[0009] Furthermore, one aspect of the present invention is a control method for an information processing device, which includes: a response result acquisition step of sending a request prompt containing the user's request to a plurality of language models and acquiring a response result for the request from each of the plurality of language models; an integration processing step of generating an integration instruction prompt that instructs the language models to integrate the response results of each of the plurality of language models; and an integrated response result acquisition step of acquiring an integrated response result that reflects the response results of the plurality of language models by sending the integration instruction prompt to the language models.

[0010] Furthermore, one aspect of the present invention is a program for causing a computer to execute the following steps: a response result acquisition step that sends a request prompt containing the user's request to a plurality of language models and acquires a response result for the request from each of the plurality of language models; an integration processing step that generates an integration instruction prompt that instructs the language models to integrate the response results of each of the plurality of language models; and an integrated response result acquisition step that sends the integration instruction prompt to the language models and acquires an integrated response result that reflects the response results of the plurality of language models. [Effects of the Invention]

[0011] According to the present invention, it is possible to provide an information processing device, a control method for the information processing device, and a program that can streamline the process of obtaining a comprehensive answer using multiple language models and improve user convenience. [Brief explanation of the drawing]

[0012] [Figure 1] This figure shows a language model utilization system to which the information processing device according to the first embodiment of the present invention is applied. [Figure 2] This is a block diagram showing the hardware configuration of the information processing device according to the first embodiment. [Figure 3] This is a functional block diagram showing an example of the functional configuration of the information processing device according to the first embodiment. [Figure 4] This figure shows an example of a request input screen displayed on the user terminal of the first embodiment. [Figure 5] This figure shows an example of a language model selection screen displayed on the user's terminal. [Figure 6] This figure shows an example of a screen displaying the answer results on the user's terminal. [Figure 7] This figure shows an example of the configuration of an integrated instruction prompt. [Figure 8] This flowchart shows an example of the flow of the process of providing answers to a user, which is performed by an information processing device. [Figure 9] It is a functional block diagram showing an example of the functional configuration of an information processing apparatus according to the second embodiment. [Figure 10] It is a schematic diagram showing the flow of data in the answer presentation process by the information processing apparatus of the second embodiment. [Figure 11] It is a diagram showing an example of a request input screen displayed on the user terminal of the second embodiment.

Embodiments for Carrying Out the Invention

[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0014] <First Embodiment> First, the overall system configuration will be described. FIG. 1 is a diagram showing a language model utilization system 100 to which an information processing apparatus 1 according to the first embodiment of the present invention is applied. The language model utilization system 100 provides an idea brainstorming service that generates and presents answers that serve as clues for sorting ideas and solving problems in response to requests from users, using a plurality of language models 3-1 to 3-n. Note that the purpose of the service is merely an example, and the present invention can also be applied to services other than the idea brainstorming service.

[0015] The language model utilization system 100 is realized by an information processing apparatus 1 that communicates with the user terminal 2 via a communication network such as the Internet.

[0016] The user terminal 2 is an information processing apparatus used by the user. The user terminal 2 is composed of a personal computer, a tablet, a smartphone, etc. The user terminal 2 may exchange various information through a web browser, or may exchange various information with the information processing apparatus 1 by a pre-installed program.

[0017] The information processing device 1 is a server that communicates with a plurality of language models 3-1 to 3-n with respect to the user terminal 2 and generates an answer to the user's request content. The information processing device 1 cooperates with the language models 3-1 to 3-n, which are external services, through, for example, an API (Application Programming Interface) or the like.

[0018] Note that the language models 3-1 to 3-n are not necessarily limited to those provided as external services. For example, part or all of the language models 3-1 to 3-n may be implemented in the information processing device 1, or they may be implemented in other servers within the same system as the information processing device 1. Also, among the plurality of language models 3-1 to 3-n, part may be implemented in the information processing device 1, and the other part may be used as an external service. In any configuration, the basic processes of transmitting the request prompt by the answer result acquisition unit 33 and

[0019] acquiring the answer result are executed in the same manner. It may be a configuration.

[0020] The language models 3-1 to 3-n each represent a different generative AI language model. The language models 3-1 to 3-n may be LLM (Large Language Model) or SLM (Small Language Model). The language models 3-1 to 3-n in FIG. 1 indicate candidates for use, and in this example, the language models 3-1 to 3-3 are used for the answer presented to the user.

[0021] In the following explanation, if a feature is common to language models 3-1 to 3-n and does not require distinction, it may simply be referred to as language model 3.

[0022] <Hardware Configuration> Next, an example of the hardware constituting the information processing device 1 will be described. Figure 2 is a block diagram showing the hardware configuration of the information processing device 1 according to the first embodiment. The information processing device 1 includes a CPU (Central Processing Unit) 11 as a processor, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a bus 14, an input / output interface 15, an output unit 16, an input unit 17, a storage unit 18, a communication unit 19, and a drive 20.

[0023] The CPU 11 executes various processes according to the program stored in the ROM 12 or the program loaded from the storage unit 18 into the RAM 13. The RAM 13 also stores data necessary for the CPU 11 to execute various processes. The CPU 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output interface 15 is also connected to this bus 14.

[0024] The input / output interface 15 is connected to an output unit 16, an input unit 17, a storage unit 18, a communication unit 19, and a drive 20. The output unit 16 consists of a display, speakers, etc., and outputs various information as images and sounds. The input unit 17 consists of a keyboard, mouse, etc., and inputs various information. The storage unit 18 consists of a hard disk, DRAM (Dynamic Random Access Memory), etc., and stores various data. The communication unit 19 communicates with other devices via a network, including the Internet.

[0025] The drive 20 is appropriately equipped with removable media 21, which may consist of a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory. Programs read from the removable media 21 by the drive 20 are installed in the storage unit 18 as needed. The removable media 21 can also store various types of data stored in the storage unit 18, just like the storage unit 18.

[0026] The hardware configuration described here is merely an example. The computer described in this embodiment, including the information processing device 1, may have the same configuration as that shown in Figure 2, or it may have a different configuration. Furthermore, the computer may consist of two or more computers. The user terminal 2 in Figure 1 is, for example, a personal computer, tablet, or personal computer having a configuration similar to the hardware configuration shown in Figure 2.

[0027] <Functional configuration> Next, the functional configuration of the information processing device 1 will be described. Figure 3 is a functional block diagram showing an example of the functional configuration of the information processing device 1 according to the first embodiment. As shown in Figure 3, the information processing device 1 includes an input / output processing unit 31, a language model specification unit 32, an answer result acquisition unit 33, an integration target setting unit 34, and an integration processing unit 35 as functional units realized on the processor (CPU 11).

[0028] The input / output processing unit 31 performs processes to receive information input from the user and to present information to the user. In this embodiment, the input / output processing unit 31 performs processes to receive information input from the user through a website on the internet accessed by the user via the user terminal 2, and also performs processes to present information to the user.

[0029] For example, the input / output processing unit 31 performs processes to receive information input from the user and to present information to the user. Specifically, the input / output processing unit 31 generates display data for various screens displayed on the user terminal 2 (some or all of the request input screen, language model selection screen, response result presentation screen, etc., described later) and sends it to the user terminal 2 to control the display of these screens. In addition, the input / output processing unit 31 allows the user to confirm and edit the integrated instruction prompt generated by the integrated processing unit 35. It also executes controls to display the content in any possible format.

[0030] Figure 4 shows an example of a request input screen displayed on the user terminal 2 of the first embodiment. The screen in Figure 4 displays the language model display unit 101, the request content input field 102, the language model specification operation unit 103, the user information display unit 104, etc. The user connects to a website that displays the request input screen by accessing a URL (Uniform Resource Locator), for example.

[0031] The Language Model Display Unit 101 displays the language model 3 used when the user requests a response. In this example, ChatGPT-4.1nano, Gemini 2.0Flash, and Gemini 2.5Flash are specified as the language model 3 to be used. The Request Input Field 102 is where the user enters the request in natural language text. In this example, the user has entered the request, "Please provide name suggestions for a new function using generative AI." When the send button 124 in the Request Input Field 102 is operated, the request is sent to the Information Processing Device 1. The Language Model Specification Operation Unit 103 is an operation button that transitions to a screen where the user can specify the language model 3 to be used. The Language Model Specification Operation Unit 103 can also save combinations of language model 3 that the user frequently uses. The User Information Display Unit 104 displays the user ID and information about the tickets the user possesses to use the language model 3. Tickets are granted by the Information Processing Device 1 in exchange for money, points, etc.

[0032] The language model selection unit 32 executes the process of setting the language model 3 specified by the user from among a plurality of candidate language models 3 as the destination for sending the request prompt. In this embodiment, the language model 3 is specified based on the user's selection result on the language model selection screen that is displayed when the language model selection operation unit 103 on the request input screen in Figure 4 is selected. The language model selection unit 32 obtains the user's language model 3 selection result through the input / output processing unit 31.

[0033] Figure 5 shows an example of a language model selection screen displayed on the user terminal 2. The screen in Figure 5 displays a basic language model selection unit 105, an advanced language model selection unit 106, a confirmation instruction unit 107, etc. The basic language model selection unit 105 displays a list of language models 3 that can be selected even if the user does not possess tickets. The advanced language model selection unit 106 displays a list of language models 3 that can be selected if the user possesses tickets. In this example, the language models 3 are displayed in three stages according to the amount of tickets consumed. The top row of the advanced language model selection unit 106 displays language models 3 that consume 1 ticket, the middle row displays language models 3 that consume 25 tickets, and the bottom row displays language models 3 that consume 80 tickets. The user decides which language model 3 to use by selecting a language model 3 displayed in the basic language model selection unit 105 or the advanced language model selection unit 106. In this example, the user can select up to 6 language models 3 to use simultaneously.

[0034] The confirmation instruction unit 107 is an on-screen operation button used to instruct the system to set the language model 3 selected by the basic language model selection unit 105 or the advanced language model selection unit 106 as the language model 3 to be used. By operating the confirmation instruction unit 107 with a language model 3 selected, the user's specified language model 3 is sent to the input / output processing unit 31. When the input / output processing unit 31 receives the information of the language model 3 specified by the user, the language model specification unit 32 sets the language model 3 to be used as the language model 3 specified by the user. In this example, ChatGPT-4.1nano, Gemini 2.0Flash, and Gemini 2.5Flash are selected.

[0035] The response result acquisition unit 33 sends a request prompt containing the user's request to multiple language models 3 and executes a process to acquire a response result for the request from each of the multiple language models 3. In this embodiment, "transmission" is not limited to transmission to an external source via a communication network, but is used as a broad concept that includes data transfer between different modules within the same device, inter-process communication, and data transfer to other servers within the system.

[0036] In this specification, “prompt” means input information used to request processing from a language model or other natural language processing system. A prompt is not limited to instructional text in natural language, but is a broad concept that includes any form of input information that a language model can interpret, such as structured data, parameters, control signals, API call arguments, audio data, image data, or combinations thereof.

[0037] The method for obtaining the user's request details will be explained. In this embodiment, the user's request details are obtained by the input / output processing unit 31 through the request details input field 102 on the request input screen in Figure 4. The response result acquisition unit 33 sends a prompt to the designated language model 3 requesting a response to the request details obtained by the input / output processing unit 31. In the above example, a prompt containing the request details "Suggest names for a new function using generation AI" is sent to ChatGPT-4.1nano (language model 3-1), Gemini 2.0Flash (language model 3-2), and Gemini 2.5Flash (language model 3-3), respectively.

[0038] The response result acquisition unit 33 acquires response results from each language model 3. In the example above, it acquires response results for the request from the three specified language models 3-1 to 3-3. The response results acquired by the response result acquisition unit 33 are presented to the user via the website by the input / output processing unit 31.

[0039] Figure 6 shows an example of a response result display screen shown on the user terminal 2. The screen in Figure 6 displays response result display sections 111a to 111c, individual additional request input instruction sections 112a to 112c, batch additional request input field 113, response result selection sections 114a to 114c, practice instruction section 115, etc.

[0040] The response result display sections 111a to 111c are the parts that display the response results for each language model 3. The number of response result display sections 111a to 111c is the number of language models 3 specified in advance. In this example, three language models 3-1 to 3-3 were specified, so three response result display sections 111a to 111c will be displayed. Response result display section 111a corresponds to language model 3-1, response result display section 111b corresponds to language model 3-2, and response result display section 111c corresponds to language model 3-3.

[0041] Each of the response result presentation units 111a to 111c displays text indicating the user's request at the beginning, and below the text indicating the request, the responses for language models 3-1 to 3-3 are displayed. In this example, response result presentation unit 111a corresponds to ChatGPT-4.1nano (language model 3-1), response result presentation unit 111b corresponds to Gemini 2.0Flash (language model 3-2), and response result presentation unit 111c corresponds to Gemini 2.5Flash (language model 3-3).

[0042] Furthermore, when the response result acquisition unit 33 receives a request from the user for further consideration of the response result, it sends a review instruction prompt including the review request to one or more language models 3 and executes a process to acquire a revised response result after further consideration. In this embodiment, the input / output processing unit 31 receives an additional review request from the user via the website, and the response result acquisition unit 33 sends a review instruction prompt including the received review request to the language models 3.

[0043] The individual additional request input instruction units 112a to 112c are on-screen operation buttons for the user to input further consideration requests for each language model 3. Each individual additional request input instruction unit 112a to 112c is set for each answer result presentation unit 111a to 111c. When the individual additional request input instruction units 112a to 112c are operated, an input field 116 for entering a question to the language model 3 is displayed. In this example, the individual additional request input instruction unit 112b is selected by the user, and an input field 116 for making a further consideration request is displayed. When the user enters a consideration request in the input field 116, the answer result acquisition unit 33 acquires the consideration request through the input / output processing unit 31 and sends a consideration instruction prompt including the consideration request to the corresponding language model 3 (in this example, Gemini 2.0Flash (language model 3-2)). The revised response result acquired from the language model 3 is displayed on the user terminal 2 in a format following the initial response result of the answer result presentation unit 111a to 111c.

[0044] The batch add request input field 113 is where the user enters a review request for all of the language models 3. When the user enters a review request in the batch add request input field 113 and operates the send button 125, the response result acquisition unit 33 acquires the review request through the input / output processing unit 31 and sends a review instruction prompt containing the review request to all of the language models 3 (in this example, language models 3-1 to 3-3). The revised response results acquired from each language model 3 are displayed by the input / output processing unit 31 in a format following the first response result of the corresponding response result presentation unit 111a to 111c.

[0045] The integration target setting unit 34 excludes answer results excluded by the user from the integration instructions issued by the integration processing unit 35, or performs processing to include only the answer results specified by the user in the integration. In this embodiment, when the integration target setting unit 34 detects an operation to exclude answer results by the user via the website, it excludes the excluded answer results from the integration target of the integration processing unit 35. The operation to exclude answer results is performed through the answer result selection units 114a to 114c.

[0046] The answer result selection units 114a to 114c are on-screen operation buttons that allow the user to select the answer result to use. Each answer result selection unit 114a to 114c is set for each answer result presentation unit 111a to 111c. In this example, they are indicated by an "x" mark, and when the user selects one, the selected answer result presentation unit 111a to 111c is deleted, and the integration target setting unit 34 excludes the selected answer result presentation unit 111a to 111c from the integration target. For example, if the answer result selection unit 114a corresponding to answer result presentation unit 111a is selected, answer result presentation unit 111a is deleted from the screen, only answer result presentation unit 111b and answer result presentation unit 111c are displayed, and the answer result of language model 3-1 is excluded from the integration target.

[0047] In the example in Figure 6, the integration target setting unit 34 removes the answer results from the integration target through the selection of the answer result selection units 114a to 114c, but the process is not limited to this. For example, checkboxes could be placed for each of the answer result presentation units 111a to 111c, and the answer results with the checkboxes checked could be designated as the answer results specified by the user, with the answer results with the checkboxes checked being the target for integration.

[0048] The integration processing unit 35 generates an integration instruction prompt that instructs the integration of the response results of the multiple language models 3. In this embodiment, the integration processing unit 35 generates the integration instruction prompt based on the user selecting the wall-hitting instruction unit 115 shown on the screen in Figure 6. In this embodiment, "integration" refers to the process of providing multiple response results to the language model and causing it to generate a new response by referring to them. The specific method of integration is left to the judgment of the language model, and as a result, it is intended that a comprehensive response reflecting the perspectives of the multiple response results is obtained.

[0049] Figure 7 shows an example of the configuration of an integrated instruction prompt. The integrated instruction prompt includes a guidance transmission unit 120 that provides guidance for responses to a specified number of language models 3, and response result transmission units 121 to 123 that transmit the responses of each language model.

[0050] The guidance transmission unit 120 includes instructions such as, "Analyze the output of the AIs shown below, extract the most important and useful information, and create one concise and comprehensive answer." It also includes the first policy, "Read the output of each AI and extract text that eliminates incorrect or inaccurate information," the second guideline, "Identify the main points of each text," the third guideline, "Eliminate redundant information and integrate the information to form a consistent message," and the fourth guideline, "If necessary, restructure the information, extract the key points in a clear and fluent manner, and summarize them."

[0051] The response result transmission units 121-123 are sections that show the response results generated from language model 3. In this example, response result transmission unit 121 shows the response result from language model 3-1, response result transmission unit 122 shows the response result from language model 3-2, and response result transmission unit 123 shows the response result from language model 3-3. Thus, the integrated instruction prompt includes the response results created by each language model 3, along with the creation guidelines.

[0052] In this embodiment, the input / output processing unit 31 displays the integrated instruction prompt generated by the integration processing unit 35 in the batch add request input field 113. By displaying the integrated instruction prompt shown in Figure 7 in the batch add request input field 113, the user can confirm the contents of the integrated instruction prompt. When the user operates the send button 125 in the batch add request input field 113, the response result acquisition unit 33 sends the integrated instruction prompt to each of the specified language models 3. As described above, if the response result acquisition unit 33 has performed a review, the revised response results will be included in the integration target.

[0053] <Processing flow> Next, referring to Figure 8, we will explain the process flow for presenting a response using the language model 3 in response to a user's request. Figure 8 is a flowchart showing an example of the process flow for presenting a response to a user, executed by the information processing device 1. Note that the flowchart in Figure 8 is merely an example, and the order and content of the processing may be changed.

[0054] In step S1, the language model selection unit 32 sets the language model 3 specified by the user from among several candidate language models 3 as the destination for sending the request prompt. The user may set the language model 3 to be used from the language model selection screen shown in Figure 5 displayed on the user terminal 2, or they may use the pre-specified language model 3 as is.

[0055] In step S2, the response result acquisition unit 33 acquires the request details entered by the user. The request details are the information entered in the request details input field 102 on the request input screen in Figure 4, and are acquired via the input / output processing unit 31.

[0056] In step S3, the response result acquisition unit 33 sends a request prompt containing the acquired request details to the specified multiple language models 3.

[0057] In step S4, the response result acquisition unit 33 acquires the response results from each language model 3 that sent the request prompt, and transmits the response results to the user terminal 2 via the input / output processing unit 31 for presentation to the user. In this embodiment, the response results from each language model 3 are displayed in each of the response result presentation units 111a to 111c in Figure 6.

[0058] In step S5, if a request for reconsideration is received from the user through the input fields 116 of the individual add request input instruction units 112a to 112c or the batch add request input field 113 in Figure 6 (step S5; Yes), the response result acquisition unit 33 executes the process of step S6, which sends a review instruction prompt containing the content of the reconsideration request to the corresponding language model 3. In this process of step S6, if a review request is entered in the input fields 116 of the individual add request input instruction units 112a to 112c, the review instruction prompt is sent only to the corresponding language model 3, and if a review request is entered in the batch add request input field 113, the review instruction prompt is sent to all specified language model 3. After this process, the process proceeds to step S7.

[0059] If no review request is received from the user in step S5 (step S5; No), the process proceeds to step S7 without going through step S6.

[0060] In step S7, if the user requests the exclusion of an answer result through the answer result selection units 114a to 114c in Figure 6 (step S7; Yes), the integration target setting unit 34 executes the process in step S8, which removes the answer result selected by the user from the integration target. In this process in step S8, the answer result presentation units 111a to 111c are also removed from the screen. After this process, the process proceeds to step S9.

[0061] If the user does not request the exclusion of the response in step S7 (step S7; No), the process proceeds to step S9 without going through step S8.

[0062] If the user performs the selection operation of the wall-hitting instruction unit 115 in Figure 6 in step S9, the process proceeds to step S10 (step S9; Yes). If the user does not perform the selection operation of the wall-hitting instruction unit 115 in step S9, the process returns to step S5 (step S9; No).

[0063] In step S10, the integration processing unit 35 generates an integration instruction prompt that instructs the specified multiple language models 3 to integrate the response results. In this embodiment, the integration instruction prompt is displayed in the batch add request input field 113, allowing the user to confirm the contents of the integration instruction prompt on the user terminal 2.

[0064] In step S11, the response result acquisition unit 33 sends an integrated instruction prompt to the specified number of language models 3. The sending of the integrated instruction prompt may be triggered by the user's operation of the send button 125 as described above, or it may be automatically executed when the user operates the wall-hitting instruction unit 115, triggered by the generation of the integrated instruction prompt.

[0065] In step S12, the response result acquisition unit 33 acquires the integrated response result from each language model 3 and transmits the integrated response result to the user terminal 2 via the input / output processing unit 31 for presentation to the user. After the processing in step S12, the process ends. In this embodiment, the integrated response result from each language model 3 is displayed in each of the response result presentation units 111a to 111c in Figure 6.

[0066] As described above, the information processing device 1 of this embodiment includes a response result acquisition unit 33 that sends a request prompt containing the user's request to a plurality of language models 3 (for example, language models 3-1 to 3-3) and acquires a response result for the request from each of the plurality of language models 3, and an integration processing unit 35 that generates an integration instruction prompt that instructs the language model 3 to integrate the response results of each of the plurality of language models 3, and the response result acquisition unit 33 acquires an integrated response result that reflects the response results of the plurality of language models 3 by sending the integration instruction prompt to the language model 3.

[0067] The flowchart in Figure 8 shows the basic processing flow from the user's initial request to obtaining the integrated response result. Even after the integrated response result is presented in step S12, the user can continue to interact with the language model 3. For example, the user can send additional questions or requests for consideration individually to each language model 3 through the individual additional request input instruction units 112a to 112c, or send additional requests for consideration to all language models 3 at once through the batch additional request input field 113. Furthermore, by operating the brainstorming instruction unit 115 again, the user can perform the integration process again based on the new response result or revised response result. In this way, the information processing device 1 of this embodiment is configured to enable continuous dialogue even after the integrated response result is presented, allowing the user to repeatedly interact with the language model 3 until they obtain a satisfactory result. This continuous dialogue function allows the user to confirm the integrated response result, ask further in-depth questions or questions from new perspectives, and obtain a more accurate answer.

[0068] Furthermore, the control method for the information processing device 1 of this embodiment includes: a response result acquisition step of sending a request prompt containing the user's request to a plurality of language models 3 and acquiring a response result for the request from each of the plurality of language models 3; an integration processing step of generating an integration instruction prompt that instructs the language models 3 to integrate the response results of each of the plurality of language models 3; and an integrated response result acquisition step of acquiring an integrated response result that reflects the response results of the plurality of language models 3 by sending an integration instruction prompt to the language models 3.

[0069] Furthermore, the program of this embodiment causes the computer to execute the following steps: a request prompt containing the user's request to multiple language models 3 and an answer result acquisition step which obtains an answer result for the request from each of the multiple language models 3; an integration processing step which generates an integration instruction prompt which instructs the language models 3 to integrate the answer results of each of the multiple language models 3; and an integrated answer result acquisition step which obtains an integrated answer result that reflects the answer results of the multiple language models 3 by sending the integration instruction prompt to the language models 3.

[0070] In this way, with the information processing device 1, its control method, or program configured, the user can simultaneously ask questions to multiple language models 3 and compare their responses. By comparing the outputs of different language models 3, it is possible to obtain responses that leverage the strengths of each language model 3, and furthermore, to obtain integrated responses, which are more comprehensive responses that combine these responses. For example, even with the same prompt, GPT-4o might suggest mobile development or themed sessions, while Claude might suggest a speed challenge rally. By integrating such differences, a broader perspective can be obtained, and the user can acquire ideas from different perspectives during brainstorming. Furthermore, since integrated instruction prompts are automatically generated, the user can easily perform the necessary tasks to utilize the characteristics of multiple language models 3.

[0071] Furthermore, the information processing device 1 of this embodiment further includes a language model designation unit 32 that sets a language model 3 specified by the user from among a plurality of candidate language models 3 as the destination for sending request prompts, and the response result acquisition unit 33 sends a request prompt to the language model 3 set by the language model designation unit 32.

[0072] This allows users to individually select language model 3, enabling them to obtain more appropriate and integrated response results by considering the content they want to request from language model 3 and the characteristics and features of each language model 3. For example, some language model 3s have a web search function, and if the user wants to include web search results in their response, they can select such a language model 3 to generate response results and integrated response results that are more tailored to the user's purpose.

[0073] Furthermore, the information processing device 1 of this embodiment further includes an integrated target setting unit 34 that excludes answer results excluded by the user from the target of the integrated instruction prompt, or that targets only the answer results specified by the user for the integrated instruction prompt.

[0074] This allows users to review the results, remove unnecessary ones, and integrate only useful responses, thereby enabling the generation of more accurate integrated response results.

[0075] Furthermore, in this embodiment, when the response result acquisition unit 33 receives a request from the user for further consideration of the response result, it sends a review instruction prompt including the review request to one or more language models 3 to acquire a revised response result after further consideration, and the integration processing unit 35 generates an integration instruction prompt that instructs the integration of the response result, the revised response result, or both thereof.

[0076] This allows the language model 3 to perform further analysis after reviewing the response results to improve their accuracy, and then generate integrated response results from the language model 3. Furthermore, the accuracy of each response can be improved by individually digging deeper into the response results of each language model 3. Additionally, it is possible to perform a simultaneous re-evaluation on all multiple language models 3, saving the effort required to perform the same re-evaluation on each language model 3, thus improving user convenience.

[0077] <Second Embodiment> Next, the configuration of the second embodiment, which differs in configuration from the first embodiment, will be described. Figure 9 is a functional block diagram showing an example of the functional configuration of the information processing device 1a according to the second embodiment. In the description of the second embodiment, components that are common or similar to those in the first embodiment may be denoted by the same reference numerals, and detailed descriptions may be omitted.

[0078] The information processing device 1a of the second embodiment is configured with the same hardware as the first embodiment (see Figure 2). The information processing device 1a of the second embodiment includes an input / output processing unit 31, a language model specification unit 32, an answer result acquisition unit 33, and an integration processing unit 35 as functional units implemented on the processor (CPU 11). The information processing device 1a of this second embodiment differs from the information processing device 1 of the first embodiment in that it does not include an integration target setting unit 34.

[0079] In the second embodiment, the process from obtaining the response results of multiple language models 3 to integrating them is performed on the backend of the information processing device 1a, which is different from the first embodiment. The response presentation process in the second embodiment will be described with reference to Figure 10. Figure 10 is a schematic diagram showing the data flow in the response presentation process by the information processing device 1a in the second embodiment.

[0080] In step S21, the user displays a request input screen on the user terminal 2. Figure 11 shows an example of the request input screen displayed on the user terminal 2 in the second embodiment. The request input screen in the second embodiment is basically the same as that in the first embodiment. In addition to the language model display unit 101, the request content input field 102, the language model specification operation unit 103, and the user information display unit 104, the request input screen in the second embodiment also displays a setting instruction unit 201 for the language model 3, which is labeled "Super AI".

[0081] In step S22, when the user operates the setting instruction unit 201, the language model specification unit 32 sets one of the pre-configured language models 3 as the destination for sending the request prompt. The language model 3 to be used may be specified by the user through the language model specification operation unit 103 or a language model selection screen as shown in Figure 5.

[0082] In step S23, the input / output processing unit 31 obtains the request details from the user. The request details are entered by the user through the request details input field 102, similar to the first embodiment.

[0083] In step S24, the response result acquisition unit 33 sends a request prompt containing the request details to multiple language models 3 that have been configured. In this example, the response result acquisition unit 33 sends the request prompt to three language models 3-1 to 3-3.

[0084] In step S25, the response result acquisition unit 33 acquires response results from each of the language models 3-1 to 3-3. Since these response results are not transmitted to the user terminal 2, unlike in the first embodiment, the user is not shown the response results before integration.

[0085] In step S26, the integration processing unit 35 generates an integration instruction prompt, and the response result acquisition unit 33 transmits the integration instruction prompt to the language model 3 that is set. In the second embodiment, this language model 3 is set in advance as the language model 3 to be integrated. Note that the language model 3 to be integrated may be selected by the user.

[0086] In step S27, the response result acquisition unit 33 acquires the integrated response result from the language model 3. In step S28, the input / output processing unit 31 sends information for displaying the integrated response result to the user terminal 2. In step S29, the integrated response result is displayed on the user terminal 2.

[0087] In the second embodiment, as in the first embodiment, the arrangement of language models 3-1 to 3-n can be flexibly configured. That is, some or all of language models 3-1 to 3-n may be implemented within the information processing device 1a, and the language model 3 that performs integration is not limited to an external service but may also be placed within the system.

[0088] In the second embodiment, the processing from step S24 to step S28 is performed automatically in the backend, so to the user it appears as if a single AI (language model 3) is providing the answer.

[0089] Although one embodiment of the present invention has been described above, the present invention is not limited to the embodiments described above, and any modifications, improvements, etc. that can achieve the objectives of the present invention are included in the present invention.

[0090] An information display function that visually displays the characteristics and strengths of language model 3 may be added to the configuration of the first or second embodiment. For example, an icon indicating the characteristics may be displayed in the name of language model 3 on the request input screen or language model selection screen, or the characteristics of language model 3 may be displayed using a hover function when the mouse cursor is placed over it.

[0091] Additionally, the response results and integrated response results may be configured to highlight the conclusions and specific keywords. These specific keywords can be extracted from words in the request or pre-configured.

[0092] Furthermore, automatic coordination (sequential execution) between the three language models may be implemented. For example, in an integrated instruction prompt, the response may be passed in a relay format from language model 3-1, language model 3-2, and language model 3-3 in that order to generate an integrated response result. Alternatively, in an integrated instruction prompt, specific roles may be assigned to each of the three language models, and the integrated response result may be generated in a discussion format.

[0093] Furthermore, the series of processes described above can be executed by hardware or by software. In other words, the functional configuration described above is merely illustrative and not particularly limiting. That is, it is sufficient that the information processing device 1 is equipped with the functionality to execute the series of processes described above as a whole, and the type of functional block used to realize this functionality is not particularly limited to the example above. Also, the location of the functional block is not particularly limited and can be arbitrary. For example, the functional block of the information processing device 1 may be transferred to another device, etc. Conversely, the functional block of another device may be transferred to a server, etc. Also, a single functional block may be composed of hardware alone, software alone, or a combination of both.

[0094] When a series of processes are executed by software, the programs that make up that software are installed on a computer or other device from a network or storage medium. The computer may be a computer built into dedicated hardware. Alternatively, the computer may be a computer capable of performing various functions by installing various programs, such as a server, a general-purpose smartphone, or a personal computer.

[0095] Such recording media containing programs may consist not only of removable media (not shown) distributed separately from the main device to provide the programs, but also of recording media provided pre-installed in the main device. Since programs can be distributed via a network, the recording media may be installed on or accessible from a computer connected to or capable of connecting to a network.

[0096] In this specification, the step of describing a program to be recorded on a recording medium includes not only processes that are performed chronologically in that order, but also processes that are not necessarily performed chronologically, but are executed in parallel or individually. Furthermore, in this specification, the term "system" refers to an overall system composed of multiple devices, means, etc. [Explanation of symbols]

[0097] 1, 1a Information Processing Device 2 User terminals 3-1~3-n Language Models 31 Input / Output Processing Unit 32 Language Model Specification Section 33 Answer result acquisition part 34 Integration Target Setting Section 35 Integrated Processing Unit 100 Language Model Utilization Systems

Claims

1. A response result acquisition unit that sends a request prompt containing the user's request to multiple language models and acquires a response result for the request from each of the multiple language models, An integration processing unit that generates an integration instruction prompt that instructs the language model to integrate the response results of each of the multiple language models, An integrated target setting unit that excludes the answer results excluded by the user from the target of the integrated instruction prompt, or that targets only the answer results specified by the user for the integrated instruction prompt, Equipped with, The response result acquisition unit acquires an integrated response result that reflects the response results of multiple language models by transmitting the integrated instruction prompt to the language model. Information processing device.

2. A response result acquisition unit that sends a request prompt containing the user's request to a plurality of language models and acquires a response result for the request from each of the plurality of language models, An integration processing unit that generates an integration instruction prompt that instructs the language model to integrate the response results of each of the multiple language models, Equipped with, The aforementioned unit for obtaining response results, When the user requests further consideration of the aforementioned response result, one or more language models are sent a review instruction prompt including the review request to obtain a revised response result after further consideration. By sending the integrated instruction prompt to the language model, an integrated response result is obtained that reflects the response results of multiple language models. The integration processing unit generates the integration instruction prompt that instructs the integration of the answer result, the revised answer result, or both thereof. Information processing device.

3. The language model specification unit further includes setting the language model specified by the user from among a plurality of candidate language models as the destination for sending the request prompt, The aforementioned unit for obtaining response results, The request prompt is sent to the language model set by the language model specification unit. The information processing apparatus according to claim 1 or 2.

4. A response result acquisition step involves sending a request prompt containing the user's request details to multiple language models and obtaining response results for the request details from each of the multiple language models. An integration processing step that generates an integration instruction prompt that instructs the language model to integrate the response results of each of the multiple language models, A set integration target step which excludes the answer results excluded by the user from the target of the integrated instruction prompt, or which sets only the answer results specified by the user as the target of the integrated instruction prompt, The integrated response result acquisition step involves sending the integrated instruction prompt to the language model to acquire an integrated response result that reflects the response results of multiple language models, A control method for an information processing device, including the device itself.

5. A response result acquisition step involves sending a request prompt containing the user's request details to multiple language models and obtaining response results for the request details from each of the multiple language models. An integration processing step that generates an integration instruction prompt that instructs the language model to integrate the response results of each of the multiple language models, A set integration target step which excludes the answer results excluded by the user from the target of the integrated instruction prompt, or which sets only the answer results specified by the user as the target of the integrated instruction prompt, The integrated response result acquisition step involves sending the integrated instruction prompt to the language model to acquire an integrated response result that reflects the response results of multiple language models, A program that causes a computer to perform a process that includes [a specific action].

6. A response result acquisition step of sending a request prompt containing the user's request to a plurality of language models and obtaining a response result for the request from each of the plurality of language models, An integration processing step that generates an integration instruction prompt that instructs the language model to integrate the response results of each of the multiple language models, The integrated response result acquisition step involves sending the integrated instruction prompt to the language model to acquire an integrated response result that reflects the response results of multiple language models, Includes, In the step of obtaining the response result, if the user requests further consideration of the response result, one or more language models are sent a review instruction prompt including the review request to obtain a revised response result after further consideration. In the integration processing step, an integration instruction prompt is generated that instructs the integration of the answer result, the revised answer result, or both thereof. A method for controlling an information processing device.

7. A response result acquisition step of sending a request prompt containing the user's request to a plurality of language models and obtaining a response result for the request from each of the plurality of language models, An integration processing step that generates an integration instruction prompt that instructs the language model to integrate the response results of each of the multiple language models, The integrated response result acquisition step involves sending the integrated instruction prompt to the language model to acquire an integrated response result that reflects the response results of multiple language models, Includes, In the step of obtaining the response result, if the user requests further consideration of the response result, one or more language models are sent a review instruction prompt including the review request to obtain a revised response result after further consideration. In the integration processing step, an integration instruction prompt is generated that instructs the integration of the answer result, the revised answer result, or both thereof. A program that causes a computer to perform a process.

Images