Information provision device, information collection system, information processing system, method for generating information, and program

Abstract

To allow proposition of more appropriate response information to a user in response to request information from the user.SOLUTION: A terminal device includes: a reception unit 111 for receiving input information input by a user; and an acquisition unit 113 for acquiring response information generated for request information according to the input information, the response information being generated in response to at least one piece of first output information generated by a trained first generation model for a first prompt including the request information and at least one piece of second output information generated by a trained second generation model for a second prompt including the request information.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to a technique for presenting more appropriate response information to request information from a user.

Background Art

[0002] In recent years, a technique has been proposed for outputting information corresponding to information input by a user using a large language model capable of natural language processing. When a large language model is fine-tuned using a dialogue dataset, dialogue-based communication becomes possible. A user can obtain a response sentence to a question sentence by inputting the question sentence into such a large language model. Specifically, a prompt including the question sentence from the user is input into the large language model.

[0003] The response sentence output by the large language model greatly depends on the content of the question sentence input by the user. For example, if the vocabulary selection or the way of speaking in the question sentence is different, the response from the large language model may also be different. Therefore, depending on the content of the question sentence, there may be a case where a response sentence that is not the response desired by the user is output. As described above, there is a problem that the output from the large language model may not be appropriate.

[0004] Therefore, various techniques for making the output from the large language model appropriate have been proposed. For example, Patent Document 1 discloses a configuration in which reference information related to a question sentence input by a user is generated, and a prompt including the question sentence and the reference information is input into the large language model to obtain a response sentence to the question sentence. In Patent Document 1, it is considered that a more specific response sentence can be generated as compared with a configuration for generating a prompt including only the question sentence.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, from the perspective of presenting an answer sentence with appropriate content to the user for the question sentence, there is still room for further improvement in the technology of Patent Document 1. In view of the above circumstances, an object of the present invention is to present more appropriate response information to the user for the request information from the user.

Means for Solving the Problems

[0007] A terminal device according to an example of the present invention includes a reception unit that receives input information input by a user, and an acquisition unit that acquires response information generated for request information corresponding to the input information, where the response information is information generated according to one or more first output information generated by a trained first generation model for a first prompt including the request information, and one or more second output information generated by a trained second generation model for a second prompt including the request information.

[0008] An information providing device according to an example of the present invention includes a response generation unit that generates response information for the request information according to one or more first output information generated by a trained first generation model for a first prompt including request information corresponding to input information input by a user, and one or more second output information generated by a trained second generation model for a second prompt including the request information.

[0009] An information collection system according to an example of the present invention is a system capable of communicating with the aforementioned terminal device, and includes a first acquisition unit that acquires an evaluation of the validity specified for the response information, the request information, and the response information, and a collection unit that stores the evaluation, the request information, and the response information in a storage device.

[0010] An information collection system according to an example of the present invention is a system capable of communicating with the aforementioned terminal device, and includes a first acquisition unit that acquires related information related to the input information generated by a trained fourth generation model for a fourth prompt including the input information, a second acquisition unit that acquires response information generated by a trained fifth generation model for a fifth prompt including the related information, and a collection unit that stores the related information and the response information in a storage device.

[0011] An information processing system according to the present invention includes a reception unit that receives input information input by a user, and a response generation unit that generates response information for the request information according to one or more first output information generated by a trained first generation model for a first prompt including the request information corresponding to the input information and one or more second output information generated by a trained second generation model for a second prompt including the request information, and an acquisition unit that acquires the response information.

[0012] An information acquisition method according to an example of the present invention is realized by a computer system that receives input information input by a user, acquires response information generated for request information corresponding to the input information, and the response information is information generated according to one or more first output information generated by a trained first generation model for a first prompt including the request information and one or more second output information generated by a trained second generation model for a second prompt including the request information.

[0013] A program according to an example of the present invention functions a computer system as a reception unit that receives input information input by a user and an acquisition unit that acquires response information generated for request information corresponding to the input information, and the response information is information generated according to one or more first output information generated by a trained first generation model for a first prompt including the request information and one or more second output information generated by a trained second generation model for a second prompt including the request information.

Brief Description of the Drawings

[0014]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Modes for Carrying Out the Invention

[0015] <First Embodiment> FIG. 1 is a block diagram exemplifying the configuration of an information processing system 100 according to the first embodiment. The information processing system 100 according to the first embodiment is a computer system including a terminal device 10 and an information management system 20. The information management system 20 provides various types of information to the user U of the terminal device 10. The information management system 20 includes an information generation system 30 and an information providing system 40.

[0016] Generally speaking, the information management system 20 provides the user U with information (hereinafter referred to as "response information") R for the request information from the user U. Typically, the request information is information for requesting the information desired by the user U, and the response information R is information indicating a response to the request information (typically the information requested by the request information). In the following description, a case where the request information is an inquiry about a predetermined target and the response information R is an answer to the inquiry will be exemplified. The predetermined target is, for example, various products and services. In the first embodiment, a case where a specific product (hereinafter referred to as "target product") is used as the predetermined target will be exemplified for convenience. The target product is, for example, software used for a predetermined purpose.

[0017] The terminal device 10 and the information management system 20 can communicate with each other via a communication network 900 such as a mobile communication network or the Internet, for example.

[0018] The terminal device 10 is realized by, for example, a portable information device such as a smartphone or a tablet terminal, or a portable or stationary information device such as a personal computer. On the other hand, the information generation system 30 and the information providing system 40 are realized by, for example, a server system.

[0019] [Terminal device 10] FIG. 2 is a block diagram illustrating the configuration of the terminal device 10. As illustrated in FIG. 2, the terminal device 10 includes a control device 11, a storage device 12, a communication device 13, a display device 14, and an operation device 15.

[0020] The control device 11 is one or more processors that control the operation of the terminal device 10. Specifically, for example, the control device 11 is constituted by one or more types of processors such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an SPU (Sound Processing Unit), a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), or an ASIC (Application Specific Integrated Circuit).

[0021] The storage device 12 is one or more memories that store the programs executed by the control device 11 and various data used by the control device 11. For example, known recording media such as semiconductor recording media and magnetic recording media, or a combination of multiple types of recording media are used as the storage device 12. Note that, for example, a portable recording medium that is detachable from the terminal device 10, or a recording medium (e.g., cloud storage) that can be accessed by the control device 11 via the communication network 900 may be used as the storage device 12.

[0022] The communication device 13 is a communication device that communicates with the information management system 20 via the communication network 900. Note that a communication device 13 that is separate from the terminal device 10 may be connected to the terminal device 10 by wire or wirelessly.

[0023] The display device 14 displays various information under the control of the control device 11. For example, various display panels such as a liquid crystal display panel or an organic EL panel are used as the display device 14. Note that an operating device 15 or a display device 14 that is separate from the information processing system 100 may be connected to the terminal device 10 by wire or wirelessly.

[0024] The operation device 15 is an input device that receives operations by the user U. For example, an operator operated by the user U or a touch panel that detects contact by the user U is used as the operation device 15. Note that an operation device 15 separate from the terminal device 10 may be connected to the terminal device 10 by wire or wirelessly. The operation device 15 receives various pieces of information input by the user U through operations on the operation device 15. Note that the operation device 15 may be a sound collection device (microphone) that receives input from the user U by voice.

[0025] FIG. 2 is a block diagram illustrating functions realized by the control device 11 executing a program stored in the storage device 12. The control device 11 realizes a plurality of functions (reception unit 111, prompt generation unit 112, acquisition unit 113, presentation unit 114) for providing answers to inquiries.

[0026] The reception unit 111 receives various pieces of information input by the user U through operations on the operation device 15. First, the reception unit 111 receives input information input by the user U. The input information is information for requesting information desired by the user U. The input information in the first embodiment is information indicating an inquiry regarding the target product. The content of the inquiry (information desired by the user U) is arbitrary, and for example, assumed to be the usage method of the target product, a specific function of the target product, or content regarding an error that occurred while using the target product. The user U inputs a sentence (text) indicating the inquiry through an operation on the operation device 15.

[0027] Second, the reception unit 111 receives request information input by the user U. The request information in the first embodiment is information selected by the user U from among the input information and one or more pieces of related information G. That is, the input information and one or more pieces of related information G are candidates for the request information.

[0028] The related information G is, like the input information, information for the user U to request the information desired, and is information related to the input information. The related information G is typically information indicating the same type of content (e.g., an inquiry) as the input information. It can also be said that the content of the input information and the content of the related information G are similar. The requested information, which is the information selected by the user U from the input information and one or more pieces of related information G related to the input information, is information corresponding to the input information. Details of the related information G will be described later.

[0029] Thirdly, the reception unit 111 receives the identification information input by the user U. The identification information is, for example, various information such as the name, type, version, or product number of the target product. Note that the identification information may include information for identifying the user U of the target product.

[0030] The prompt generation unit 112 generates a prompt P including various information input by the user U. The prompt P is a request to the information management system 20. The prompt P generated by the prompt generation unit 112 is transmitted from the communication device 13 to the information management system 20.

[0031] First, the prompt generation unit 112 generates a prompt P1 (an exemplification of the "third prompt") including the input information. Specifically, the prompt P1 is a prompt for requesting one or more pieces of related information G related to the input information. The prompt P1 is transmitted to the information management system 20 (information generation system 30). Note that the prompt P1 including the input information includes both the case where the input information is directly included and the case where the input information is indirectly included as information corresponding to the input information (e.g., information obtained by processing or modifying the input information).

[0032] Second, the prompt generation unit 112 generates prompts Pz2 and P3 including the requested information. Note that the prompts Pz2 and P3 including the requested information include both the case where the requested information is directly included and the case where the requested information is indirectly included as information corresponding to the requested information (e.g., information obtained by processing or modifying the requested information).

[0033] Specifically, each of prompt Pz2 and prompt P3 is a prompt for requesting output information W for the request information. The output information W is information that serves as a candidate for the response information R (i.e., the answer). That is, it can also be said that prompts Pz2 and P3 are prompts for requesting the response information R.

[0034] Prompts Pz2 and P3 are transmitted to the information management system 20 (information providing system 40). One or more pieces of output information (hereinafter referred to as "first output information") W1 are generated for prompt Pz2, and one or more pieces of output information (hereinafter referred to as "second output information") W2 are generated for prompt P3. Using the plurality of pieces of output information W (W1, W2) generated by the information management system 20, response information R for the request information is generated. Details of the output information W and the response information R will be described later.

[0035] In the first embodiment, actually, one or more pieces of first output information W1 are generated for prompt P2 generated from prompt Pz2. The generation of prompt P2 is performed in the information providing system 40. In the first embodiment, a configuration is exemplified in which the information included in prompt P2 is different from the information included in prompt P3. Commonly included in prompt P2 and prompt P3 are the request information and the identification information.

[0036] On the other hand, prompt P2 includes reference information that is referred to in the generation of the first output information W1 in addition to the request information and the identification information. In the information providing system 40 described later, prompt P2 is generated by adding reference information to prompt Pz2 transmitted from the terminal device 10. The reference information is information that the information processing system 100 refers to when generating the first output information W1 (candidate for the response information). Specifically, the reference information is, for example, information that details the target product. For example, a document describing the use, purpose, function, configuration, technical specifications, and usage / setting method of the target product is exemplified as the reference information. Note that in the first embodiment, prompt P2 is an example of the "first prompt", and prompt P3 is an example of the "second prompt".

[0037] The acquisition unit 113 acquires the information generated by the various prompts P1 - P3 described above. The acquisition unit 113 in the first embodiment acquires the information generated by the prompts P1 - P3 from the information management system 20 by receiving it with the communication device 13.

[0038] First, the acquisition unit 113 acquires one or more pieces of related information G generated for the prompt P1 including the input information. Second, the acquisition unit 113 acquires the response information R generated for the requested information which is the information selected by the user U from among the input information and the one or more pieces of related information G.

[0039] The presentation unit 114 presents various information (for example, related information G and response information R) to the user U. Specifically, the presentation unit 114 presents various information to the user U by displaying it with the display device 14. However, the presentation of information by the presentation unit 114 is not limited to presentation by display, and may be, for example, presentation by voice.

[0040] First, the presentation unit 114 presents one or more pieces of related information G generated by the prompt P1 to the user U. Specifically, one or more pieces of related information G are presented to the user U together with the input information.

[0041] [Information generation system 30] FIG. 4 is a block diagram illustrating the configuration of the information generation system 30. The information generation system 30 is a device for generating one or more pieces of related information G related to the input information. For example, a server system capable of communicating with the terminal device 10 via the communication network 900 is exemplified as the information generation system 30.

[0042] As illustrated in FIG. 4, the information generation system 30 includes, for example, a control device 31, a storage device 32, and a communication device 33.

[0043] The control device 31 is one or more processors that control the operation of the information generation system 30. Specifically, for example, the control device 31 is composed of one or more types of processors such as a CPU (Central Processing Unit), GPU (Graphics Processing Unit), SPU (Sound Processing Unit), DSP (Digital Signal Processor), FPGA (Field Programmable Gate Array), or ASIC (Application Specific Integrated Circuit).

[0044] The storage device 32 is one or more memories that store the programs executed by the control device 31 and various data used by the control device 31. For example, known recording media such as semiconductor recording media and magnetic recording media, or a combination of multiple types of recording media, are used as the storage device 32. For example, a portable recording medium that can be attached to and detached from the information generation system 30, or a recording medium (such as cloud storage) that can be accessed by the control device 31 via the communication network 900, may be used as the storage device 32.

[0045] The communication device 33 is a communication device that communicates with the terminal device 10 via the communication network 900. Note that a communication device 33 that is separate from the information generation system 30 may be connected to the information generation system 30 by wire or wirelessly.

[0046] FIG. 5 is a block diagram illustrating functions realized by the control device 31 executing a program stored in the storage device 32. The control device 31 according to the first embodiment realizes a plurality of functions (acquisition unit 311, related information generation unit 312) for generating one or more pieces of related information G related to the input information.

[0047] The acquisition unit 311 acquires the input information transmitted from the terminal device 10 by receiving it with the communication device 33. Specifically, the acquisition unit 311 acquires the prompt P1 including the input information.

[0048] The related information generation unit 312 generates one or more pieces of related information G related to the input information. In the first embodiment, the trained generation model M1 is used to generate the related information G. The related information generation unit 312 generates one or more pieces of related information G by inputting a prompt P1 including the input information to the generation model M1. The prompt P1 can also be paraphrased as a prompt that instructs the generation model M1 to generate one or more pieces of related information G related to the input information.

[0049] The one or more pieces of related information G generated by the related information generation unit 312 are transmitted to the terminal device 10 by the communication device 33 and presented to the user U. As described above, the one or more pieces of related information G are presented to the user U together with the input information.

[0050] The generation model M1 is composed of, for example, a Transformer model (sequence conversion model). However, the generation model M1 may be composed of any other form of neural network (for example, a recurrent neural network or a convolutional neural network). Further, the generation model M1 may be composed of a combination of multiple types of deep neural networks.

[0051] In the first embodiment, a configuration is exemplified in which a large language model (LLM) that has been pre-trained using a large amount of text data and in which the appearance probability of words is modeled is used as a generation model M1. The large language model is capable of natural language processing. In particular, it is preferable to use, as the generation model M, a dialogue-type large language model obtained by fine-tuning the large language model using a dialogue dataset. According to the dialogue-type large language model, it is possible to generate text in a dialogue format. As the dialogue-type large language model, for example, any known model such as ChatGPT (manufactured by OpenAI), Bard (manufactured by Google), Bing AI (manufactured by Microsoft) can be used as the generation model M1. The generation model M1 is preferably fine-tuned by a dataset including a plurality of combinations of assumed input information (prompt P1) and related information G (answer) for the input information. Note that, as the generation model M1, a learned model obtained by learning the relationship between the input information and the related information G by machine learning may be used.

[0052] As understood from the above description, the generation model M1 outputs statistically reasonable related information G for the input information. Note that a plurality of parameters defining the generation model M1 are set by machine learning (particularly deep learning) using a plurality of training data and are stored in the storage device 32. Note that it is possible for the generation model M1 to refer to the database D2 described later in the third embodiment and to use the information in the database D2 as training data for the generation model M1.

[0053] [Information providing system 40] FIG. 6 is a block diagram illustrating the configuration of the information providing system 40. The information providing system 40 is a device that generates response information R for request information corresponding to input information. That is, an answer to an inquiry is generated by the information providing system 40. For example, a computer system capable of communicating with the terminal device 10 via the communication network 900 is exemplified as the information providing system 40.

[0054] As illustrated in FIG. 6, the information providing system 40 includes a first processing device 41, a second processing device 42, an information providing device 43, and an information acquisition device 44. The information providing device 43 can communicate with each of the first processing device 41 and the second processing device 42 via a communication network 900.

[0055] The information acquisition device 44 is a server system for generating a prompt P2 from a prompt Pz2 transmitted from the terminal device 10. The information acquisition device 44 according to the first embodiment generates the prompt P2 by adding reference information to the prompt Pz2. That is, the prompt P2 is a prompt including request information, identification information, and reference information. The information acquisition device 44 can communicate with the terminal device 10 and the first processing device 41 via the communication network 900. For example, the information acquisition device 44 is composed of a control device 441, a storage device 442, and a communication device 443.

[0056] The communication device 443 receives the prompt Pz2 transmitted from the terminal device 10 and transmits the prompt P2 generated from the prompt Pz2 to the first processing device 41.

[0057] The control device 441 generates the prompt P2 from the prompt Pz2. Specifically, reference information corresponding to the request information and the identification information included in the prompt Pz2 is acquired, and the prompt P2 is generated by adding the reference information to the prompt Pz2. The method of acquiring the reference information is not particularly limited, but in the following description, a configuration in which the control device 441 acquires the reference information using a data table (hereinafter referred to as a "reference information table") Dq pre-registered in the storage device 442 is exemplified.

[0058] In the reference information table Dq, for example, reference information is registered in association with each combination of identification information and a character string (hereinafter referred to as "target character string") that may be included in the request information. The target character string that may be included in the request information is, for example, a character string (including words) representing functions that may be the subject of an inquiry about the target product. As illustrated in FIG. 7, in the reference information table Dq, different reference information is associated with each combination of identification information and the target character string. Note that common reference information may be registered for a plurality of combinations, or a plurality of reference information may be registered for one combination. Then, in the reference information table Dq, reference information in which the character string in the request information included in the prompt Pz2 matches (or has a high similarity) with the identification information is acquired. However, the control device 441 may acquire reference information by inputting the request information and the identification information to the generation model. The generation model is composed of, for example, a Transformer model or a neural network of any form, similar to the generation model M1.

[0059] FIG. 8 is a block diagram illustrating the configuration of the first processing device 41. The first processing device 41 is a server system that generates one or more first output information W1 for the prompt P2. The first output information W1 is, as described above, a candidate for the response information R (answer). Specifically, the first processing device 41 includes a control device 411, a storage device 412, and a communication device 413.

[0060] The control device 411 is one or more processors that control the operation of the first processing device 41. Specifically, for example, the control device 411 is composed of one or more types of processors such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an SPU (Sound Processing Unit), a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), or an ASIC (Application Specific Integrated Circuit).

[0061] The storage device 412 is one or more memories that store the programs executed by the control device 411 and various data used by the control device 411. For example, known recording media such as semiconductor recording media and magnetic recording media, or a combination of multiple types of recording media, are used as the storage device 412. For example, a portable recording medium that is detachable from the information providing system 40, or a recording medium (e.g., cloud storage) that can be accessed by the control device 411 via the communication network 900, may be used as the storage device 412.

[0062] The communication device 413 is a communication device that communicates with the terminal device 10 via the communication network 900. Note that a communication device 413 that is separate from the information providing system 40 may be connected to the first processing device 41 by wire or wirelessly.

[0063] FIG. 9 is a block diagram illustrating functions realized by the control device 411 executing the programs stored in the storage device 412. The control device 411 of the first embodiment realizes a plurality of functions (acquisition unit 141, processing unit 142) for generating one or more pieces of first output information W1 according to the request information.

[0064] The acquisition unit 141 acquires the request information transmitted from the information acquisition device 44 by receiving it with the communication device 413. Specifically, the acquisition unit 141 acquires the prompt P2 transmitted from the information acquisition device 44. As described above, the prompt P2 includes request information, identification information, and reference information.

[0065] The processing unit 142 generates one or more pieces of first output information W1 for the request information. A trained generation model M2 is used to generate the first output information W1. Specifically, the processing unit 142 generates one or more pieces of first output information W1 by inputting the prompt P2 acquired by the acquisition unit 141 into the generation model M2. The generation model M2 of the first embodiment generates the first output information W1 for the request information in consideration of the identification information and the reference information. The prompt P2 can also be paraphrased as a prompt that instructs to generate one or more pieces of first output information W1 for the request information based on the identification information and the reference information. Details of the generation model M2 will be described later. The first output information W1 generated by the processing unit 142 is transmitted to the information providing device 43 by the communication device 413.

[0066] FIG. 10 is a block diagram illustrating the configuration of the second processing device 42. The second processing device 42 is a server system that generates one or more pieces of second output information W2 for the prompt P3. As described above, the second output information W2 is a candidate for the response information R (answer), similar to the first output information W1. Specifically, the second processing device 42 includes a control device 421, a storage device 422, and a communication device 423.

[0067] The control device 421, the storage device 422, and the communication device 423 adopt, for example, the same configurations as those described for the control device 411, the storage device 412, and the communication device 413 of the first processing device 41.

[0068] FIG. 11 is a block diagram illustrating the functions realized by the control device 421 executing the program stored in the storage device 422. The control device 421 of the first embodiment realizes a plurality of functions (acquisition unit 241, processing unit 242) for generating one or more pieces of second output information W2 according to the request information.

[0069] The acquisition unit 241 acquires the request information transmitted from the terminal device 10 by receiving it with the communication device 423. Specifically, the acquisition unit 241 acquires the prompt P3 transmitted from the terminal device 10. As described above, the prompt P3 includes the request information and the identification information. Note that in the first embodiment, the reference information is not included in the prompt P3.

[0070] The processing unit 242 generates one or more pieces of second output information W2 for the request information. A trained generation model M3 is used to generate the second output information W2. Specifically, the processing unit 242 inputs the prompt P3 acquired by the acquisition unit 241 into the generation model M3 to generate one or more pieces of second output information W2. The generation model M3 of the first embodiment generates one or more pieces of second output information W2 for the request information in consideration of the identification information. It can also be said that the prompt P3 is a prompt that instructs to generate one or more pieces of second output information W2 for the request information based on the identification information. The second output information W2 generated by the processing unit 242 is transmitted to the information providing device 43 by the communication device 423.

[0071] Hereinafter, the generation models M2 and M3 will be described. The generation models M2 and M3 are configured by, for example, a Transformer model, similar to what was described above for the generation model M1. In the first embodiment, a configuration is exemplified in which a large language model that is pre-trained in advance using a large amount of text data (training data) and in which the appearance probability of words is modeled is used as the generation models M2 and M3.

[0072] Large language models are capable of natural language processing. In particular, it is preferable to use dialogue-based large language models fine-tuned using a dialogue dataset as generation models M2 and M3. According to the dialogue-based large language model, it is possible to generate text in dialogue form. As the dialogue-based large language model, for example, any known model such as ChatGPT (manufactured by OpenAI), Bard (manufactured by Google), Bing AI (manufactured by Microsoft) can be used as generation models M2 and M3. The generation model M2 and the generation model M3 may be of the same type or different types. Note that various parameters defining the generation models M2 and M3 are set by machine learning (especially deep learning) using a plurality of training data and are held in the storage devices 412 and 422.

[0073] For the establishment of the generation model M2, various data related to the target product are used as training data. In the first embodiment, for example, a dataset including a plurality of combinations of assumed requirement information (prompt P2) and first output information W1 (answer) for the requirement information is used as training data. In particular, highly reliable information related to the target product is preferably used as training data. Note that the generation model M2 may be established by fine-tuning a generation model pre-trained with a large amount of text data using training data related to the target product.

[0074] Highly reliable information related to the target product is, for example, information collected in the information collection system 50 described later (information stored in the databases D1 and D2). Also, for example, information publicly available or collected by the manufacturer or seller of the target product is also highly reliable information related to the target product. Note that it is also possible for the generation model M2 to refer to the information collected in the information collection system 50 in the generation of the first output information W1.

[0075] To establish the generation model M3 of the first embodiment, mainly various types of data that are generally publicly available (for example, publicly available on websites, books, academic papers, and newspapers) are used as training data. The training data used to establish the generation model M3 may include information on the target product that is generally publicly available, but is information covering a wide range of content other than the target product. The generation model M3 of the first embodiment is trained with a wider range of information than the generation model M2, but it can also be said that it is not trained with highly reliable information on the target product as the training data like the generation model M2.

[0076] As understood from the above description, in the first embodiment, the training data of the generation model M2 and the generation model M3 is different. For the generation model M2, information specialized for the target product (that is, information selected as related to a specific target) is used as training data, and for the generation model M3, a wide range of information that is generally publicly available is used as training data. It can also be said that the training data used for training the generation model M2 is information that is more limited (content-wise) than the training data used for training the generation model M3. While the generation model M2 can generate an answer (the first output information W1) specialized for the information of the target product, the generation model M3 can generate an answer (the second output information W2) taking into account a wide range of information other than the target product. However, in establishing the generation model M2, in addition to various types of data related to the target product, a wide range of information that is generally publicly available, similar to the generation model M3, may be used as training data.

[0077] FIG. 12 is a block diagram illustrating the configuration of the information providing apparatus 43. The information providing apparatus 43 is a server system that generates response information R for request information. Specifically, the information providing apparatus 43 includes a control device 431, a storage device 432, and a communication device 413.

[0078] The control device 431, the storage device 432, and the communication device 433 adopt, for example, the same configuration as that described for the control device 411, the storage device 412, and the communication device 413 of the first processing device 41.

[0079] FIG. 13 is a block diagram illustrating functions realized by the control device 431 executing a program stored in the storage device 432. The control device 431 of the first embodiment realizes a plurality of functions (acquisition unit 341, index calculation unit 342, response generation unit 343) for generating response information R for request information.

[0080] The acquisition unit 341 acquires by receiving, via the communication device 433, one or more pieces of first output information W1 transmitted from the first processing device 41 and one or more pieces of second output information W2 transmitted from the second processing device 42.

[0081] The index calculation unit 342 specifies an index K indicating the validity of the request information for each of the one or more pieces of first output information W1 acquired by the acquisition unit 341 and each of the one or more pieces of second output information W2 acquired by the acquisition unit 341. That is, the index K is specified for each output information W (for each first output information W1 and for each second output information W2).

[0082] The index K is an index indicating how appropriate (suitable) the output information W is for the request information, and is, for example, a numerical value (for example, a score). For example, the higher the score, the higher the validity of the output information W, and the lower the score, the lower the validity of the output information W.

[0083] In the first embodiment, the generation model M4 is used for calculating the index K. Specifically, the index calculation unit 342 inputs the request information and the first output information W1 for each of the first output information W1 into the generation model M4, and inputs the request information and the second output information W2 for each of the second output information W2 into the generation model M4, thereby specifying the index K for each output information W.

[0084] The generation model M4 is composed of a Transformer model, similar to that described above for the generation model M1. For example, it may use a large language model or a trained model that has learned the relationship between the output information W, the request information, and the index K through machine learning. When the generation model M4 is a Transformer model, it is preferably trained by a dataset including a plurality of combinations of, for example, assumed output information W and an index K for the output information W. As can be understood from the above description, the generation model M4 outputs a statistically reasonable index K for the output information W. Note that a plurality of parameters defining the generation model M4 are set by machine learning (especially deep learning) using a plurality of training data and are held in the storage device 432.

[0085] The response generation unit 343 generates response information R for the request information. Specifically, the response generation unit 343 generates response information R for the request information according to one or more first output information W1 generated by the generation model M2 for the prompt P2 and one or more second output information W2 generated by the generation model M3 for the prompt P3.

[0086] The response generation unit 343 of the first embodiment generates response information R according to the index K specified for each of the one or more first output information W1 and each of the one or more second output information W2. Specifically, the response generation unit 343 generates response information R according to the result of comparing the index K specified for each of the one or more first output information W1 and each of the one or more second output information W2 with a threshold value. In this embodiment, the method of generating response information R is made different according to the number of output information W for which the index K exceeds the threshold value among the one or more first output information W1 and the one or more second output information W2.

[0087] The method for generating the response information R differs depending on whether there is one output information W for which the index K exceeds the threshold among one or more pieces of first output information W1 and one or more pieces of second output information W2, whether there are multiple pieces of output information W for which the index K exceeds the threshold among one or more pieces of first output information W1 and one or more pieces of second output information W2, and whether there is no output information W for which the index K exceeds the threshold among one or more pieces of first output information W1 and one or more pieces of second output information W2. For the above three cases, the specific method by which the response generation unit 343 generates the response information R will be described later. Note that the threshold is arbitrarily set by the administrator of the information providing system 40. The response information R specified by the response generation unit 343 is transmitted to the terminal device 10 by the communication device 433.

[0088] As understood from the above description, the response information R is information generated according to one or more pieces of first output information W1 generated by the trained generation model M2 for the prompt P2 including the request information and one or more pieces of second output information W2 generated by the trained generation model M3 for the prompt P3 including the request information.

[0089] FIG. 14 is a flowchart illustrating a specific procedure of processing executed by the entire information processing system 100. The user U of the terminal device 10 inputs input information for requesting (inquiring) desired information about the target product by operating the operating device 15. FIG. 15 illustrates an example of the input information input by the user U. FIG. 15 illustrates input information indicating an inquiry about an error that occurred in the function X of the target product.

[0090] The reception unit 111 of the terminal device 10 receives the input information input by the user U (SA1). The prompt generation unit 112 generates a prompt P1 including the input information (SA2) and transmits the prompt P1 from the communication device 13 to the information generation system 30 (SA3). As described above, the prompt P1 is a prompt P for requesting one or more pieces of related information G related to the input information.

[0091] The acquisition unit 311 of the information generation system 30 acquires the prompt P1 transmitted from the terminal device 10 by receiving it through the communication device 33 (SB1). The related information generation unit 312 generates one or more pieces of related information G for the prompt P1 acquired by the acquisition unit 311 (SB2). Statistically appropriate related information G (related information G that is highly likely to be related to the content of the input information) is generated from the content of the input information. The related information generation unit 312 transmits the generated one or more pieces of related information G from the communication device 33 to the terminal device 10 (SB3).

[0092] FIG. 15 also illustrates an example of the related information G generated by the related information generation unit 312. In FIG. 15, a case where three pieces of related information G (G1, G2, G3) are generated is exemplified. As illustrated in FIG. 15, the related information G is information related (similar) to the input information and is information indicating an inquiry regarding the function X similar to the input information. For example, related information G1 that further specifies the content of the input information (for example, an error that occurred in function X or information requested) and related information G2 and G3 regarding content related to the content of the input information (for example, content regarding errors that occurred in functions X1 and X2 related to function X) are assumed.

[0093] The acquisition unit 113 of the terminal device 10 receives one or more pieces of related information G transmitted by the information generation system 30 through the communication device 13 (SA4). The presentation unit 114 presents the input information and one or more pieces of related information G to the user U by displaying them on the display device 14 (SA5). That is, the input information and the three pieces of related information G (G1, G2, G3) exemplified in FIG. 15 are displayed on the display device 14. The user U selects the information considered to be the most appropriate as the requested information from the input information and the three pieces of related information G displayed on the display device 14. The selection of the requested information is performed by an operation on the operation device 15. The reception unit 111 receives the requested information input (selected) by the user U through an operation on the operation device 15 (SA6).

[0094] Note that the user U may use the information selected from the input information and one or more pieces of related information G after modifying it as the request information. Alternatively, the user U may use the information selected by the user U after processing (e.g., conversion or combination with other information) in the terminal device 10 as the request information. That is, the request information may be the input information or the related information G itself, or information corresponding to the input information or the related information G.

[0095] Furthermore, the user U inputs identification information for identifying the inquiry target (target product) by operating the operation device 15. For example, the user U may input the identification information (e.g., the version of the target product) by selecting it from a plurality of prepared options, or the user U may input arbitrary identification information. The plurality of options are displayed by the display device 14. The reception unit 111 receives the identification information input by the user U by operating the operation device 15 (SA7).

[0096] The prompt generation unit 112 generates prompts Pz2, P3 including the request information and the identification information (SA8). The prompt generation unit 112 transmits the generated prompts Pz2, P3 from the communication device 13 to the information providing system 40 (SA9).

[0097] When the information providing system 40 (the first processing device 41, the second processing device 42, the information providing device 43, the information acquisition device 44) receives the prompts Pz2, P3 transmitted from the terminal device 10, it executes a process of generating response information R (SC1). FIG. 16 is a flowchart illustrating the detailed procedure of step SC1.

[0098] As illustrated in FIG. 16, the control device 441 of the information acquisition device 44 acquires the prompt Pz2 transmitted from the terminal device 10 by receiving it with the communication device 443 (SC01). Next, the control device 441 generates the prompt P2 from the prompt Pz2 (SC02). Specifically, the control device 441 generates the prompt P2 including the request information, the identification information, and the reference information by attaching the reference information to the prompt Pz2. As described above, the reference information is acquired using the reference information table Dq. For example, the reference information in which the target character string that matches (or has a high similarity) the character string (hereinafter referred to as "search character string") included in the reference information of the prompt P2 and the identification information that matches (or has a high similarity) the identification information of the prompt P2 are associated is acquired from the reference information table Dq. For example, when the request information is "Switched from product A to product B. Please teach me the procedure for enabling function X.", the character string "Procedure for enabling function X" can be used as the search character string. Note that a plurality of different search character strings may be used to acquire a plurality of pieces of reference information. Then, the control device 441 transmits the prompt P2 to the first processing device 41 (SC03).

[0099] The acquisition unit 141 of the first processing device 41 acquires the prompt P2 transmitted from the information acquisition device 44 by receiving it with the communication device 413 (SC11). The processing unit 142 generates one or more pieces of first output information W1 for the prompt P2 acquired by the acquisition unit 141 (SC12). Specifically, the processing unit 142 generates one or more pieces of first output information W1 by inputting the prompt P2 to the generation model M2. The processing unit 142 transmits the generated one or more pieces of first output information W1 to the information providing device 43 (SC13).

[0100] The acquisition unit 241 of the second processing device 42 acquires the prompt P3 transmitted from the terminal device 10 by receiving it with the communication device 423 (SC21). The processing unit 242 generates one or more second output information W2 for the prompt P3 acquired by the acquisition unit 241 (SC22). Specifically, the processing unit 242 generates one or more second output information W2 by inputting the prompt P3 into the generation model M3. The acquisition unit 241 transmits the generated one or more second output information W2 to the information providing device 43 (SC23). Note that steps SC11 - SC13 and steps SC21 - SC23 may be performed in parallel or sequentially.

[0101] FIG. 17 illustrates the first output information W1 and the second output information W2. In FIG. 17, a case where two first output information W1 (W11, W12) are generated and one second output information W2 is generated is illustrated. The output information W1, W2 are candidates for the response information R (answer).

[0102] The acquisition unit 341 of the information providing device 43 acquires one or more first output information W1 transmitted from the first processing device 41 and one or more second output information W2 transmitted from the second processing device 42 by receiving them with the communication device 433 (SC31, SC32). Note that the order of steps SC31 and SC32 is arbitrary and they may be performed in parallel.

[0103] The index calculation unit 342 specifies an index K indicating the validity with respect to the request information for each of the one or more first output information W1 and each of the one or more second output information W2 (SC33). In the following description, a case where the index K is a score (out of 100 points) is illustrated. The generation model M4 is used to specify the index K. Specifically, the index K is specified by inputting the request information and the output information W into the generation model M4 for each of the plurality of output information W (W11, W12, W2). As illustrated in FIG. 17, the index K (score) is specified for each output information W.

[0104] The response generation unit 343 executes a process of generating response information R (hereinafter referred to as "response generation process") according to one or more pieces of first output information W1 and one or more pieces of second output information W2 acquired by the acquisition unit 341 (SC34). In the response generation process of the present embodiment, response information R is generated according to an index K specified for each of one or more pieces of first output information W1 and each of one or more pieces of second output information W2. FIG. 18 is a flowchart illustrating a specific procedure of the response generation process.

[0105] First, the response generation unit 343 determines whether there is output information W for which the index K exceeds a threshold value among a plurality of pieces of output information W (W1, W2) (SC34-1). When there is output information W for which the index K exceeds the threshold value among the plurality of pieces of output information W (SC34-1; YES), the response generation unit 343 generates response information R (SC34-2). When there is one piece of output information W for which the index K exceeds the threshold value, response information R is generated according to the one piece of output information W. The response information R generated according to the output information W is assumed to be, in addition to the case where it is the output information W itself, for example, information obtained by processing or modifying the output information W, or information obtained by adding other information to the output information W.

[0106] When there are a plurality of pieces of output information W for which the index K exceeds the threshold value, response information R is generated according to the plurality of pieces of output information W. The response information R generated according to the plurality of pieces of output information W is assumed to be, for example, in addition to the case where it is the information itself obtained by combining the plurality of pieces of output information W, for example, information obtained by processing or modifying the combined information, or information obtained by adding other information to the plurality of pieces of output information W.

[0107] On the other hand, when there is no output information W among the plurality of output information W for which the index K exceeds the threshold value (SC34-1; NO), the response generation unit 343 determines whether the process of SC34-1 has been repeated a predetermined number of times (for example, 2 times) (SC34-3). If the process of SC34-1 has not been repeated for a predetermined number of times (SC34-3; NO), the process returns to step SC02 in FIG. 16, and the prompt P2 is generated again. Specifically, the information acquisition device 44 generates a prompt P2 in which the reference information is changed to other reference information. The information acquisition device 44 of the first embodiment uses the reference information table Dq to change the search string and then acquires the reference information again. For example, when the request information is "Switched from product A to product B. Please teach me the activation procedure for function X.", if the search string was the character string "activation procedure for function X" in the previous step SC02, the character string "activation procedure for function X in product B" can be used as the search string. That is, the search string may be a character string formed by combining a plurality of words included in the request information. The changed search string is, for example, a character string formed by combining other words or character strings with the previous search string (that is, a character string partially overlapping before and after the change), or a character string completely different from the previous search string. Note that the changed prompt P2 may include the reference information of the prompt P2 before the change together with the newly acquired reference information. On the other hand, the request information and the identification information are maintained. The changed prompt P2 is transmitted to the first processing device 41 (SC03).

[0108] Next, the first processing device 41 performs steps SC11-SC13 in FIG. 16 again. Then, the information providing device 43 acquires one or more first output information W1 generated for the changed prompt P2 (SC31), and executes steps SC33 and SC34. The process of changing the prompt P2 can also be paraphrased as a process of updating the first output information W1. Note that the step SC32 of receiving the second output information W2 and the process of calculating the index K for the second output information W2 in step SC33 are omitted because there is no change in the prompt P3. That is, the second output information W2 is not updated and is maintained.

[0109] For each of the one or more updated first output information W1, the response generation unit 343 of the information providing device 43 executes step SC34-1 in FIG. 18 for the specified index K and the index K of each of the one or more maintained second output information W2 (that is, the index K specified in the previous step SC33). That is, when there is no output information W for which the index K exceeds the threshold value, the response generation unit 343 generates response information R according to the index K for each of the one or more first output information W1 newly generated by the generation model M2 for the prompt P2 in which the reference information has been changed to other reference information, and each of the one or more maintained second output information W2.

[0110] In the example of FIG. 17, for example, when the threshold value is set to 80 points, response information R is generated according to the first output information W11. When the threshold value is set to 70 points, response information R is generated according to the first output information W11 and the first output information W12. When the threshold value is set to 95 points, the prompt P2 is updated without generating response information R.

[0111] On the other hand, when the process of SC34-1 in FIG. 18 is repeated a predetermined number of times (SC34-3; YES), the response generation unit 343 generates response information R (SC34-2). In this case, a response indicating that appropriate information for the request information could not be generated (that is, a response indicating that the desired information could not be provided to the user U) is generated as the response information R. Note that when there is no output information for which the index K exceeds the threshold value, a configuration is also adopted in which the process proceeds to step SC34-2 without performing the process of step SC34-3 to generate response information R (a response indicating that appropriate information for the request information could not be generated).

[0112] As understood from the above description, the response generation unit 343 generates response information R according to the result of comparing the index K specified for each of one or more pieces of first output information W1 and one or more pieces of second output information W2 with a threshold value. Then, the method of generating the response information R is made different according to the number of pieces of output information W for which the index K exceeds the threshold value. Specifically, the method of generating the response information R is different depending on whether there is one piece of output information W for which the index K exceeds the threshold value among one or more pieces of first output information W1 and one or more pieces of second output information W2, whether there are a plurality of pieces of output information W exceeding the threshold value, and whether there is no piece of output information W exceeding the threshold value.

[0113] As illustrated in FIG. 14, the response information R generated by the response generation unit 343 is transmitted from the communication device 433 to the terminal device 10 (SC2). The acquisition unit 113 of the terminal device 10 acquires it by receiving the response information R by the communication device 13 (SA10). Then, the presentation unit 114 presents it to the user U by displaying the response information R acquired by the acquisition unit 113 on the display device 14 (SA11).

[0114] In the first embodiment, since the response information R generated according to one or more pieces of first output information W1 generated by the generation model M2 for the prompt P2 and one or more pieces of second output information W2 generated by the generation model M3 for the prompt P3 can be acquired, for example, compared with a configuration in which the response information R generated according to the output information W generated by one generation model is acquired and provided to the user U, it becomes possible to present the user U with more appropriate response information R for the request information from the user U.

[0115] <Second Embodiment> The second embodiment will be described. For elements whose functions are the same as those in the first embodiment in each of the embodiments illustrated below, the reference numerals used in the description of the first embodiment are reused, and the detailed description of each is appropriately omitted.

[0116] FIG. 19 is a block diagram illustrating the configuration of the information processing system 100 according to the second embodiment. The information processing system 100 according to the second embodiment has a configuration in which an information collection system 50 is added to the information management system 20 of the first embodiment. Other configurations are the same as those of the first embodiment. The information management system 20 is a computer system for collecting various types of information after response information R is presented to the user U.

[0117] In the second embodiment, after response information R to the request information is presented, the validity of the response information R is specified. In the second embodiment, for example, the user U specifies (inputs) an evaluation of the validity (hereinafter referred to as "feedback evaluation") of the response information R generated in response to the request information by operating the operating device 15. The feedback evaluation includes, for example, an index (hereinafter referred to as "evaluation index") indicating how appropriate the response information R is for the request information, and a text (hereinafter referred to as "evaluation text") that the user U can arbitrarily input regarding the validity of the response information R. The evaluation index may be, for example, an index selected from options (such as "good", "normal", "bad") that divide the evaluation of validity into multiple levels, or a score that the user U can arbitrarily input. However, the feedback evaluation may be performed by a processing device (not shown) instead of the user U. The processing device outputs, for example, a feedback evaluation for the response information R using a generation model. For such a generation model, for example, a model in which the relationship between the response information R and the feedback evaluation is machine-learned or a large language model is used.

[0118] The reception unit 111 of the terminal device 10 receives the specified feedback evaluation. Then, the terminal device 10 transmits information (hereinafter referred to as "evaluation information") F including the feedback evaluation to the information collection system 50. The evaluation information F includes the response information R that is the target of the evaluation and the request information corresponding to the response information R. Further, the evaluation information F of the second embodiment includes identification information. Note that the evaluation information F may include other information (such as input information, related information G, reference information, and output information W).

[0119] FIG. 20 is a block diagram illustrating the configuration of the information collection system 50. The information collection system 50 is a computer system that collects various types of information after presenting response information R to the user U. Specifically, the information collection system 50 includes a control device 51, a storage device 52, and a communication device 53.

[0120] The control device 51 is one or more processors that control the operation of the information collection system 50. The storage device 52 is one or more memories that store the programs executed by the control device 51 and various types of data used by the control device 51. The communication device 53 is a communication device that communicates with the terminal device 10 via the communication network 900. The control device 51, the storage device 52, and the communication device 53 adopt a configuration similar to that described for the control device 411, the storage device 412, and the communication device 413 of the first processing device 41, for example.

[0121] FIG. 21 is a block diagram illustrating the functions realized when the control device 51 executes the programs stored in the storage device 52. The control device 51 realizes a plurality of functions (acquisition unit 511, collection unit 512) for collecting various types of information after presenting the response information R to the user U.

[0122] The acquisition unit 511 receives the evaluation information F transmitted from the terminal device 10 via the communication device 53. That is, the acquisition unit 511 (an example of the "first acquisition unit") functions as an element that acquires the evaluation of validity (feedback evaluation) specified for the response information R generated for the request information corresponding to the input information input by the user U, the request information, and the response information R.

[0123] The collection unit 512 stores the evaluation information F acquired by the acquisition unit 511 in the storage device 60. The feedback evaluation, the request information, the response information R, and the identification information included in the evaluation information F are stored in the storage device 60. For example, in the database D1 stored in the storage device 60, the feedback evaluation, the request information, the response information R, and the identification information are registered in association with each other.

[0124] The storage device 60 is, for example, a recording medium (e.g., cloud storage) accessible by the information collection system 50 via the communication network 900 and is used as the storage device 60. The storage device 60 is preferably accessible by the processing unit 142 (generation model M2) of the first processing device 41 as described above in the first embodiment. Note that the storage device 52 of the information collection system 50 may be used as the storage device 60.

[0125] Before storing the evaluation information F in the storage device 60, there is a process of checking the content of the evaluation information F. First, it is checked whether the feedback evaluation is appropriate. For example, based on the request information, identification information, and response information R, it is checked whether the feedback evaluation is appropriate (i.e., whether the evaluation is valid). The check of whether the feedback evaluation is appropriate may be performed, for example, by the administrator of the information collection system 50 or by a processing device composed of a computer system. The processing device (not shown) checks whether the feedback evaluation is appropriate using, for example, a generation model. For such a generation model, for example, a model in which the relationship between the request information, identification information, response information R, and feedback evaluation and the validity of the feedback evaluation is machine-learned or a large language model is used.

[0126] When it is determined that the feedback evaluation is not appropriate, the feedback evaluation is appropriately corrected. The correction of the feedback evaluation may be performed, for example, by the administrator of the information collection system 50 or by the processing device. The processing device (not shown) corrects the feedback evaluation using, for example, a generation model. For such a generation model, for example, a model in which the relationship between the request information, identification information, response information R, and feedback evaluation and the corrected feedback evaluation is machine-learned or a large language model is used.

[0127] Second, it is confirmed whether the response information R is appropriate for the request information. For example, based on the request information, identification information, and feedback evaluation, it is confirmed whether the response information R is appropriate (that is, whether the response information R is valid). The confirmation of whether the response information R is appropriate may be performed, for example, by the administrator of the information collection system 50 or by a processing device composed of a computer system. The processing device (not shown) confirms whether the response information R is appropriate using, for example, a generation model. For such a generation model, for example, a model in which the relationship between the request information, identification information, and response information R and the validity of the response information R is machine-learned, or a large language model is used.

[0128] When it is determined that the response information R is not appropriate, the response information R is appropriately corrected. The correction of the response information R may be performed, for example, by the administrator of the information collection system 50 or by the processing device. The processing device (not shown) corrects the response information R using, for example, a generation model. For such a generation model, for example, a model in which the relationship between the request information, identification information, feedback evaluation, and response information R and the corrected response information R is machine-learned, or a large language model is used.

[0129] If it is determined that the feedback evaluation of the collection unit 512 of the present embodiment is not appropriate, the appropriately corrected feedback evaluation is stored in the storage device 60. If it is determined that the response information R is not appropriate for the input information, the appropriately corrected response information R is stored in the storage device 60. When the feedback evaluation and the response information R are appropriate, they are stored in the storage device 60 without correction.

[0130] Note that the collection unit 512 may further store in the storage device 60 whether the feedback evaluation is appropriate and the feedback evaluation before correction (when the feedback evaluation is not appropriate). Similarly, the collection unit 512 may store in the storage device 60 whether the response information R is appropriate and the response information R before correction (when the response information R is not appropriate).

[0131] According to the configuration of the third embodiment, when it is determined that the feedback evaluation is inappropriate, the appropriately corrected feedback evaluation is stored in the storage device 60. When it is determined that the response information R is inappropriate for the request information, the appropriately corrected response information R is stored in the storage device 60. Then, the information stored in the storage device 60 (database D1) can be used as training data for the generation model M2 of the first processing device 41 or can be referred to by the first processing device 41. Therefore, the first processing device 41 can generate the first output information W1 with higher accuracy. For example, a combination of the request information and the response information R (i.e., the first output information W1) is used as training data for the generation model M2. Also, the generation model M2 may refer to the database D1 to obtain the first output information W1 for the request information.

[0132] <Third Embodiment>

[0133] FIG. 22 is a block diagram illustrating functions realized by the control device 51 executing a program stored in the storage device 52 of the information collection system 50 according to the third embodiment.

[0134] In addition to the acquisition unit 511 and the collection unit 512, the control device 51 of the third embodiment realizes a prompt generation unit 513 and an information generation unit 514 as a plurality of functions for collecting various information after presenting the response information R to the user U. Note that the function of the information generation unit 514 may be mounted on a separate device capable of communicating with the information collection system 50.

[0135] First, the acquisition unit 511 (an example of the "first acquisition unit") of the third embodiment acquires the input information input by the user U. The prompt generation unit 513 generates a prompt P4 (an example of the "fourth prompt") including the input information. Specifically, the prompt P4 is information that requests to generate related information related to the input information. Note that in the prompt P4, it may be requested to generate a plurality of related information. The related information in the third embodiment is the same information as the related information G described in the first embodiment. That is, the related information is information for requesting the information desired by the user U, similar to the input information.

[0136] The prompt P4 generated by the prompt generation unit 513 is input to the information generation unit 514, and related information related to the input information is generated. The information generation unit 514 (an example of the "second acquisition unit") generates (acquires) the related information using the trained generation model M5 (an example of the "fourth generation model").

[0137] Next, the prompt generation unit 513 generates a prompt P5 (an example of the "fifth prompt") including the related information. Specifically, the prompt P5 is information for requesting to generate response information for the related information. The response information is information indicating a response to the related information (typically the information requested by the related information). When the related information is information indicating an inquiry, the response information is information indicating an answer to the inquiry. The response information has the same content as the first output information W1 (the answer to the inquiry).

[0138] The prompt P5 generated by the prompt generation unit 513 is input to the information generation unit 514, and response information for the related information is generated. The information generation unit 514 (an example of the "third acquisition unit") generates (acquires) the response information using the trained generation model M6 (an example of the "fifth generation model").

[0139] In the third embodiment, the information generation unit 514 that generates related information functions as an element (second acquisition unit) that acquires the related information related to the input information generated by the generation model M5 for the prompt P4 including the input information, and the information generation unit 514 that generates the response information functions as an element (third acquisition unit) that acquires the response information generated by the generation model M6 for the prompt P5 including the related information. However, when the function of the information generation unit 514 is mounted on an external device, the elements that acquire the related information and the response information generated by the device by receiving them function as the second acquisition unit and the third acquisition unit, respectively.

[0140] The collection unit 512 stores the input information acquired by the acquisition unit 511, the related information and the response information generated by the information generation unit 514 in the storage device 60. For example, in the database D2 stored in the storage device 60, the input information, the related information, and the response information are registered in association with each other. However, it is not essential to store the input information in the storage device 60.

[0141] Before storing the input information, the related information, and the response information in the storage device 60, it is confirmed whether the response information is appropriate. For example, based on the content of the related information, it is confirmed whether the response information is appropriate (that is, whether the response information is valid). The confirmation of whether the response information is appropriate may be performed, for example, by the administrator of the information collection system 50 or by a processing device configured by a computer system. The processing device (not shown) confirms whether the response information is appropriate using, for example, a generation model. For such a generation model, for example, a model in which the relationship between the related information and the response information and the validity of the response information is machine-learned, or a large language model is used.

[0142] When it is determined that the response information is not appropriate, the response information is appropriately corrected. The correction of the response information may be performed, for example, by the administrator of the information collection system 50 or by the processing device. The processing device (not shown) corrects the response information using, for example, a generation model. For such a generation model, for example, a model in which the relationship between the related information and the response information and the corrected response information is machine-learned, or a large language model is used.

[0143] In the third embodiment, when it is determined that the response information is not appropriate for the related information, the collection unit 512 stores the appropriately corrected response information in the storage device 60. When the response information is appropriate, it is stored in the storage device 60 without being corrected. Note that the collection unit 512 may further store in the storage device 60 whether the response information is appropriate or not, and the response information before correction (when the response information is not appropriate).

[0144] According to the configuration of the third embodiment, when it is determined that the response information is not appropriate, the appropriately corrected response information is stored in the storage device 60. Then, the information stored in the storage device 60 (database D2) can be used as training data for the information generation system 30 (generation model M1), or can be referred to by the information generation system 30. Therefore, the information generation system 30 can generate the related information G with higher accuracy. For example, the combination of the input information and the related information is used as training data for the generation model M1. Also, since the response information has the same content as the first output information W1, the combination of the related information (i.e., the candidate for the required information) and the response information (i.e., the first output information W1) may be used as training data for the generation model M2. Further, the generation model M1 may refer to the database D2 to obtain the related information for the input information, or the generation model M2 may refer to the database D2 to obtain the first output information W1 for the required information.

[0145] Note that the generation model M5 and the generation model M6 may be a single common generation model. Also, the generation model M5 may be common with the generation model M1, or the generation model M6 may be common with the generation model M2. Also, it may be determined whether the related information generated for the input information is appropriate, and when it is determined that the related information is not appropriate, the related information may be corrected and then stored in the storage device 60.

[0146] <Modification Example> Each of the embodiments exemplified above can be variously modified. Specific modification modes are exemplified below. It is also possible to appropriately combine two or more modes arbitrarily selected from the following examples.

[0147] (1) In each of the above-described embodiments, the case where the inquiry target is a specific target product is exemplified, but the target may be a service. Services include, for example, various services related to water supply, gas, electricity, banking, medical care, accommodation, education, and transportation. The contents of the reference information and the identification information can be appropriately changed according to the type of the target. For example, when the inquiry target is a service (for example, water supply, gas, electricity), as the identification information, information for identifying the plan or the contractor with which the service is contracted is exemplified, and as the reference information, information collected by the service provider (details of the plan with which the user has contracted, past usage history, etc.), or information generally disclosed by the service provider is exemplified as the reference information. Also, the target product is not limited to software, and various products such as home appliances, games, and foods are assumed.

[0148] (2) In each of the above-described embodiments, a part of the functions of the information generation system 30 or the information providing system 40 may be mounted on the terminal device 10. For example, a part or all of the related information generation unit 312, the processing unit 142, the processing unit 242, the index calculation unit 342, and the response generation unit 343 may be mounted on the terminal device 10. For example, in a configuration where the response generation unit 343 is mounted on the terminal device 10, the acquisition unit 113 that acquires the response information R is an element that generates the response information R (that is, the response generation unit 343). That is, the acquisition unit 113 may be an element that acquires the response information R by receiving it from an external device (in the first to third embodiments), or an element that acquires the response information R by generating it (in the configuration where the response generation unit 343 is mounted on the terminal device 10). Similarly, the acquisition unit 113 may be an element that acquires the related information G by receiving it from an external device (in the first to third embodiments), or an element that acquires the related information G by generating it (in the configuration where the related information generation unit 312 is mounted on the terminal device 10).

[0149] (3) In each of the above-described forms, a configuration for generating one or more pieces of related information G is not essential. That is, the input information may be used as the request information without presenting the related information G to the user U. The request information corresponding to the input information includes the case where the input information itself (including the case where the input information is modified or processed information) and the case where the related information G related to the input information (including the case where the related information G is modified or processed information).

[0150] (4) In each of the above-described forms, it is also assumed that a part or all of the functions of the prompt generation unit 112 are mounted on a device other than the terminal device 10 (for example, the information generation system 30 or the information providing system 40). For example, when the function of generating the prompts Pz2 and P3 is mounted on the information providing system 40, information (for example, request information or identification information) necessary for generating the prompts Pz2 and P3 is transmitted from the terminal device 10 to the information providing system 40.

[0151] Also, in each of the above-described forms, the information acquisition device 44 generates the prompt P2 from the prompt Pz2, but the function of generating the prompt P2 may be provided in, for example, the terminal device 10 or the first processing device 41.

[0152] (5) In each of the above-described forms, it is not essential that the prompt P2 includes identification information and reference information. However, by including the identification information and the reference information in the prompt P2, the generation model M2 can generate the first output information W1 with high accuracy (that is, high validity with respect to the request information). When the prompt P2 does not include reference information, the information acquisition device 44 is omitted from the information providing system 40, and the prompt Pz2 (an example of the "first prompt") is directly transmitted from the terminal device 10 to the first processing device 41. Similarly, it is not essential that the prompt P3 includes identification information. Also, the prompt P3 may include reference information. Also, the assignment of the identification information to the prompts Pz2 and P3 is not limited to the terminal device 10, and the information management system 20 may perform it.

[0153] (6) In each of the above-described embodiments, it is not essential that the response generation unit 343 generates the response information R according to the index K. In the above configuration, the index calculation unit 342 is omitted. For example, a configuration in which the response information R is generated by combining all of the generated plurality of output information W, or a configuration in which the response information R is generated by combining one first output information W1 and one second output information W2 is adopted. Further, the similarity (coincidence rate) between the first output information W1 and the second output information W2 may be calculated, and the response information R may be generated according to the similarity. For example, when the similarity exceeds a predetermined threshold value, the response information R including the second output information W2 may be generated, and when the similarity is less than the predetermined threshold value, the response information R including the first output information W1 may be generated.

[0154] (7) In the above-described embodiment, the response generation unit 343 may generate the response information R according to the output information W (or a plurality of output information W up to a predetermined rank in descending order of the index K) in which the index K is the largest among the indexes K specified for each of one or more first output information W1 and one or more second output information W2. That is, it is not essential to generate the response information R according to the result of comparing the index K of each output information W with the threshold value.

[0155] (8) In each of the above-described embodiments, the user U may be able to confirm the processing flow until the input information is input and the final response information R is obtained. For example, the input information input by the user U, the related information G generated from the input information, the request information selected by the user U from the input information and the related information G, and the response information R for the request information are made visible to the user U. These pieces of information are stored in the storage device 12 of the terminal device 10 or a storage device accessible from the terminal device 10 via the communication network 900.

[0156] (9) In each of the above-described embodiments, each of the information generation system 30, the information management system 20, the first processing device 41, the second processing device 42, the information providing device 43, the information acquisition device 44, and the information collection system 50 may be a single device or a computer system composed of a plurality of devices.

[0157] (10) The information processing system (terminal device, information generation system, information management system, information acquisition device, first processing device, second processing device, information providing device, information collection system) according to the foregoing embodiment is realized by the cooperation of a computer (specifically, a control device) and a program. The program according to the foregoing embodiment can be provided in a form stored in a computer-readable recording medium and installed in the computer. The recording medium is, for example, a non-transitory recording medium, and an optical recording medium (optical disk) such as a CD-ROM is a preferred example, but it may include any known form of recording medium such as a semiconductor recording medium or a magnetic recording medium. Note that the non-transitory recording medium includes any recording medium except a transitory, propagating signal, and does not exclude a volatile recording medium. It is also possible to provide the program to the computer in a form of distribution via a communication network.

[0158] <Supplementary Note> From the embodiments exemplified above, for example, the following configurations can be grasped.

[0159] [1] A terminal device comprising a reception unit that receives input information input by a user, and an acquisition unit that acquires response information generated for request information corresponding to the input information, wherein the response information is information generated according to one or more pieces of first output information generated by a trained first generation model for a first prompt including the request information, and one or more pieces of second output information generated by a trained second generation model for a second prompt including the request information.

[0160] According to the above configuration, response information generated according to one or more pieces of first output information generated by a trained first generation model for a first prompt including request information corresponding to input information input by a user, and one or more pieces of second output information generated by a trained second generation model for a second prompt including the request information can be obtained. Therefore, for example, compared with a configuration in which response information corresponding to output information generated by one generation model for a prompt including request information is obtained, more appropriate response information can be presented to the user for the request information from the user.

[0161] [2] The request information is information indicating an inquiry about a predetermined target, and the first generation model is a model trained using highly reliable information about the target as training data for the terminal device of [1].

[0162] In the above configuration, since the first model is a model trained using highly reliable information about a predetermined target as training data, the first output information can be generated based on the highly reliable information about the target.

[0163] [3] The response information is information generated according to an index specified for each of the one or more pieces of first output information and each of the one or more pieces of second output information, and the index indicates the validity for the request information for the terminal device of [1].

[0164] According to the above configuration, response information is generated according to an index specified for each of the one or more pieces of first output information and each of the one or more pieces of second output information, and since the index indicates the validity for the request information, the response information generated taking into account the validity for the request information can be presented to the user.

[0165] [4] The response information is generated according to the result of comparing the index with a threshold value for the terminal device of [3].

[0166] In the above configuration, response information is generated according to the result of comparing the index with the threshold value. Therefore, for example, response information generated according to one or more pieces of output information (i.e., output information with high validity) in which the index exceeds the threshold value can be presented to the user.

[0167] [5] The method by which the response information is generated differs according to the number of pieces of output information in which the index exceeds the threshold value among each of the one or more pieces of first output information and each of the one or more pieces of second output information [4] of the terminal device.

[0168] In the above configuration, response information generated according to different methods according to the number of pieces of output information in which the index exceeds the threshold value among each of the one or more pieces of first output information and each of the one or more pieces of second output information can be presented to the user.

[0169] [6] The method by which the response information is generated differs depending on whether there is one piece of output information in which the index exceeds the threshold value among the one or more pieces of first output information and the one or more pieces of second output information, whether there are a plurality of pieces of output information in which the index exceeds the threshold value among the one or more pieces of first output information and the one or more pieces of second output information, and whether there is no piece of output information in which the index exceeds the threshold value among the one or more pieces of first output information and the one or more pieces of second output information [5] of the terminal device.

[0170] In the above configuration, response information generated by different methods depending on whether there is one piece of output information in which the index exceeds the threshold value among the one or more pieces of first output information and the one or more pieces of second output information, whether there are a plurality of pieces of output information in which the index exceeds the threshold value among the one or more pieces of first output information and the one or more pieces of second output information, and whether there is no piece of output information in which the index exceeds the threshold value among the one or more pieces of first output information and the one or more pieces of second output information can be presented to the user.

[0171] [7] The response information is generated according to the one piece of output information when there is one piece of output information that exceeds the threshold value, and is generated according to the plurality of pieces of output information when there are a plurality of pieces of output information that exceed the threshold value [6] of the terminal device.

[0172] In the above configuration, when there is one piece of output information exceeding the threshold value, response information is generated according to the one piece of output information. When there are multiple pieces of output information exceeding the threshold value, response information is generated according to the multiple pieces of output information. Therefore, when there are multiple pieces of highly reliable output information, it is possible to present the user with response information taking into account all of the multiple pieces of output information.

[0173] [8] The first prompt includes reference information referred to in the generation of the first output information. The response information is, when there is no output information exceeding the threshold value, for each of one or more first output information newly generated by the first generation model for the first prompt in which the reference information is changed to other reference information, and for each of the one or more second output information, generated according to the specified index. The terminal device of [7].

[0174] In the above configuration, when there is no output information exceeding the threshold value, response information is generated according to the specified index for each of one or more first output information newly generated by the first generation model for the first prompt in which the reference information is changed to other reference information, and for each of the one or more second output information. Therefore, when highly reliable output information cannot be obtained, one or more new first output information (that is, one or more updated first output information) can be used for the generation of response information.

[0175] [9] The acquisition unit causes the trained third generation model to generate one or more pieces of related information related to the input information for the third prompt including the input information, and the request information is the information selected by the user from among the input information and the one or more pieces of related information. The terminal device of [1] to [8].

[0176] In the above configuration, since the requested information is the information selected by the user from the input information and one or more related information, response information is generated with the requested information being the information desired by the user among the multiple options (input information and one or more related information). Therefore, for example, compared with a configuration in which the input information entered by the user is used as the requested information, response information that conforms to the user's intention can be obtained.

[0177]

[10] The requested information is information indicating an inquiry regarding a predetermined target, and the first prompt is the terminal device of [1] including identification information for identifying the target.

[0178] In the above configuration, since the requested information is information indicating an inquiry regarding a predetermined target and the first prompt includes identification information for identifying the target, response information for the requested information can be generated with high accuracy.

[0179]

[11] The first prompt is the terminal device of [1] to [7] including reference information referred to in the generation of the response information.

[0180] In the above configuration, since the first prompt includes reference information referred to in the generation of the response information, response information for the requested information can be generated with high accuracy.

[0181]

[12] A terminal device comprising a reception unit that receives input information entered by a user, and an acquisition unit that acquires one or more related information related to the input information, which is generated by a trained third generation model for a third prompt including the input information.

[0182] In the above configuration, since it includes a reception unit that receives input information input by a user, and an acquisition unit that generates, using a trained third generation model, one or more pieces of related information related to the input information for a third prompt including the input information, it is possible to provide the user with one or more pieces of related information related to the input information. Therefore, for example, it is possible to generate a prompt including requested information, which is information desired by the user among a plurality of options (input information and one or more pieces of related information). As a result, it is possible to present the user with response information more appropriate for the requested information.

[0183]

[13] An information providing apparatus including a response generation unit that generates response information for requested information according to one or more pieces of first output information generated by a trained first generation model for a first prompt including the requested information corresponding to input information input by a user and one or more pieces of second output information generated by a trained second generation model for a second prompt including the requested information.

[0184] In the above configuration, response information for the requested information is generated according to one or more pieces of first output information generated by a trained first generation model for a first prompt including the requested information corresponding to input information input by a user and one or more pieces of second output information generated by a trained second generation model for a second prompt including the requested information. Therefore, for example, compared with a configuration in which response information corresponding to output information generated by one generation model for a prompt including the requested information is acquired, it is possible to present the user with more appropriate information for the requested information.

[0185]

[14] An information generation system including a related information generation unit that generates one or more pieces of related information related to input information by inputting a third prompt including the input information input by a user to a trained third generation model.

[0186] In the above configuration, by inputting a third prompt including the input information entered by the user into the third generation model, one or more pieces of related information are generated, so that one or more pieces of related information related to the input information can be provided to the user. Therefore, for example, a prompt including the requested information, which is the information desired by the user among a plurality of options (input information and one or more pieces of related information), can be generated. Consequently, more appropriate response information can be presented to the user for the requested information.

[0187]

[15] An information collection system comprising: a first acquisition unit that acquires an evaluation of the validity specified for the response information generated for the requested information corresponding to the input information entered by the user, the requested information, and the response information; and a collection unit that stores the evaluation, the requested information, and the response information in a storage device.

[0188] In the above configuration, the evaluation of the validity specified for the response information generated for the requested information corresponding to the input information entered by the user, the requested information, and the response information are stored in the storage device. Therefore, the evaluation, the requested information, and the response information stored in the storage device can be used, for example, as training data for the first generation model, or can be referred to when the first generation model generates the first output information. That is, the first generation model can generate the first output information with high accuracy. Consequently, more appropriate response information can be presented to the user for the requested information.

[0189]

[16] When it is determined that the evaluation is not appropriate, the collection unit stores the appropriately corrected evaluation in the storage device, and when it is determined that the response information is not appropriate for the requested information, the collection unit stores the appropriately corrected response information in the storage device

[14] of the information collection system.

[0190] In the above configuration, when it is determined that the evaluation is inappropriate, the appropriately corrected evaluation is stored in the storage device, and when it is determined that the response information is inappropriate for the request information, the appropriately corrected response information is stored in the storage device. Therefore, the effect that the first generation model can generate the first output information with high accuracy becomes remarkable.

[0191]

[17] A first acquisition unit that generates, by a trained fourth generation model, a fourth prompt including input information input by a user and acquires related information related to the input information, a second acquisition unit that acquires response information generated by a trained fifth generation model for a fifth prompt including the related information, and a collection unit that stores the related information and the response information in a storage device. An information collection system.

[0192] In the above configuration, the trained fourth generation model generates a fourth prompt including the input information input by the user, and the related information related to the input information and the response information generated by the trained fifth generation model for the fifth prompt including the related information are stored in the storage device. Therefore, it is possible to use the related information and the response information stored in the storage device, for example, as training data for the third generation model or to refer to them when the third generation model generates related information. Therefore, the third generation model can generate related information with high accuracy.

[0193]

[18] The collection unit stores the appropriately corrected response information in the storage device when it is determined that the response information is inappropriate for the related information. The information collection system of

[16] .

[0194] In the above configuration, when it is determined that the response information is inappropriate for the related information, the appropriately corrected response information is stored in the storage device, so the effect that the third generation model can generate related information with high accuracy is remarkable.

[0195]

[19] An information processing system comprising: a reception unit that receives input information input by a user; one or more first output information generated by a first generation model trained for a first prompt including request information corresponding to the input information; and one or more second output information generated by a second generation model trained for a second prompt including the request information, a response generation unit that generates response information for the request information according to the one or more first output information and the one or more second output information; and an acquisition unit that acquires the response information generated for the request information corresponding to the input information.

[0196] In the above configuration, since the response information generated according to the one or more first output information generated by the first generation model trained for the first prompt including the request information corresponding to the input information input by the user and the one or more second output information generated by the second generation model trained for the second prompt including the request information can be obtained, for example, compared with a configuration in which response information corresponding to the output information generated by one generation model for a prompt including the request information is obtained, it is possible to present more appropriate response information to the user for the request information from the user.

[0197]

[20] An information acquisition method realized by a computer system that receives input information input by a user, acquires response information generated for request information corresponding to the input information, and the response information is information generated according to one or more first output information generated by a first generation model trained for a first prompt including the request information and one or more second output information generated by a second generation model trained for a second prompt including the request information.

[0198] In the above configuration, since response information can be obtained according to one or more pieces of first output information generated by a trained first generation model for a first prompt including request information corresponding to input information input by a user and one or more pieces of second output information generated by a trained second generation model for a second prompt including the request information, for example, compared with a configuration in which response information according to output information generated by one generation model for a prompt including request information is obtained, it is possible to present more appropriate response information to the user for the request information from the user.

[0199]

[21] A program that causes a computer system to function as a reception unit that receives input information input by a user and an acquisition unit that acquires response information generated for request information corresponding to the input information, and the response information is information generated according to one or more pieces of first output information generated by a trained first generation model for a first prompt including the request information and one or more pieces of second output information generated by a trained second generation model for a second prompt including the request information.

[0200] In the above configuration, since response information can be obtained according to one or more pieces of first output information generated by a trained first generation model for a first prompt including request information corresponding to input information input by a user and one or more pieces of second output information generated by a trained second generation model for a second prompt including the request information, for example, compared with a configuration in which response information according to output information generated by one generation model for a prompt including request information is obtained, it is possible to present more appropriate response information to the user for the request information from the user.

[0201] It is also specified as an invention of a method respectively executed by the information providing device of

[12] , the information providing device of

[13] , the information generation system of

[14] , and the information collection systems of

[15]

[17] . Further, it is also specified as an invention of a program that causes the information providing device of

[12] , the information generation system of

[13] , the information collection systems of

[15]

[17] , and the information processing system of

[19] to function respectively.

Explanation of Reference Numerals

[0202] 10: Terminal device 11: Control device 12: Storage device 13: Communication device 14: Display device 15: Operating device 20: Information management system 30: Information generation system 31: Control device 32: Storage device 33: Communication device 40: Information providing system 41: First processing device 42: Second processing device 43: Information providing device 44: Information acquisition device 50: Information collection system 51: Control device 52: Storage device 53: Communication device 60: Storage device 100: Information processing system 111: Reception unit 112: Prompt generation unit 113: Acquisition unit 114: Presentation unit 141: Acquisition unit 142: Processing unit 241: Acquisition unit 242: Processing unit 311: Acquisition unit 312: Related information generation unit 341: Acquisition unit 342: Index calculation unit 343: Response Generation Unit 411: Control Device 412: Memory Device 413: Communication Device 421: Control Device 422: Memory Device 423: Communication Device 431: Control Device 432: Memory Device 433: Communication Device 511: Acquisition Unit 512: Collection Unit 513: Prompt Generation Unit 514: Information Generation Unit 900: Communication Network F: Feedback Evaluation G: Related Information K: Index M1 - M6: Generation Model P1 - P6: Prompt W1, W2: Output Information

Claims

1. A reception unit that receives input information input by a user; an acquisition unit that acquires response information generated in response to request information corresponding to the input information, A terminal device, wherein the response information is information generated in response to one or more first output information generated by a trained first generative model in response to a first prompt including the request information, and one or more second output information generated by a trained second generative model in response to a second prompt including the request information.

2. the request information is information indicating an inquiry regarding a predetermined object, The first generative model is a model trained using reliable information about the target as training data. The terminal device of claim 1.

3. the response information is information generated in response to an index identified for each of the one or more pieces of first output information and each of the one or more pieces of second output information; The indicator indicates the validity of the requested information. The terminal device of claim 1.

4. The response information is generated in response to a result of comparing the index with a threshold value. The terminal device according to claim 3.

5. A method for generating the response information differs depending on the number of pieces of output information whose index exceeds the threshold value, among the one or more pieces of first output information and the one or more pieces of second output information. The terminal device according to claim 4.

6. The method for generating the response information includes: A case in which the index exceeds the threshold value among the one or more first output information and the one or more second output information is one output information; A case in which there is a plurality of pieces of output information in which the index exceeds the threshold value among the one or more pieces of first output information and the one or more pieces of second output information; The one or more pieces of first output information and the one or more pieces of second output information are different from a case where there is no output information in which the index exceeds the threshold value. The terminal device according to claim 5.

7. The response information is When there is only one piece of output information exceeding the threshold value, the output information is generated in response to the single piece of output information, When there are a plurality of pieces of output information exceeding the threshold value, the output information is generated according to the plurality of pieces of output information. The terminal device according to claim 6.

8. The first prompt includes reference information that is referenced in generating the first output information, and when there is no output information exceeding the threshold, the response information is generated according to the indexes identified for each of the one or more pieces of first output information newly generated by the first generative model in response to the first prompt in which the reference information is changed to other reference information and each of the one or more pieces of second output information. The terminal device according to claim 7.

9. The acquisition unit acquires one or more pieces of related information related to the input information, the one or more pieces of related information being generated by a trained third generative model in response to a third prompt including the input information; The request information is information selected by the user from the input information and the one or more pieces of related information. The terminal device of claim 1.

10. the request information is information indicating an inquiry regarding a predetermined object, The first prompt includes identification information for identifying the subject. The terminal device according to claim 1

11. The first prompt includes reference information that is referenced in generating the response information. The terminal device of claim 1.

12. a response generation unit that generates response information to a first prompt including request information corresponding to input information input by a user, in response to one or more pieces of first output information generated by a trained first generation model in response to a first prompt including the request information, and in response to one or more pieces of second output information generated by a trained second generation model in response to a second prompt including the request information; An information providing device comprising:

13. A system capable of communicating with the terminal device of claim 1, a first acquisition unit that acquires an evaluation of validity determined for the response information, the request information, and the response information; a collection unit that stores the evaluation, the request information, and the response information in a storage device; An information gathering system comprising:

14. A system capable of communicating with the terminal device of claim 1, A first acquisition unit that acquires related information related to the input information by generating a fourth generative model that has been trained for a fourth prompt including the input information; A second acquisition unit that acquires answer information generated by a trained fifth generation model in response to a fifth prompt including the related information; a collection unit that stores the related information and the answer information in a storage device; An information gathering system comprising:

15. A reception unit that receives input information input by a user; a response generation unit that generates response information to the request information in response to one or more pieces of first output information generated by a trained first generation model in response to a first prompt including request information corresponding to the input information, and one or more pieces of second output information generated by a trained second generation model in response to a second prompt including the request information; an acquisition unit for acquiring the response information; An information processing system comprising:

16. Accept the input information entered by the user, Obtaining response information generated in response to request information corresponding to the input information; An information acquisition method implemented by a computer system, wherein the response information is information generated in response to one or more first output information generated by a trained first generative model in response to a first prompt including the request information, and one or more second output information generated by a trained second generative model in response to a second prompt including the request information.

17. a reception unit that receives input information input by a user; and causing the computer system to function as an acquisition unit that acquires response information generated in response to request information corresponding to the input information; The response information is information generated in response to one or more first output information generated by a trained first generative model in response to a first prompt including the request information, and one or more second output information generated by a trained second generative model in response to a second prompt including the request information.

Images