Architecture proposed device, information processing method and program

Abstract

We provide technology for generating appropriate architectures. [Solution] The architecture proposal device according to one aspect of this disclosure receives user requirements for the architecture of an information system, generates a proposed architecture for the information system in a large-scale generation model, generates verification results of the proposed architecture in the large-scale generation model, generates inquiries regarding deficiencies in the large-scale generation model, outputs the generated inquiries to the user, receives input of answers to the inquiries from the user, updates the user requirements based on the input answers, and generates the updated proposed architecture in the large-scale generation model. The architecture proposal device improves the proposed architecture by repeating the process from verification to updating of the proposed architecture.

Description

Technical Field

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

Background Art

[0002] In recent years, for example, the technological development of large-scale generative models such as large language models, large vision-language models, and large speech models has been progressing. It is known that large-scale generative models can acquire common sense by learning a large amount of data. By using such a large-scale generative model, arbitrary information can be generated. For example, in Patent Document 1, it is proposed to use a large language model for the generation of program code.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The inventors of the present case have found the following problems. That is, it is assumed that a large-scale generative model is used to design an architecture related to an information system. In an architecture related to an information system, the patterns of components such as the types of modules, specifications, and setting methods are diverse. Therefore, it is difficult to enumerate all the requirements for the architecture at the first stage. When the requirements for the architecture are insufficient, it is difficult to generate an appropriate architecture.

[0005] In one aspect, this disclosure has been made in view of such circumstances, and one of its purposes is to provide a technique for generating an appropriate architecture.

Means for Solving the Problems

[0006] This disclosure adopts the following configuration to solve the aforementioned problems. Note that the following configurations can be combined as appropriate.

[0007] The proposed architecture device relating to one aspect of this disclosure includes a control unit. The control unit receives user requirements for the architecture of the information system from the user, generates a proposed architecture for the information system in a large-scale generation model by providing a first prompt including the input user requirements and proposal instructions, generates verification results for the proposed architecture in the large-scale generation model by providing a second prompt including the generated proposed architecture and verification instructions, generates inquiries regarding the deficiencies in the large-scale generation model if there are deficiencies in the generated verification results by providing a third prompt including the proposed architecture, verification results, and inquiry generation instructions, outputs the generated inquiries to the user, receives input of answers to the inquiries from the user, updates the user requirements based on the input answers, generates an updated proposed architecture in the large-scale generation model by providing a fourth prompt including the updated user requirements, proposed architecture, verification results, inquiries, and re-proposal instructions, and improves the proposed architecture by repeating the process of generating verification results for the proposed architecture, generating inquiries, outputting inquiries, receiving input of answers to inquiries, updating user requirements, and generating the updated proposed architecture until no more deficiencies exist, and then outputs the improved proposed architecture. It is composed.

[0008] In this configuration, an alternating dialogue between the user and the large-scale generative model takes place, involving repeated proposals, validations, and inquiries to the user. This dialogue facilitates the refinement of user requirements through repeated responses to inquiries, thereby improving the proposed architecture. As a result, this configuration can be expected to generate an appropriate architecture.

[0009] In the proposed architecture device relating to the above aspect, the verification instructions may include a verification list containing a predefined set of verification items. The deficiencies may be those verification items included in the verification list that were evaluated as deficiencies in the verification by the large-scale generative model. With this configuration, by defining the verification items in advance, it is possible to guarantee the scope of verification.

[0010] In the architecture proposal device relating to the above aspect, the third prompt may further include the user's skill level for architecture design. With this configuration, it is possible to expect the generation of inquiries that correspond to the user's skill level.

[0011] In the proposed architecture relating to the above aspect, the user's skill level may be inferred from the user's input. The user's input may reflect the user's skill level. Therefore, it is possible to infer the skill level from the user's input. With this configuration, by automatically inferring the skill level from the user's input, it is possible to reduce the effort required to have the user answer about their skill level.

[0012] In the proposed architecture relating to the above aspect, the control unit may be configured to further perform the task of having a large-scale generative model estimate the user's skill level by providing a fifth prompt that includes user input and instructions for estimating the skill level. With this configuration, it becomes unnecessary to prepare a dedicated model for estimating the skill level, and thus a reduction in cost can be expected.

[0013] In the architecture proposal device relating to the above aspect, the control unit may be configured to, if the input to an inquiry includes a response indicating that it cannot be answered, to further perform the following actions: generate a modified inquiry in a large-scale generation model by providing a sixth prompt including the inquiry, the user's skill level for architecture design, and modification instructions; output the modified inquiry to the user; and receive input from the user for the answer to the modified inquiry. Updating user requirements may be performed by updating the user requirements with the answer entered for the modified inquiry. With this configuration, when an inquiry that the user cannot answer is generated, it becomes possible to obtain an appropriate answer from the user by remaking the inquiry to match the user's skill level.

[0014] In the architecture proposal device relating to the above aspect, the control unit may be configured to further perform the function of associating and saving user input, instructions for the large-scale generative model, and the generation results of the large-scale generative model as a dialogue history in chronological order. With this configuration, it is possible to verify afterward whether the generation process of the proposed architecture is appropriate by using the obtained dialogue history.

[0015] In the architecture proposal device relating to the above aspect, the control unit may be configured to further output the generated proposed architecture to the user and to receive input from the user specifying the parts of the proposed architecture that should not be changed. The draft instructions may include a statement indicating that the specified parts should not be changed. With this configuration, the designation of parts that do not need to be changed can be accepted, and this designation can be reflected in the revised instructions, ensuring that the specified parts are not changed when resubmitting the architecture.

[0016] In the architecture proposal device relating to the above aspect, the control unit may be configured to further output the generated proposed architecture to the user and to receive input from the user specifying the parts to be changed in the proposed architecture. The re-proposal instruction may include an instruction statement indicating that the specified parts should be changed. With this configuration, it is possible to prompt the user to change the specified parts when re-proposing the architecture.

[0017] In the architecture proposal device relating to the above aspect, the control unit may be configured to further cause a large-scale generative model to generate alternatives to the proposed architecture by providing a seventh prompt that includes the proposed architecture, verification results, and instructions for generating alternatives. The third prompt may further include the generated alternatives. With this configuration, understanding of the proposed architecture can be facilitated by generating alternatives. By including the generated alternatives in the third prompt, it is possible to expect an improvement in the accuracy of generating queries.

[0018] In the architecture proposal device relating to the above aspect, the control unit may be configured to further perform the following actions: provide an eighth prompt including the proposed architecture and a summary instruction, cause the large-scale generative model to generate a summary of the proposed architecture, and output the generated summary to the user. With this configuration, the output summary makes it easier for the user to understand the proposed architecture.

[0019] In the proposed architecture device relating to the above aspect, the query generation instruction may include an instruction statement that generates a query item with restricted input format for the answer. With this configuration, it is possible to obtain a query item that is easy for the user to answer.

[0020] In the proposed architecture device relating to the above aspect, the query generation instruction may include an instruction statement that lists query items in order of priority. With this configuration, it is possible to obtain query items that are easy for the user to answer.

[0021] In the proposed architecture device relating to the above aspect, the query generation instruction may include an instruction statement that generates query items in a format that asks for the user's request. With this configuration, it can be expected that the likelihood of generating appropriate query items will be improved.

[0022] In the architecture proposal device relating to the above aspect, the control unit may be configured to manage user requirements, proposed architecture, verification results, and inquiries as structured internal information. With this configuration, the construction of each prompt is easy because the internal information is structured. Therefore, it can be expected that the implementation of the system (architecture proposal device) will be simplified.

[0023] Furthermore, the form of this disclosure is not limited to the above-described architecture-proposed device (information processing device). As another aspect of the architecture-proposed device relating to each of the above aspects, one aspect of this disclosure may be an information processing method (architecture-proposed method) that realizes all or part of the above-described configurations, or a program, or a machine-readable storage medium that stores such a program. Here, a machine-readable storage medium may be a non-temporary medium that stores information such as programs by electrical, magnetic, optical, mechanical or chemical action. A non-temporary storage medium may include storage media (CDs, DVDs, semiconductor memory, etc.), auxiliary storage devices of a computer, external storage devices connected to a computer, etc.

[0024] For example, an information processing method (architecture proposal method) according to an aspect of the present disclosure may be executed by a computer. The information processing method includes receiving an input of user requirements for an architecture related to an information system from a user, providing a first prompt including the input user requirements and a proposal instruction to cause a large-scale generation model to generate a proposed architecture of the information system, providing a second prompt including the generated proposed architecture and a verification instruction to cause the large-scale generation model to generate a verification result of the proposed architecture, in the generated verification result, if there are deficiencies, providing a third prompt including the proposed architecture, the verification result, and a generation instruction for an inquiry to cause the large-scale generation model to generate inquiry items related to the deficiencies, outputting the generated inquiry items to the user, receiving an input of an answer to the inquiry items from the user, updating the user requirements based on the input answer, providing a fourth prompt including the updated user requirements, the proposed architecture, the verification result, the inquiry items, and a re-proposal instruction to cause the large-scale generation model to generate an updated proposed architecture, repeating generating a verification result of the proposed architecture, generating inquiry items, outputting the inquiry items, receiving an input of an answer to the inquiry items, updating the user requirements, and generating an updated proposed architecture until there are no deficiencies, improving the proposed architecture, and outputting the improved proposed architecture.

[0025] Also, for example, a program according to an aspect of the present disclosure (architecture proposal program) may be a program for causing a computer to execute an information processing method. The information processing method includes receiving an input of user requirements for an architecture related to an information system from a user, providing a first prompt including the input user requirements and a proposal instruction to cause a large-scale generation model to generate a proposed architecture of the information system, providing a second prompt including the generated proposed architecture and a verification instruction to cause the large-scale generation model to generate a verification result of the proposed architecture, in the generated verification result, if there are deficiencies, providing a third prompt including the proposed architecture, the verification result, and a generation instruction for an inquiry to cause the large-scale generation model to generate inquiry items regarding the deficiencies, outputting the generated inquiry items to the user, receiving an input of an answer to the inquiry items from the user, updating the user requirements based on the input answer, providing a fourth prompt including the updated user requirements, the proposed architecture, the verification result, the inquiry items, and a re-proposal instruction to cause the large-scale generation model to generate an updated proposed architecture, repeating generating a verification result of the proposed architecture, generating inquiry items, outputting the inquiry items, receiving an input of an answer to the inquiry items, updating the user requirements, and generating an updated proposed architecture until there are no more deficiencies, improving the proposed architecture, and outputting the improved proposed architecture.

Effect of the Invention

[0026] According to an aspect of the present disclosure, it is possible to provide a technique for generating an appropriate architecture.

Brief Description of the Drawings

[0027] [Figure 1] FIG. 1 schematically shows an example of a scene to which the present disclosure is applied. [Figure 2]Figure 2 schematically shows an example of a scenario where user requirements are received and a first prompt is provided. [Figure 3] Figure 3 schematically shows an example of a scenario in which a second prompt is given. [Figure 4] Figure 4 schematically shows an example of a scenario in which a third prompt is given. [Figure 5] Figure 5 schematically shows an example of a scenario where the inquiry is output, the user is prompted to enter a response, and a fourth prompt is given. [Figure 6] Figure 6 schematically illustrates an example of a method for generating inquiries based on skill levels. [Figure 7] Figure 7 schematically shows an example of how to correct the inquiry. [Figure 8] Figure 8 schematically shows an example of a method for controlling the format of inquiry items. [Figure 9] Figure 9 schematically illustrates an example of a scenario in which alternative solutions are generated. [Figure 10] Figure 10 schematically illustrates one example of how to control the scope of architectural re-proposal. [Figure 11] Figure 11 schematically shows an example of a scenario in which a summary is generated. [Figure 12] Figure 12 schematically shows an example of a scenario in which a dialogue history is generated. [Figure 13] Figure 13 schematically shows an example of internal information. [Figure 14] Figure 14 shows a specific example of the transition of internal information. [Figure 15] Figure 15 shows a specific example of the transition of internal information. [Figure 16] Figure 16 shows a specific example of the transition of internal information. [Figure 17] Figure 17 shows a specific example of the transition of internal information. [Figure 18] Figure 18 shows a specific example of the transition of internal information. [Figure 19] Figure 19 shows a specific example of the transition of internal information. [Figure 20] Figure 20 shows a specific example of the transition of internal information. [Figure 21] Figure 21 schematically shows an example of the hardware configuration of the proposed architecture device. [Figure 22] Figure 22 schematically shows an example of the software configuration of the proposed architecture device. [Figure 23] Figure 23 is a flowchart showing an example of the processing procedure for architecture proposal by the architecture proposal device. [Figure 24] Figure 24 is a flowchart showing an example of the processing procedure for architecture proposal by the architecture proposal device. [Figure 25] Figure 25 shows the architecture initially proposed in the first experimental example. [Figure 26] Figure 26 shows the architecture initially proposed in the first experimental example. [Figure 27] Figure 27 shows the verification results of the architecture initially proposed in the first experimental example. [Figure 28] Figure 28 shows the questions and answers regarding the architecture initially proposed in the first experimental example. [Figure 29] Figure 29 shows the architecture proposed for the second time in the first experimental example (the result of the first revision of the architecture). [Figure 30] Figure 30 shows the architecture proposed for the second time in the first experimental example (the result of the first revision of the architecture). [Figure 31] Figure 31 shows the verification results of the second proposed architecture in the first experimental example. [Figure 32] Figure 32 shows the questions and answers regarding the second proposed architecture in the first experimental example. [Figure 33] Figure 33 shows the architecture proposed for the third time in the first experimental example (the result of the second revision of the architecture). [Figure 34] Figure 34 shows the architecture proposed for the third time in the first experimental example (the result of the second revision of the architecture). [Figure 35] Figure 35 shows the instruction given to the large-scale generative model as an instruction to generate inquiry items in the form of a question asking for user requests. [Figure 36] Figure 36 shows the result of generating query items by providing a prompt that does not contain instructions to generate query items in a format that asks for the user's request. [Figure 37] Figure 37 shows the result of generating query items by providing a prompt that includes instructions for generating query items in a format that asks for the user's request. [Modes for carrying out the invention]

[0028] Hereinafter, embodiments relating to one aspect of this disclosure will be described with reference to the drawings. However, the embodiments described below are merely illustrative in all respects of this disclosure. Various improvements or modifications may be made without departing from the scope of this disclosure. In implementing this disclosure, specific configurations may be adopted as appropriate depending on the embodiment. In this embodiment, the data appearing is described in natural language, but more specifically, it is specified in pseudo-language, commands, parameters, machine code, electrical signals, etc., that can be recognized by machines such as computers.

[0029] §1 Examples of Application Figure 1 schematically shows an example of a scenario to which this disclosure applies. Figures 2 to 5 schematically show an example of a scenario for each step. The architecture proposal device 1 according to this embodiment is one or more computers configured to control the interaction between user Z and large-scale generative model 2, thereby causing the large-scale generative model 2 to construct an architecture relating to an information system.

[0030] The architecture proposal device 1 according to this embodiment receives user requirements 30 (30A) for the architecture of an information system from user Z (Figure 1(A), Figure 2). The architecture proposal device 1 causes the large-scale generation model 2 to generate a proposed architecture 40 (40A) for the information system by providing a first prompt P1 that includes the input user requirements 30 (30A) and a proposal instruction 61 (Figure 1(B), Figure 2). The architecture proposal device 1 causes the large-scale generation model 2 to generate a verification result 45 of the proposed architecture 40 (40A) by providing a second prompt P2 that includes the generated proposed architecture 40 (40A) and a verification instruction 62 (Figure 1(C), Figure 3).

[0031] If any deficiencies exist in the generated verification result 45, the architecture proposal device 1 provides a third prompt P3 including the proposed architecture 40 (40A), the verification result 45, and a query generation instruction 63, causing the large-scale generation model 2 to generate a query item 49 regarding the deficiencies (Figure 1(D), Figure 4). The architecture proposal device 1 outputs the generated query item 49 to user Z and accepts input of an answer 31 to the query item 49 from user Z (Figure 1(E)(F), Figure 5). The architecture proposal device 1 updates the user requirement 30 (30B) based on the input answer 31 (Figure 5). For convenience of explanation, the user requirement 30 before the update is also referred to as "user requirement 30A," and the user requirement 30 after the update is also referred to as "user requirement 30B."

[0032] The architecture proposal device 1 causes the large-scale generation model 2 to generate the updated proposed architecture 40 (40B) by providing a fourth prompt P4 that includes the updated user requirements 30 (30B), proposed architecture 40 (40A), verification results 45, inquiry items 49, and re-proposal instruction 64 (Figure 1(G), Figure 5). For convenience of explanation, the proposed architecture 40 before the update (re-proposal) will also be referred to as "proposed architecture 40A," and the proposed architecture 40 after the update (re-proposal) will also be referred to as "proposed architecture 40B."

[0033] The architecture proposal device 1 improves the proposed architecture 40 by repeatedly generating verification results 45 of the proposed architecture 40, generating inquiry items 49, outputting inquiry items 49, accepting input of answers 31 to inquiry items 49, updating user requirements 30, and generating the updated proposed architecture 40 until no deficiencies remain. The architecture proposal device 1 outputs the improved proposed architecture 40.

[0034] Note that user requirements 30 and answers 31 may be submitted using any method. The combined item 49 may be output in any way. That is, the method of inputting and outputting information to user Z is not particularly limited and may be appropriately selected depending on the embodiment.

[0035] In one example, as shown in each figure, user Z may access the architecture proposal device 1 using user terminal U1. User terminal U1 may be a computer separate from the architecture proposal device 1. The type of computer for user terminal U1 is not particularly limited and may be appropriately selected depending on the embodiment. User terminal U1 may be, for example, a general-purpose PC (Personal Computer), a notebook PC, a terminal device, etc. The terminal device is This may include smartphones, tablet devices, etc. In this case, information input and output to user Z may be performed via user terminal U1.

[0036] In other words, the architecture proposal device 1 may indirectly receive input of user requirements 30 and answers 31 via the user terminal U1. The transmission path for the input user requirements 30 and answers 31 may be appropriately selected depending on the embodiment. The user requirements 30 and answers 31 may be transmitted to the large-scale generation model 2 via the architecture proposal device 1, or they may be transmitted to the large-scale generation model 2 without passing through the architecture proposal device 1. The architecture proposal device 1 may also output the generated inquiry items 49 to the user terminal U1. The transmission path for the inquiry items 49 may also be appropriately selected depending on the embodiment. The inquiry items 49 may be transmitted to the user terminal U1 via the architecture proposal device 1, or they may be transmitted to the user terminal U1 without passing through the architecture proposal device 1.

[0037] In another example, the user terminal U1 and the architecture proposal device 1 may be configured as a single computer. That is, the architecture proposal device 1 may also function as the user terminal U1. In this case, the architecture proposal device 1 may directly accept input of user requirements 30 and answers 31 via an input device. The architecture proposal device 1 may directly output inquiry items 49 via an output device. The user terminal U1 may be omitted from each figure.

[0038] In this embodiment, the architecture proposal device 1 repeatedly initiates an alternating dialogue between user Z and large-scale generative model 2, proposing an architecture (generating and updating the proposed architecture 40), verifying it (generating verification results 45), and querying user Z (receiving input of user requirements 30, output of inquiry items 49, and input of answers 31). In this dialogue, the repeated output of answers 31 to inquiry items 49 promotes the refinement of user requirements 30. By refining user requirements 30, the proposed architecture 40 obtained from the large-scale generative model 2 can be polished. As a result, according to this embodiment, it can be expected that an appropriate architecture (proposed architecture 40) will be generated.

[0039] [Information Systems / Architecture] An information system may include any system relating to information technology. An information system may include, for example, cloud services, on-premise systems, computer programs, etc. The form of an information system may include any known form. For example, cloud services may include AWS (Amazon Web Services), Cycloud, etc.

[0040] The architecture may include any components necessary to achieve the objectives of the information system. The types of components of the information system are not particularly limited and may be appropriately selected depending on the embodiment. Components of the information system may include, for example, software, hardware, network, data, security, etc. Furthermore, the architecture may include the basic structure between the components. The architecture may include, for example, hardware configuration, software configuration, network configuration, data structure, security method, interface design, etc. It may include.

[0041] [User Requirements] User requirement 30 may be any requirement that user Z has for the architecture of the information system. User requirement 30 may include user Z's wishes and requests. User requirement 30 may consist of dynamic information directly related to the architecture. User requirement 30 may include one or more requirements such as, for example, "I want to create a matching app that connects people through shared hobbies," "I want to place a 'like' button on the profile page," "I want to be able to send messages," and "I want to be able to make restaurant reservations."

[0042] User requirements 30 may be entered in any data format, such as text, audio, or images. Audio data may be used as is, or it may be converted to text by speech analysis before being used as user requirements 30. At least a portion of user requirements 30 may be specified in natural language. User requirements 30 are updated by the response 31 to inquiry 49. Updating user requirements 30 may include adding one or more new requirements and modifying one or more existing requirements. New requirements may be requirements not included in user requirements 30 before the update. Existing requirements may be requirements included in user requirements 30 before the update. Modifications to existing requirements may include deletion, change, and replacement. For example, the updated user requirements 30 (30B) may be configured to include the entered response 31 as is.

[0043] [First prompt] The first prompt P1 is a prompt given to the large-scale generation model 2 when generating the proposed architecture 40. The configuration of the first prompt P1 is not particularly limited, and may be determined as appropriate depending on the embodiment, as long as it includes the user requirements 30 and the proposed instructions 61. The first prompt P1 may further include any information other than the user requirements 30 and the proposed instructions 61.

[0044] The configuration of the proposal instruction 61 is not particularly limited and may be determined as appropriate depending on the embodiment, as long as it is possible to instruct the large-scale generative model 2 to propose an architecture. In one example, the proposal instruction 61 may include an instruction statement that instructs the model to propose an architecture that satisfies the given user requirements 30. The instruction statement may be given in natural language, for example, "Please propose an architecture that satisfies the user requirements." The data format of the instruction statement in the proposal instruction 61 is not particularly limited and may be selected as appropriate depending on the embodiment. The generated proposed architecture 40 is the architecture proposed by the large-scale generative model 2.

[0045] [Second prompt] The second prompt P2 is a prompt given to the large-scale generative model 2 when generating the verification result 45 of the proposed architecture 40. The configuration of the second prompt P2 is not particularly limited, as long as it includes the proposed architecture 40 and the verification instructions 62, and may be appropriately determined depending on the embodiment. The second prompt P2 may further include any information other than the proposed architecture 40 and the verification instructions 62.

[0046] The configuration of the verification instruction 62 is not particularly limited and may be determined as appropriate depending on the embodiment, as long as it is possible to instruct the large-scale generative model 2 to verify the proposed architecture 40. In one example, the verification instruction 62 may include an instruction statement that instructs the verification of the proposed architecture 40. The instruction statement may be given in natural language, for example, "Please verify the proposed architecture." The data format of the instruction statement in the verification instruction 62 is not particularly limited and may be selected as appropriate depending on the embodiment.

[0047] The generated verification result 45 is the result of verifying the proposed architecture 40 using the large-scale generation model 2. In one example, the verification result 45 may be output to user Z. The verification result 45 may be output in the same manner as inquiry item 49, etc. Furthermore, the verification items (verification items) may be defined as appropriate depending on the embodiment. At least a part of the verification items may be specified in any way. Alternatively, all of the verification items may be left to the large-scale generation model 2.

[0048] For example, as illustrated in Figure 3, the verification instruction 62 may include a verification list 620 containing a plurality of predefined verification items. The verification items may be defined as appropriate depending on the embodiment. The verification items may include, for example, security, cost (operational costs, etc.), reliability (fault tolerance, etc.), scalability, performance, maintainability, manageability, etc. Maintainability may include CI (Continuous Integration), CD (Continuous Delivery / Continuous Deployment), etc. Manageability may include IAC (Infrastructure as Code), etc. For example, the verification instruction 62 may include the verification list 620 along with an instruction statement that instructs the verification of the proposed architecture 40.

[0049] Accordingly, the deficiencies discovered during verification may be those verification items included in the verification list 620 that were evaluated as having deficiencies during verification by the large-scale generative model 2. According to one example of this embodiment, by defining the verification items in advance, it is possible to guarantee the scope of verification by the large-scale generative model 2.

[0050] The deficiencies may be any matter in the proposed architecture 40 that is deficient or potentially deficient. The level of deficiency is not particularly limited and may be defined as appropriate depending on the embodiment. The deficiencies may include concerns, areas for improvement, points to note, points to be aware of, issues, pending items, etc. If one or more deficiencies exist in the verification results, the generation of inquiry items 49 by the third prompt P3 may be performed. At least a portion of the one or more inquiry items 49 that are generated may relate to at least one of the one or more deficiencies. The one or more deficiencies included in the verification results 45 may include deficiencies other than those included in the verification list 620.

[0051] [Third prompt] The third prompt P3 is a prompt given to the large-scale generation model 2 when generating the inquiry items 49 regarding the deficiencies. The configuration of the third prompt P3 is not particularly limited, and may be determined as appropriate depending on the embodiment, as long as it includes the proposed architecture 40, the verification results 45, and the query generation instructions 63. The third prompt P3 may further include any information other than the proposed architecture 40, the verification results 45, and the query generation instructions 63.

[0052] The configuration of the query generation instruction 63 is not particularly limited and may be determined as appropriate depending on the embodiment, as long as it is possible to instruct the large-scale generation model 2 to generate inquiry items 49 regarding deficiencies in the proposed architecture 40. In one example, the query generation instruction 63 may include an instruction statement (first instruction statement) for generating inquiry items 49 to user Z. The instruction statement (first instruction statement) may be given in natural language, for example, "Please generate an inquiry to the user." The data format of the instruction statement (first instruction statement) is not particularly limited and may be selected as appropriate depending on the embodiment.

[0053] The generated inquiry items 49 may include any items that inquire about to user Z regarding the deficiencies. One or more generated inquiry items 49 may include inquiries about items other than deficiencies. If the verification items are given as a verification list 620, one or more generated inquiry items 49 may include inquiries about items other than the verification items. But that's fine. For example, generating inquiry 49 may additionally involve at least one of the following four options:

[0054] (1) Generating inquiries according to skill level Figure 6 schematically shows an example of a method for generating inquiry items 49 according to the skill level in this embodiment. In one example, the third prompt P3 may further include the user Z's skill level 351 for designing the architecture. The skill level (skill level 351) may be the ability to understand terminology related to the proposed architecture (proposed architecture 40). The method of expressing the skill level is not particularly limited and may be determined as appropriate depending on the embodiment. The skill level may be given as a rank, for example, beginner, intermediate, expert (advanced), or as any other arbitrary numerical indicator. The number of levels in the skill level is not particularly limited and may be defined as appropriate depending on the embodiment. When expressing the skill level as a rank, names such as "beginner" do not have to be given. According to this example, it can be expected that inquiry items 49 corresponding to the user Z's skill level 351 will be generated. Furthermore, by adjusting the content of the inquiry items 49 according to the skill level 351, it can be expected that the possibility of obtaining an appropriate answer 31 will improve.

[0055] (How to acquire skill levels) Skill level 351 may be acquired by any means. Skill level 351 may be given in advance by user Z's own or another person's answers, or it may be obtained by inferring it by any means. User Z's input may reflect user Z's skill level 351. Therefore, it is possible to infer skill level 351 from user Z's input. For example, skill level 351 may be inferred from user Z's input 300.

[0056] Input 300 may include any responses from user Z. For example, input 300 may include at least a portion of user requirements 30. Once the proposed architecture 40 has been updated (architecture re-proposed) one or more times, input 300 may include at least a portion of the response 31 to inquiry 49. Response 31 may include the response 315 indicating that it is not possible to answer, as described later (Figure 7). Input 300 may include at least one of the following: a response evaluating alternatives 451, an input specifying parts to be kept unchanged 318, and an input specifying parts to be changed 319 (Figures 9 and 10, described later).

[0057] Input 300 may include responses other than the above-mentioned responses entered during the architecture design (user requirements 30, answers 31, responses evaluating alternatives 451, input specifying parts 318 not to be changed, input specifying parts 319 to be changed, etc.). For example, the architecture proposal device 1 may output a special question to user Z to measure user Z's skill level 351 and receive input of answers to the special question from user Z. The content of the special question may be determined as appropriate depending on the embodiment. The special question may be output in the same manner as inquiry item 49. Input of the answer may be received in the same manner as user requirements 30, etc. Other responses may include this answer. Input 300 may consist of at least one of the above-mentioned responses and other responses entered during the architecture design.

[0058] According to one example of this embodiment, by automatically inferring the skill level 351 from the user Z's input 300, it is possible to reduce the effort required to get user Z to answer with the skill level 351. Furthermore, in this example, since it does not rely on user Z's subjective answer, the objectivity of the resulting skill level 351 can be ensured. Therefore, it is possible to improve the accuracy of generating inquiry items 49 that are suitable for user Z by using the third prompt P3 which includes this skill level 351.

[0059] (Guessing method) The method for inferring skill level 351 is not particularly limited and may be appropriately selected depending on the embodiment. Skill level 351 may be inferred by any method, such as language analysis, rule-based methods, or the use of a pre-trained model. The pre-trained model may be appropriately generated by machine learning. As shown in Figure 6, in one example, the architecture proposal device 1 may cause the large-scale generative model 2 to infer user Z's skill level 351 by providing a fifth prompt P5 which includes user Z's input 300 and a skill level 351 inference instruction 65.

[0060] The fifth prompt P5 is a prompt given to the large-scale generative model 2 when inferring user Z's skill level 351 from user Z's input 300. The configuration of the fifth prompt P5 is not particularly limited, as long as it includes the input 300 and the inference instruction 65, and may be appropriately determined depending on the embodiment. The fifth prompt P5 may further include any information other than the input 300 and the inference instruction 65.

[0061] The configuration of the estimation instruction 65 is not particularly limited and can be appropriately determined depending on the embodiment, as long as it is possible to instruct the large-scale generative model 2 to estimate the skill level 351. In one example, the estimation instruction 65 may include an instruction statement that instructs to estimate the skill level 351 from the input 300. The instruction statement of the estimation instruction 65 may be given in natural language, for example, "Estimate the user's skill level from the user's input." The data format of the instruction statement of the estimation instruction 65 is not particularly limited and can be appropriately selected depending on the embodiment. The skill level 351 obtained by estimation (estimate result) may be appropriately reflected in the third prompt P3.

[0062] The estimation of skill level 351 by the large-scale generative model 2 may be performed at any time after an arbitrary response is received from user Z. For example, the estimation of skill level 351 may be performed at any time after user requirement 30 has been entered but before the construction of the third prompt P3 is completed. Assuming a stage after the proposed architecture 40 has been updated, the estimation of skill level 351 may be performed at any time after the answer 31 to inquiry 49 has been entered but before the construction of the third prompt P3 is completed. In another example, if the answer to a specific question is used as input 300, the estimation of skill level 351 may be performed in advance before the input of user requirement 30 is received.

[0063] In one example of this embodiment, by using the large-scale generative model 2 to estimate skill level 351, it is not necessary to prepare a dedicated model for estimating skill level 351. Therefore, cost reduction can be expected. In another example, by using the large-scale generative model 2, skill level 351 can be estimated while taking into account a wide range of contexts. Therefore, a high accuracy in the estimated skill level 351 can be expected.

[0064] Furthermore, at least a portion of the process for estimating skill level 351 may be executed on the architecture proposal device 1, or it may be executed on a computer other than the architecture proposal device 1. If the entire process for estimating skill level 351 is performed by another computer, estimating skill level 351 may consist of instructing the other computer to estimate skill level 351 and obtaining the estimation result of skill level 351 from the other computer.

[0065] (2) Correction of inquiries User Z may not necessarily be able to appropriately answer all of the one or more inquiries 49 that are generated. The same applies to the following. For example, the architecture proposal device 1 may accept a response indicating that it cannot answer inquiry 49 (including inquiry 49 generated according to the above skill level 351). If a response indicating that it cannot answer is given, the architecture proposal device 1 may have the large-scale generation model 2 modify inquiry 49.

[0066] Figure 7 schematically shows an example of a method for modifying inquiry 49 in this embodiment. In one example, the architecture proposal device 1 may accept the input of the answer 31 to inquiry 49 in any format. In one example, the input format of the answer 31 may be determined according to inquiry 49, or it may be predetermined on the architecture proposal device 1. In either case, the input format of the answer 31 may be configured to allow the input of a response 315 indicating that it is impossible to answer.

[0067] For example, if the input format of the response 31 is determined according to the inquiry 49, the inquiry generation instruction 63 may include an instruction that allows the input of a response 315 indicating that it is impossible to answer. The instruction that allows for an impossible answer may be incorporated into the first instruction, or it may be given separately from the first instruction. This may generate an inquiry 49 that allows for a response 315 indicating that it is impossible to answer. Also, for example, if the input format of the response 31 is predetermined, an input format that allows for a response 315 indicating that it is impossible to answer may be prepared in advance. The input format may be defined in a program (program 81 in Figure 21), or it may be held as separate data in a memory resource. The proposed architecture device 1 may accept the input of the response 311 in this input format, thereby allowing the input of a response 315 indicating that it is impossible to answer.

[0068] The response 315 indicating that an answer is unavailable may be entered in any way. For example, the response 315 indicating that an answer is unavailable may be entered as a free-form answer such as "I don't know." Alternatively, the response 315 indicating that an answer is unavailable may be entered as a selection-type answer such as a checkbox. Furthermore, the unit in which the response 315 indicating that an answer is unavailable is entered is not particularly limited and may be determined as appropriate depending on the embodiment. For example, the response 315 indicating that an answer is unavailable may be entered for each inquiry item 49, or it may be entered for multiple inquiry items 49 at once.

[0069] If the input to inquiry 49 includes a response 315 indicating that it cannot be answered, the architecture proposal device 1 may cause the large-scale generation model 2 to generate a modified inquiry 49 (49B) by providing a sixth prompt P6 that includes inquiry 49 (49A), user Z's skill level 351 for architecture design, and modification instructions 66. Skill level 351 may be acquired by any method as described above. For the sake of explanation, the inquiry 49 before modification will also be referred to as "inquiry 49A," and the inquiry 49 after modification will also be referred to as "inquiry 49B."

[0070] The sixth prompt P6 is a prompt given to the large-scale generation model 2 when modifying inquiry item 49 to match user Z's skill level 351. The configuration of the sixth prompt P6 is not particularly limited, and may be determined as appropriate depending on the embodiment, as long as it includes inquiry item 49(49A), skill level 351, and modification instruction 66. The sixth prompt P6 may further include any information other than inquiry item 49(49A), skill level 351, and modification instruction 66.

[0071] The configuration of the modification instruction 66 is not particularly limited and may be determined as appropriate depending on the embodiment, as long as it is possible to instruct the large-scale generation model 2 to modify the inquiry item 49 according to the skill level 351. In one example, the modification instruction 66 may include an instruction statement that instructs the modification of the inquiry item 49. The instruction statement of the modification instruction 66 may be, for example, "according to the user's skill level The instructions may be given in natural language, such as "Please correct the inquiry." The data format of the instruction text in the correction instruction 66 is not particularly limited and may be appropriately selected depending on the embodiment.

[0072] The architecture proposal device 1 may output the modified inquiry item 49(49B) to user Z. The architecture proposal device 1 may receive input of the response 31 to the modified inquiry item 49(49B) from user Z. Updating user requirements 30 may consist of updating user requirements 30 with the response 31 input to the modified inquiry item 49(49B). The output of inquiry item 49(49B) and the acceptance of response 31 may be processed in the same way as for the inquiry item 49(49A) before modification. If a response 315 indicating that it is impossible to answer is input for the modified inquiry item 49(49B), the modification of inquiry item 49 may be repeated.

[0073] In one example of this embodiment, when an inquiry item 49 that user Z cannot answer is generated, the possibility of eliciting an appropriate answer 31 from user Z can be obtained by modifying the inquiry item 49 to match user Z's skill level 351. In one example, by allowing a response 315 indicating that the inquiry item is unanswerable, it is possible to suppress the input of a perfunctory answer 31 based on vague knowledge, even though an appropriate answer 31 is difficult to provide. This increases the likelihood that the update of the proposed architecture 40 by the fourth prompt P4 using the input answer 31 will proceed in the correct direction.

[0074] If a response 315 indicating that a part of the original inquiry 49(49A) cannot be answered is entered, and appropriate answers 31 are entered for the remainder, the entered answers 31 may be handled as is. For example, the entered answers 31 may be reflected in the update of user requirement 30. That is, updating user requirement 30 may include updating user requirement 30 with the answers 31 entered for the original inquiry 49(49A). For another example, the answers 31 entered for the original inquiry 49(49A) may be discarded.

[0075] The method for modifying inquiry 49 is not limited to the above example. The configuration of the sixth prompt P6 is not limited to the above example and may be modified as appropriate depending on the embodiment. In another example, the sixth prompt P6 may further include a response 315 indicating that it is not possible to answer. In another example, inquiry 49 other than the inquiry 49 for which a response 315 indicating that it is not possible to answer has been entered may be omitted from the sixth prompt P6. Also, in another example, the sixth prompt P6 may be configured to instruct modification to a simpler expression without specifying a skill level 351. For example, the modification instruction 66 may include an instruction to simplify (lower the level of) inquiry 49. The instruction to simplify inquiry 49 may be incorporated into the instruction to modify inquiry 49, or it may be given separately from the instruction to modify inquiry 49. The wording of the instruction to simplify the text may be determined as appropriate depending on the embodiment, for example, "Please make the expression easier to understand," or "Please reduce the difficulty of the terminology used." In this case, skill level 351 may be omitted from the sixth prompt P6.

[0076] (3) Format control of inquiry items For example, the format of the generated query item 49 may be controlled as appropriate by including an arbitrary instruction statement in the query generation instruction 63. The instruction statement may be configured to instruct the generation of the query item 49 in the target format. The content of the instruction statement may be defined as appropriate depending on the control content. Below are three examples of controlling the format of the query item 49.

[0077] Figure 8 shows a method for controlling the format of the generated query items 49 in this embodiment. An example is schematically shown. In this example, the query generation instruction 63 may include at least one of the following: instruction 631 (second instruction) which generates query items 49 with restricted input formats for the answer 31; instruction 633 (third instruction) which lists the query items 49 in order of priority; and instruction 635 (fourth instruction) which generates query items in a format that asks for user Z's requirements. Each instruction (631, 633, 635) may be incorporated into the first instruction or may be given separately from the first instruction.

[0078] In one example, the above form may be adopted in which the query generation instruction 63 includes at least one of the instruction statements (631, 633, 635) and generates the query item 49 according to the skill level 351. That is, the third prompt P3 may further include the skill level 351. This skill level 351 may be omitted as appropriate. In another example, the above form may be adopted in which, in the query generation instruction 63 includes at least one of the instruction statements (631, 633, 635), the input of a response 315 indicating that the generated query item 49 cannot be answered is allowed, and the query item 49 is modified. In this case, the modification instruction 66 when modifying the query item 49 may also include at least one of the instruction statements (631, 633, 635), similar to the generation instruction 63. In another example, the generation instruction 63 includes at least one of each instruction statement (631, 633, 635), whereas in the modification instruction 66, each instruction statement (631, 633, 635) may be omitted.

[0079] (Input format restrictions) When the generation instruction 63 includes an instruction statement 631, the restrictions on the input format may be determined as appropriate depending on the embodiment. The restricted input format may include, for example, a selection format, a numerical input, etc. The selection format may include single selection, multiple selection, range selection, date selection, etc. In the case of numerical input, a numerical range (upper and lower limits) may or may not be specified. The instruction statement 631 may be given in natural language, for example, "Please generate the inquiry in selection format or numerical input format." The data format of the instruction statement 631 is not particularly limited and may be selected as appropriate depending on the embodiment. According to one example of this embodiment, by including the instruction statement 631 in the generation instruction 63, user Z can obtain an inquiry 49 that is easy to answer 31.

[0080] (Priority) When the generation instruction 63 includes an instruction statement 633, the priority can be set in any way. For example, the priority may be specified by user Z, inferred from user Z's input (input 300, etc.), pre-set in the program, or left to the large-scale generation model 2. The instruction statement 633 may be given in natural language, for example, "Please list them in order of priority." The data format of the instruction statement 633 is not particularly limited and may be appropriately selected depending on the embodiment. According to one example of this embodiment, by including an instruction statement 633 in the generation instruction 63, the generated inquiry items 49 can be arranged in order of priority. This is expected to make it easier for user Z to answer 31.

[0081] Furthermore, when designing an architecture, it is more efficient to design the architecture by starting with aspects that affect the whole system (for example, whether to use servers or serverless architecture) rather than deciding on detailed aspects (for example, how to handle logging). This narrows the scope of the design. For example, priorities can be set according to their impact on the architecture design. In this case, since inquiry items 49 are arranged in order of priority, the answers 31 can be decided starting with the items with the highest impact. This is expected to improve the efficiency of architecture design.

[0082] (Format of asking for a request) When the generation instruction 63 includes instruction statement 635, user Z's request is a wish, request, intention, etc. It may include. The format for asking for a request may be appropriately defined depending on the embodiment. The format for asking for a request may include, for example, question formats such as "Do you want to...?" or "Do you want...?". Instruction 635 may be given in natural language, for example, "Please make the inquiry a question to confirm the user's request." The data format of instruction 635 is not particularly limited and may be appropriately selected depending on the embodiment. In the second experimental example described later, it was found that by giving a prompt instructing the generation of inquiry items in the form of asking for the user's request, there is a high possibility that appropriate inquiry items will be generated by the large-scale generative model. According to one example of this embodiment, based on this finding, it can be expected that the probability of generating appropriate inquiry items 49 will be improved by including instruction 635 in the generation instruction 63.

[0083] (others) The instruction statements included in the query generation instruction 63 are not limited to the above example and may be modified as appropriate depending on the embodiment. In another example, the query generation instruction 63 may include an instruction to avoid asking questions about technical necessity, either in place of or together with at least one of the instruction statements (631, 633, 635). This instruction may be incorporated into the first instruction statement or provided separately from the first instruction statement. The instruction may be provided in natural language, for example, "Please avoid asking questions about technical necessity." The data format of the instruction statement is not particularly limited and may be selected as appropriate depending on the embodiment.

[0084] (4) Reflection of alternatives Figure 9 schematically shows an example of a scenario in which an alternative 451 is generated in this embodiment. In one example, the architecture proposal device 1 may cause the large-scale generation model 2 to generate an alternative 451 of the proposed architecture 40 by providing a seventh prompt P7 which includes the proposed architecture 40, the verification result 45, and a generation instruction 67 for the alternative 451.

[0085] The seventh prompt P7 is a prompt given to the large-scale generation model 2 when generating an alternative 451 of the proposed architecture 40. The configuration of the seventh prompt P7 is not particularly limited, as long as it includes the proposed architecture 40, the verification results 45, and the generation instructions 67, and may be appropriately determined depending on the embodiment. The seventh prompt P7 may further include any information other than the proposed architecture 40, the verification results 45, and the generation instructions 67.

[0086] The configuration of the generation instruction 67 is not particularly limited and may be determined appropriately depending on the embodiment, as long as the large-scale generation model 2 can be instructed to generate an alternative 451 to the proposed architecture 40 in response to the deficiencies found during verification. For example, the generation instruction 67 may include an instruction statement that instructs the generation of an alternative 451 to the proposed architecture 40. The instruction statement of the generation instruction 67 may be given in natural language, for example, "Based on the verification results, please generate an alternative to the proposed architecture." The data format of the instruction statement of the generation instruction 67 is not particularly limited and may be selected appropriately depending on the embodiment.

[0087] The obtained alternative 451 may be used as is. In one example, the obtained alternative 451 may be used to generate query 49. That is, the third prompt P3 may further include the generated alternative 451. In this case, the generation of the alternative 451 may be performed at any time after the verification result 45 is obtained and before the construction of the third prompt P3 is completed. According to one example of this embodiment, the generation of the alternative 451 can facilitate understanding of the proposed architecture 40. By including the generated alternative 451 in the third prompt P3, an improvement in the accuracy of generating query 49 can be expected.

[0088] Note that the use of alternative 451 is not limited to this example. Another example is: Alternative 451 does not have to be used to generate query 49 (construction of the third prompt P3). In this case, the generation of alternative 451 may be performed at any time after the verification result 45 has been obtained.

[0089] Furthermore, the architecture proposal device 1 may output the generated alternative 451 to user Z, similar to the inquiry item 49. The architecture proposal device 1 may also accept input from user Z for an evaluation response to the alternative 451. The method of providing the evaluation response is not particularly limited and may be appropriately selected depending on the embodiment. For example, the evaluation response may be expressed as a binary value, "good" or "bad". The number of evaluation ranks is not limited to binary values ​​and may be appropriately changed depending on the embodiment. Also, for example, the evaluation response may be obtained in free form. In one example, the evaluation response to this alternative 451 may be interpreted as being included in the response 31 to inquiry item 49. That is, the evaluation response to the alternative 451 may be reflected in the update of user requirement 30. As a result, the fourth prompt P4 may further include the evaluation response to the alternative 451. According to one example of this embodiment, the large-scale generative model 2 can be made to generate an updated proposed architecture 40 (re-propose the architecture) after further reflecting the evaluation response of user Z to the alternative 451. Therefore, an improvement in the generation accuracy of the updated proposed architecture 40 can be expected.

[0090] Furthermore, in one example, when adopting a form that uses alternative 451 to generate inquiry 49, the above form of generating inquiry 49 according to the skill level 351 may be further adopted. In this case, the third prompt P3 may further include skill level 351 and alternative 451. This skill level 351 may be omitted as appropriate. In one example, when adopting a form that uses alternative 451 to generate inquiry 49, the inquiry generation instruction 63 may include at least one of the above instruction statements (631, 633, 635). In one example, when adopting a form that uses alternative 451 to generate inquiry 49, the above form may be adopted that allows input of a response 315 indicating that the generated inquiry 49 cannot be answered, and modifies inquiry 49. In this case, the sixth prompt P6 when modifying inquiry 49 may also further include alternative 451, similar to the third prompt P3. In another example, the third prompt P3 further includes alternative 451, whereas in the sixth prompt P6, alternative 451 may be omitted.

[0091] [Fourth prompt] The fourth prompt P4 is a prompt given to the large-scale generative model 2 when updating the proposed architecture 40 (re-proposing the architecture). Updating the proposed architecture 40 may include modifying the proposed architecture 40. The configuration of the fourth prompt P4 is not particularly limited and may be determined as appropriate depending on the embodiment, as long as it includes the updated user requirements 30 (30B), proposed architecture 40 (40A), verification results 45, inquiries 49, and re-proposal instructions 64. The fourth prompt P4 may further include any information other than the updated user requirements 30 (30B), proposed architecture 40 (40A), verification results 45, inquiries 49, and re-proposal instructions 64.

[0092] If it is possible to instruct the large-scale generative model 2 to update the proposed architecture 40 based on user Z's response 31 to inquiry 49, the configuration of the re-proposal instruction 64 is not particularly limited and may be determined as appropriate depending on the embodiment. For example, the re-proposal instruction 64 may include an instruction statement (first instruction statement) that instructs the re-proposal of the architecture. The instruction statement of the re-proposal instruction 64 may be given in natural language, for example, "Please re-proposal the architecture based on the user's response." The data format of the instruction statement of the re-proposal instruction 64 is not particularly limited and may be selected as appropriate depending on the embodiment.

[0093] The generation process by this fourth prompt P4 updates the proposed architecture 40. Updating the proposed architecture 40 may include adding one or more new components and modifying one or more existing components. The new components may be components not included in the proposed architecture 40 (40A) before the update. The existing components may be components that were also included in the proposed architecture 40 (40A) before the update. Modification of existing components may include at least one of deletion, alteration, and replacement.

[0094] In this embodiment, the dialogue from verification to updating of the proposed architecture 40 may be repeated any number of times until no deficiencies remain. This allows for improvement of the proposed architecture 40. Note that the absence of deficiencies may include not only the absence of deficiencies found in the verification results 45, but also the user Z deciding to discontinue updating the proposed architecture 40.

[0095] (Update scope) The scope of updating the proposed architecture 40 is not particularly limited and may be determined as appropriate depending on the embodiment. The scope of updating the proposed architecture 40 may be left to the large-scale generative model 2, or it may be determined by any information processing. Furthermore, the scope of updating the proposed architecture 40 may be specified by user Z.

[0096] Figure 10 schematically shows an example of a method for controlling the scope of updating (re-proposing) the proposed architecture 40 in this embodiment. In one example, the architecture proposal device 1 may output the generated proposed architecture 40 (40A) to user Z. The proposed architecture 40 (40A) may be output in the same manner as inquiry items 49, etc.

[0097] The architecture proposal device 1 may receive input from user Z specifying at least one of the parts 318 to be kept unchanged and the parts 319 to be changed within the proposed architecture 40 (40A). The input specifying at least one of the parts 318 to be kept unchanged and the parts 319 to be changed may be received in the same manner as user requirements 30, etc. The parts 318 to be kept unchanged and the parts 319 to be changed may be appropriately selected from the components of the proposed architecture 40 (40A). The method for specifying each part (318, 319) is not particularly limited and may be appropriately determined depending on the embodiment. The reception of the output of the proposed architecture 40 (40A) and the input specifying each part (318, 319) may be performed at any timing after generating the proposed architecture 40 (40A) but before completing the construction of the fourth prompt P4.

[0098] If an input specifying a portion 318 to be left unchanged is received, the re-proposal instruction 64 may include an instruction statement 648 indicating that the specified portion 318 will not be changed. If an input specifying a portion 319 to be changed is received, the re-proposal instruction 64 may include an instruction statement 649 indicating that the specified portion 319 will be changed. If an input specifying both a portion 318 to be left unchanged and a portion 319 to be changed is received, the re-proposal instruction 64 may include instruction statements 648 and 649.

[0099] The structure of the instruction statement 648 is not particularly limited and may be appropriately determined depending on the embodiment, as long as it is possible to instruct the large-scale generation model 2 not to change the specified part 318. The instruction statement 648 may be given in natural language, for example, "Do not change the part specified by the user." Also, the structure of the instruction statement 649 is not particularly limited and may be appropriately determined depending on the embodiment, as long as it is possible to instruct the large-scale generation model 2 to change the specified part 319. The instruction statement 649 may be given in natural language, for example, "Do not change the part specified by the user." Instructions may be given in natural language, such as "Please change the part that has been changed." Each instruction (648, 649) may be incorporated into the first instruction of the re-proposal instruction 64, or it may be given separately from the first instruction. The data format of each instruction (648, 649) is not particularly limited and may be appropriately selected depending on the embodiment.

[0100] In one example of this embodiment, if a designation of a part 318 that does not need to be changed is received, this designation can be reflected in the re-proposal instruction 64, thereby preventing the designated part 318 from being changed during the architecture re-proposal. Also, if a designation of a part 319 to be changed is received, this designation can be reflected in the re-proposal instruction 64, thereby prompting the change of the designated part 319 during the architecture re-proposal.

[0101] [Generate summary] Figure 11 schematically shows an example of a scenario in which a summary 41 is generated in this embodiment. In one example, the architecture proposal device 1 may cause the large-scale generation model 2 to generate a summary 41 of the proposed architecture 40 by providing an eighth prompt P8 which includes the proposed architecture 40 and a summary instruction 68.

[0102] The eighth prompt P8 is a prompt given to the large-scale generative model 2 when generating a summary of the proposed architecture 40. The configuration of the eighth prompt P8 is not particularly limited, as long as it includes the proposed architecture 40 and the summary instruction 68, and may be appropriately determined depending on the embodiment. The eighth prompt P8 may further include any information other than the proposed architecture 40 and the summary instruction 68.

[0103] The configuration of the summary instruction 68 is not particularly limited and can be appropriately determined depending on the embodiment, as long as it is possible to instruct the large-scale generative model 2 to generate the summary 41. In one example, the summary instruction 68 may include an instruction statement that instructs the generation of a summary 41 of the proposed architecture 40. The instruction statement of the summary instruction 68 may be given in natural language, for example, "Please generate a summary of the proposed architecture." The data format of the instruction statement of the summary instruction 68 is not particularly limited and can be appropriately selected depending on the embodiment. The generation of the summary 41 may be performed at any time after the generation of the proposed architecture 40.

[0104] The generated summary 41 may be used as desired. For example, the architecture proposal device 1 may output the generated summary 41 to user Z. The summary 41 may be output in the same manner as the inquiry items 49, etc. The output of the summary 41 may be performed at any time after the summary 41 has been generated. For example, if a form is adopted in which the proposed architecture 40 is output, the summary 41 may be output together with the proposed architecture 40. In addition, the summary 41 may be output separately from the proposed architecture 40, or it may be output on its own.

[0105] In one example of this embodiment, the output summary 41 makes it easier for user Z to understand the proposed architecture 40. Outputting the summary 41 together with the proposed architecture 40 is particularly effective. The configuration of the eighth prompt P8 may be modified as appropriate depending on the embodiment. In another example, the eighth prompt P8 may further include a skill level 351. In one example of this embodiment, a summary 41 corresponding to the skill level 351 can be generated.

[0106] [Large-scale generative models] The large-scale generative model 2 may be a trained model generated by machine learning using a large amount of training data. The large-scale generative model 2 may include, for example, a large-scale language model, a large-scale visual language model, a large-scale speech model, etc. The large-scale generative model 2 may also include a multimodal model. The configuration and number of parameters of the large-scale generative model 2 depend on the embodiment. The large-scale generative model 2 may be determined as appropriate. For example, a Transformer, a diffusion model, etc. It may have any structure. The large-scale generative model 2 may be fine-tuned according to its intended use. For example, the large-scale generative model 2 may use a known model such as Claude, GPT, or CyberAgentLM.

[0107] The large-scale generation model 2 may be deployed on any computer. In one example, the large-scale generation model 2 may be deployed within the architecture proposal device 1. In this case, the model data of the large-scale generation model 2 may be held in the memory resources of the architecture proposal device 1. The architecture proposal device 1 may construct each prompt P1 to P8 and execute each generation process by providing each constructed prompt P1 to P8 to the large-scale generation model 2. In another example, the large-scale generation model 2 may be deployed on a computer other than the architecture proposal device 1 (such as an external server). In this case, the architecture proposal device 1 may instruct the other computer to execute the generation process based on each prompt P1 to P8 in the large-scale generation model 2. The architecture proposal device 1 causing the large-scale generation model 2 to generate each piece of information (such as the proposed architecture 40) may include the architecture proposal device 1 instructing the other computer to execute the generation process of the large-scale generation model 2. In this case, at least part of the process of constructing each prompt P1 to P8 may be executed on the architecture proposal device 1 or the other computer. The proposed architecture device 1 may individually instruct the generation process based on each prompt P1 to P8, or it may instruct at least two or more generation processes at once.

[0108] The data format of the information provided as each prompt P1 to P8 is not particularly limited and may be appropriately selected depending on the embodiment. For example, each prompt P1 to P8 may consist of data such as text, images, sounds, and tokens. Tokens may be obtained by converting data such as text with a tokenizer. Each prompt P1 to P8 may consist of multiple types of data. Each prompt P1 to P8 may further include any information other than the above-mentioned components.

[0109] [Dialogue History] Figure 12 schematically shows an example of a scenario in which the dialogue history 5 is generated in this embodiment. In one example, the architecture proposal device 1 may associate and save the user Z's input 305, the instructions 600 for the large-scale generation model 2, and the generation result 400 of the large-scale generation model 2 as the dialogue history 5 in chronological order.

[0110] User Z's input 305 may include any of User Z's responses. The configuration of input 305 is not particularly limited and may be determined as appropriate depending on the embodiment, provided that the history of User Z's responses can be verified retrospectively. For example, input 305 may include answers 31 to user requirements 30 and inquiries 49. If a form is adopted that accepts input of answers to specific questions for inferring skill level 351, input 305 may include answers to specific questions. If a form is adopted that accepts evaluations of alternatives 451, input 305 may include responses of evaluations of said alternatives 451. If a form is adopted that accepts input specifying at least one of the parts to be kept unchanged 318 and the parts to be changed 319, input 305 may include the specification (answer) of at least one of the parts to be kept unchanged 318 and the parts to be changed 319.

[0111] Instruction 600 may include any instructions for the large-scale generation model 2. The configuration of instruction 600 is not particularly limited and may be determined as appropriate depending on the embodiment, as long as the history of the generation process of the large-scale generation model 2 can be verified retrospectively. For example, instruction 600 may include each prompt P1 to P8 as is. Instruction 600 may also include each instruction 61 to 68. Instruction 600 simplifies the generation process by each prompt P1 to P8 (each instruction 61 to 68). It may be configured as shown in a simplified manner.

[0112] The generated result 400 may include all results of the generation process by the large-scale generation model 2. The composition of the generated result 400 is not particularly limited and may be determined as appropriate depending on the embodiment, as long as the history of the results of the generation process by the large-scale generation model 2 can be verified retrospectively. For example, the generated result 400 may include the proposed architecture 40, verification results 45, and inquiry items 49. If a form is adopted in which inquiry items 49 are modified, the generated result 400 may include the modified inquiry items 49 (49B). If a form is adopted in which the large-scale generation model 2 is made to infer skill levels 351, the generated result 400 may include the inferred skill levels 351. If a form is adopted in which the large-scale generation model 2 is made to generate alternative proposals 451, the generated result 400 may include alternative proposals 451. If a form is adopted in which the large-scale generation model 2 is made to generate summaries 41, the generated result 400 may include summaries 41.

[0113] The storage area for saving the dialogue history 5 is not particularly limited and may be appropriately selected depending on the embodiment. The dialogue history 5 may be stored in at least one of the memory resources within the proposed architecture device 1, a storage medium, and an external storage device (including the memory resources of an external computer).

[0114] The timing for updating the dialogue history 5 (generating the history for each processing step) is not particularly limited and may be determined as appropriate depending on the embodiment. For example, the dialogue history 5 may be updated for each processing step. That is, each time a process such as user Z's input or generation process is executed, the corresponding history may be added to the dialogue history 5. Alternatively, the dialogue history 5 may be updated to add the history of multiple processing steps at once. The unit for generating the history may be determined as appropriate depending on the embodiment.

[0115] According to one example of this embodiment, the resulting dialogue history 5 can be used to retrospectively verify whether the generation process of the proposed architecture 40 is appropriate. Note that the use of the dialogue history 5 is not limited to verification. The dialogue history 5 may be used in any situation other than verification. For example, at least a portion of the dialogue history 5 may be used to construct each prompt P1 to P8. That is, each prompt P1 to P8 may include at least a portion of the dialogue history 5 obtained up to that point.

[0116] [Internal Information] The method for managing various types of information, such as user Z's input and the generation results of the large-scale generation model 2, is not particularly limited and may be determined as appropriate depending on the embodiment. The various types of information may be internally stored as structured information. For example, the architecture proposal device 1 may manage user requirements 30, proposed architecture 40, verification results 45, and inquiries 49 as structured internal information.

[0117] If a method is adopted to obtain user attributes including skill level 351, the structured internal information may include user attributes. If a method is adopted to generate alternative 451, the structured internal information may include alternative 451. If a method is adopted to generate summary 41, the structured internal information may include summary 41. If a method is adopted to generate dialogue history 5, the structured internal information may include dialogue history 5.

[0118] Figure 13 schematically shows an example of internal information 125 according to this embodiment. In one example, the internal information 125 may include user state 3, architecture state 4, and dialogue history 5.

[0119] User state 3 at the target time may be composed of any information about user Z obtained up to that point in time. For example, user state 3 may be user requirement 30(30A, 3 0B) and user attributes may be included. For example, user requirement 30 may include the answer 31 to inquiry 49. If a format is adopted that accepts evaluation of alternative 451, user requirement 30 may include the response to the evaluation of said alternative 451. If a format is adopted that accepts input specifying at least one of the parts to be changed 318 and the parts to be changed 319, user requirement 30 may include the specification (response) of at least one of the parts to be changed 318 and the parts to be changed 319.

[0120] User attributes may consist of any information about user Z that may be involved in the design of the architecture or interaction with the large-scale generative model 2. User attributes may include user Z's skill level 351. In addition to skill level 351, user attributes may include any constraints imposed by user Z (e.g., hobbies, preferences, usage environment, etc.). Each prompt P1 to P8 may include user attributes. For example, by including user attributes in the first prompt P1, user Z's constraints can be reflected in the generation of the proposed architecture 40. By including user attributes in the fourth prompt P4, user Z's constraints can be reflected in the updating of the proposed architecture 40. By including user attributes in the second prompt P2, user Z's constraints can be reflected in the verification of the proposed architecture 40. Furthermore, by including user attributes in the fourth prompt P4, user Z's constraints can be reflected in the generation of inquiry items 49.

[0121] The architecture state 4 at the target time may consist of any information about the architecture obtained up to that point. For example, architecture state 4 may include the proposed architecture 40 (40A, 40B), verification results 45, and inquiries 49 (49A, 49B). If a form for generating a summary 41 is adopted, architecture state 4 may include the generated summary 41. If a form for generating an alternative 451 is adopted, architecture state 4 (verification results 45) may include the generated alternative 451. The dialogue history 5 at the target time may consist of the history of dialogues obtained up to that point between user Z and the large-scale generation model 2. The dialogue history 5 may include the update history of user state 3 and architecture state 4.

[0122] The format of the structure is not particularly limited and may be appropriately selected depending on the embodiment. For example, a known format such as XML (extensible markup language) may be used as the format of the structure. Furthermore, the storage area for holding the internal information 125 may be arbitrarily selected. The internal information 125 may be held in at least one of the memory resources within the architecture proposed device 1, a storage medium, and an external storage device (including the memory resources of an external computer). Each prompt P1 to P8 may include at least a portion of the internal information 125 obtained up to that point, in addition to the above components.

[0123] (Specific examples of internal information transitions) Figures 14 to 20 show specific examples of the transitions of internal information 125 in this embodiment. Figures 14 to 20 assume an example scenario in which the large-scale generative model 2 is made to estimate the skill level 351 before verification, the summary 41 is generated before verification, and the alternative plan 451 is generated. Internal information X in Figures 14 to 20 is an example of internal information 125. User state A is an example of user state 3. Architecture state B is an example of architecture state 4. Dialogue history C is an example of dialogue history 5. The symbols (0, 1, 2, etc.) after each piece of information (X, A, B, C) represent the transition.

[0124] (A) Initial state ~ Proposed architecture As illustrated in Figure 14, in the initial state, internal information X0 may be empty. That is, user state A0, architecture state B0, and dialogue history C0 may each be empty. In response to receiving input Z1 of user requirement AA1, the architecture proposal device 1 may update internal information X from X0 to X1. User state A is the input user requirement A A0 may be updated to A1 to include A1 (input Z1). The dialogue history C may be updated from C0 to C1 to include the history that user Z provided the system (large-scale generative model 2) with input Z1 of user requirement AA1, and that user state A was updated from A0 to A1.

[0125] A first prompt P1 may be constructed using the obtained user requirement AA1. By providing the constructed first prompt P1 to the large-scale generation model 2, the proposed architecture BA1 may be generated. The generated proposed architecture BA1 may be output to user Z. In response to the generation of the proposed architecture BA1, the architecture proposal device 1 may update the internal information X from X1 to X2. The architecture state B may be updated from B0 to B1 to include the generated proposed architecture BA1. The dialogue history C may be updated from C1 to C2 to include the history that the large-scale generation model 2 executed the process of generating the proposed architecture BA1, that the architecture state B was updated from B0 to B1, and that the proposed architecture BA1 was output.

[0126] (B) Estimation of skill level / generation of summary As illustrated in Figure 15, when the skill level is to be estimated by the large-scale generative model 2, the fifth prompt P5 may be constructed using the obtained user input (input Z1) from the internal information X2. By providing the constructed fifth prompt P5 to the large-scale generative model 2, the skill level ABA1 may be estimated. In response to the estimated skill level ABA1, the architecture proposal device 1 may update the internal information X from X2 to X2_1. The user state A may be updated from A1 to A1_1 to include the estimated skill level ABA1 (user attribute AB1). The dialogue history C may be updated from C2 to C2_1 to include the history that the large-scale generative model 2 performed the skill level ABA1 estimation process and that the user state A was updated from A1 to A1_1.

[0127] Furthermore, when generating summary 41, the eighth prompt P8 may be constructed using the obtained proposed architecture BA1 from the internal information X2 (X2_1). By providing the constructed eighth prompt P8 to the large-scale generation model 2, summary BB1 may be generated. The generated summary BB1 may be output to user Z. In response to the generation of summary BB1, the architecture proposal device 1 may update the internal information X from X2 (X2_1) to X2_2. The architecture state B may be updated from B1 to B1_1 to include the generated summary BB1. The dialogue history C may be updated from C2 (C2_1) to C2_2 to include the history that the large-scale generation model 2 executed the summary BB1 generation process, that the architecture state B was updated from B1 to B1_1, and that summary BB1 was output. Either the estimation of skill level ABA1 or the generation of summary BB1 may be performed first.

[0128] (C) Verification of the proposed architecture As illustrated in Figure 16, a second prompt P2 may be constructed using the proposed architecture BA1 from the internal information X2 (X2_1, X2_2). By providing the constructed second prompt P2 to the large-scale generation model 2, a verification result BC1 of the proposed architecture BA1 may be generated. The generated verification result BC1 may be output to user Z. In response to the generation of the verification result BC1, the architecture proposal device 1 may update the internal information X from X2 (X2_1, X2_2) to X3. The architecture state B may be updated from B1 (B1_1) to B2 to include the generated verification result BC1. The dialogue history C may be updated from C2 (C2_1, C2_2) to C3 to include the history that the large-scale generation model 2 executed the process of generating the verification result BC1, that the architecture state B was updated from B1 (B1_1) to B2, and that the verification result BC1 was output.

[0129] (D) Generation of alternatives As illustrated in Figure 17, the seventh prompt P7 may be constructed using the proposed architecture BA1 and verification result BC1 from internal information X3. By providing the constructed seventh prompt P7 to the large-scale generation model 2, an alternative BC1_1 of the proposed architecture BA1 may be generated. The generated alternative BC1_1 may be output to user Z. In response to the generation of alternative BC1_1, the architecture proposal device 1 may update internal information X from X3 to X3_1. The architecture state B may be updated from B2 to B2_1 to include the generated alternative BC1_1. The dialogue history C may be updated from C3 to C3_1 to include the history that the large-scale generation model 2 executed the generation process for alternative BC1_1, that the architecture state B was updated from B2 to B2_1, and that alternative BC1_1 was output. When input is received regarding the evaluation response for alternative BC1_1, user state A (user requirements) may be updated to include the evaluation response for alternative BC1_1. Dialogue history C may be updated to include the history that user Z provided the evaluation response and that user state A was updated.

[0130] (E) Generating query items As illustrated in Figure 18, the third prompt P3 may be constructed using the proposed architecture BA1 and verification result BC1 from the internal information X3 (X3_1). The third prompt P3 may further include at least one of the user attribute AB1 (skill level ABA1) and alternative BC1_1. By providing the constructed third prompt P3 to the large-scale generation model 2, query item BD1 may be generated. The generated query item 49 may be output to user Z. In response to the generation of query item BD1, the architecture proposal device 1 may update the internal information X from X3 (X3_1) to X4. The architecture state B may be updated from B2 (B2_1) to B3 to include the generated query item BD1. The dialogue history C may be updated from C3(C3_1) to C4 to include the history that the large-scale generation model 2 executed the generation process for query item BD1, that the architecture state B was updated from B2(B2_1) to B3, and that query item BD1 was output.

[0131] (F) Acceptance of responses to inquiries As illustrated in Figure 19, input Z2 of user Z's response 31 may be obtained for the outputted inquiry item BD1. In response to obtaining input Z2 of response 31, the architecture proposal device 1 may update internal information X from X4 to X5. User requirement AA may be updated from AA1 to AA2 to include input Z2 of response 31. Accordingly, user state A may be updated from A1(A1_1) to A2. Dialogue history C may be updated from C4 to C5 to include the history that user Z provided the system (large-scale generation model 2) with input Z2 of response 31 for inquiry item BD1, and that user state A was updated from A1(A1_1) to A2.

[0132] If input is received specifying at least one of the parts 318 to be kept unchanged and the parts 319 to be changed for the proposed architecture BA1, user state A (user requirement AA) may be updated to include the specification (response) of at least one of the parts 318 to be kept unchanged and the parts 319 to be changed. Dialogue history C may be updated to include the history that user Z provided a specification of at least one of the parts 318 to be kept unchanged and the parts 319 to be changed, and that user state A was updated.

[0133] Furthermore, if the input Z2 of response 31 includes a response 315 indicating that it is impossible to answer, a sixth prompt P6 may be constructed using inquiry item BD1. By providing the sixth prompt P6 to the large-scale generation model 2, a modified inquiry item may be generated. The modified inquiry item may be output to user Z. Input of a response to the modified inquiry item may be accepted. Accordingly, the architecture proposal device 1 internal Department information X may be updated. User state A (user requirement AA) may be updated to include the answer to the modified inquiry. Architecture state B may be updated to include the modified inquiry. Dialogue history C may be updated to include the history that the large-scale generation model 2 executed the generation process for the modified inquiry, that architecture state B was updated, that the modified inquiry was output to user Z, that user Z provided input for the answer to the modified inquiry, and that the user state was updated.

[0134] (G) Re-proposing the architecture As illustrated in Figure 20, the fourth prompt P4 may be constructed using user requirements AA2, proposed architecture BA1, verification results BC1, and inquiry items BD1 from internal information X5. By providing the constructed fourth prompt P4 to the large-scale generation model 2, the updated proposed architecture BA2 may be generated. That is, the proposed architecture BA may be updated from BA1 to BA2. The updated proposed architecture BA2 may be output to user Z. In response to obtaining the updated proposed architecture BA2, the architecture proposal device 1 may update internal information X from X5 to X6. The architecture state B may be updated from B3 to B4 to include the updated proposed architecture BA2. The dialogue history C may be updated from C5 to C6 to include the history that the large-scale generation model 2 executed the process of generating the updated proposed architecture BA2, that the architecture state B was updated from B3 to B4, and that the updated proposed architecture BA2 was output. From this point onward, until the architecture re-proposal is stopped, the architecture proposal device 1 may update internal information X as appropriate.

[0135] In one example of this embodiment, as shown in the specific example above, the internal information 125 is structured, making it easy to construct each prompt P1 to P8 by extracting information from the target field. Therefore, it is expected that the implementation of the system (architecture proposal device) will be simplified. In one example, the structured internal information 125 is also expected to simplify post-verification.

[0136] §2 Example Configuration [Hardware configuration] Figure 21 schematically shows an example of the hardware configuration of the architecture proposed device 1 according to this embodiment. In one example, the architecture proposed device 1 may be configured as a computer in which a control unit 11, a storage unit 12, and a communication module 13 are electrically connected.

[0137] The control unit 11 may include a hardware processor such as a CPU, RAM (Random Access Memory), and ROM (Read Only Memory), and is configured to perform information processing based on a program and various data. The control unit 11 (CPU) is an example of a processor resource.

[0138] The storage unit 12 may include, for example, a hard disk drive, a solid-state drive, or a semiconductor memory, and is configured to hold arbitrary data. The storage unit 12, RAM, and ROM are examples of memory resources of the architecture proposal device 1. In this example embodiment, the storage unit 12 may store various information such as program 81 and internal information 125. Program 81 is a program that causes the architecture proposal device 1 to execute information processing related to the architecture proposal (Figures 23 and 24 described later). Program 81 includes a series of instructions for said information processing. Program 81 is an example of the program (architecture proposal program) of this disclosure.

[0139] In one example, at least one of the program 81 and the internal information 125 is stored in the storage unit 12. Alternatively, or together with the storage unit 12, the information may be stored in the storage medium 91. The storage medium 91 is configured to store various types of information (stored programs, etc.) by electrical, magnetic, optical, mechanical, or chemical means so that a machine such as a computer can read the information. The storage unit 12 and the storage medium 91 are examples of non-temporary storage media. The proposed architecture device 1 may obtain the program 81 from the storage medium 91. The storage medium 91 may be a disk-type storage medium (CD, DVD, etc.) or a non-disk-type storage medium such as semiconductor memory (flash memory, etc.). Any drive device may be used to read the information stored in the storage medium 91. The type of drive device may be selected according to the storage medium 91. The drive device may be connected to the proposed architecture device 1 by any method. The storage medium 91 may include an external storage device.

[0140] The communication module 13 is configured to perform wired or wireless communication over a network. The communication module 13 may consist of, for example, a wired LAN (Local Area Network) module, a wireless LAN module, etc. The network standard is not particularly limited and may be appropriately selected depending on the embodiment. For example, the type of network may be appropriately selected from the Internet, wireless communication network, mobile communication network, telephone network, dedicated network, etc. The proposed architecture device 1 may perform data communication with other computers (for example, a user terminal U1, an external computer, etc.) via the communication module 13.

[0141] Regarding the specific hardware configuration of the proposed architecture device 1, components can be omitted, replaced, and added as appropriate, depending on the embodiment. For example, the control unit 11 may include multiple hardware processors. The hardware processors may consist of a microprocessor, FPGA (field-programmable gate array), DSP (digital signal processor), GPU (Graphics Processing Unit), ASIC (application specific integrated circuit), etc. The proposed architecture device 1 may further include an input device and an output device. The input device is configured to accept information input. The input device may consist of, for example, a mouse, keyboard, or control. The output device is configured to output information. The output device may consist of, for example, a display or speaker. The input device and the output device may be integrated at least in part by a touch panel display or the like. The input device and the output device may be connected to the proposed architecture device 1 as appropriate. The communication module 13 may be omitted. At least one of the program 81 and the internal information 125 may be stored in an external storage device such as a NAS (Network Attached Storage). External storage devices are also an example of non-temporary storage media. The proposed architecture device 1 may consist of multiple computers. In this case, the hardware configurations of each computer may or may not be identical. The proposed architecture device 1 may consist of computers designed specifically for the services provided, as well as general-purpose server devices, general-purpose PCs, notebook PCs, terminal devices, etc. Terminal devices may include smartphones, tablet devices, etc.

[0142] [Software Configuration] Figure 22 schematically shows an example of the software configuration of the architecture proposal device 1 according to this embodiment. The control unit 11 of the architecture proposal device 1 executes instructions contained in the program 81 stored in the storage unit 12 using the CPU. As a result, the architecture proposal device 1 operates as a computer equipped with the following software modules: requirement receiving unit 1101, first request unit 1102, second request unit 1103, third request unit 1104, output processing unit 1105, response receiving unit 1106, specification receiving unit 1107, fourth request unit 1108, level estimation unit 1109, sixth request unit 1110, seventh request unit 1111, eighth request unit 1112, and history generation unit 1113. In other words, in this example, each software module of the architecture proposal device 1 is implemented by the control unit 11 (CPU).

[0143] The requirements receiving unit 1101 is configured to receive user requirements 30 for the architecture of the information system from user Z. The first request unit 1102 is configured to cause the large-scale generation model 2 to generate a proposed architecture 40 for the information system by providing a first prompt P1 that includes the input user requirements 30 and a proposal instruction 61. The second request unit 1103 is configured to cause the large-scale generation model 2 to generate a verification result 45 of the proposed architecture 40 by providing a second prompt P2 that includes the generated proposed architecture 40 and a verification instruction 62. The third request unit 1104 is configured to cause the large-scale generation model 2 to generate an inquiry item 49 by providing a third prompt P3 that includes the proposed architecture 40, the verification result 45, and a query generation instruction 63 if there are any deficiencies.

[0144] The output processing unit 1105 is configured to output the generated inquiry items 49 to user Z. The response receiving unit 1106 is configured to receive the response 31 to the inquiry items 49 from user Z and update the user requirements 30 based on the input response 31. The fourth request unit 1108 is configured to cause the large-scale generation model 2 to generate the updated proposed architecture 40 by providing a fourth prompt P4 that includes the updated user requirements 30, proposed architecture 40, verification results 45, inquiry items 49, and re-proposal instruction 64.

[0145] The generation of verification results 45 by the second request unit 1103, the generation of inquiry items 49 by the third request unit 1104, the output processing unit 1105, the acceptance of input of answer 31 by the answer receiving unit 1106, the updating of user requirements 30 by the answer receiving unit 1106, and the updating of proposed architecture 40 by the fourth request unit 1108 may be repeated until no deficiencies remain. This may improve the proposed architecture 40. The output processing unit 1105 is configured to output the improved proposed architecture 40.

[0146] The output processing unit 1105 may be further configured to output the generated proposed architecture 40 to user Z. The designation receiving unit 1107 is configured to receive input from user Z specifying at least one of the parts 318 to be kept unchanged and the parts 319 to be changed in the proposed architecture 40. The fourth request unit 1108 may be configured to cause the large-scale generation model 2 to generate an updated proposed architecture 40 using a re-proposal instruction 64 which includes at least one of an instruction statement 648 indicating that the specified part 318 will not be changed and an instruction statement 649 indicating that the specified part 319 will be changed.

[0147] The level estimation unit 1109 is configured to estimate user Z's skill level 351 from user Z's input 300. The level estimation unit 1109 may include a fifth request unit 11091. The fifth request unit 11091 is configured to cause the large-scale generative model 2 to estimate user Z's skill level 351 by providing a fifth prompt P5 which includes user Z's input 300 and an instruction 65 to estimate the skill level 351.

[0148] The sixth request unit 1110 is configured to cause the large-scale generation model 2 to generate a modified inquiry item 49 by providing a sixth prompt P6 including the inquiry item 49, skill level 351, and modification instruction 66, if the input for the inquiry item 49 includes a response 315 indicating that it cannot be answered. The output processing unit 1105 may be configured to output the modified inquiry item 49 to user Z. The response receiving unit 1106 may be configured to receive the input of the response 31 for the modified inquiry item 49 from user Z and update the user requirements 30 with the input response 31.

[0149] The seventh request unit 1111 provides a seventh prompt P7 which includes a generation instruction 67 for the proposed architecture 40, verification results 45, and alternative 451, thereby generating an alternative for the proposed architecture 40. The configuration is such that 451 is generated by the large-scale generation model 2. The third request unit 1104 may be configured to generate the inquiry item 49 by using the third prompt P3 which further includes the generated alternative 451, thereby causing the large-scale generation model 2 to generate the inquiry item 49.

[0150] The eighth request unit 1112 is configured to cause the large-scale generation model 2 to generate a summary 41 of the proposed architecture 40 by providing an eighth prompt P8 which includes the proposed architecture 40 and a summary instruction 68. The output processing unit 1105 may be further configured to output the generated summary 41 to user Z. The history generation unit 1113 is configured to associate and save user Z's input 305, the instruction 600 to the large-scale generation model 2, and the generation result 400 of the large-scale generation model 2 as a dialogue history 5 in chronological order.

[0151] §3 Example of Operation Figures 23 and 24 are flowcharts illustrating an example of the processing procedure for proposing an architecture by the architecture proposal device 1 according to this embodiment. The following processing procedure is an example of the information processing method (architecture proposal method) of the present disclosure. However, the following processing procedure is merely an example, and each step may be modified as much as possible. Furthermore, depending on the embodiment, steps in the following processing procedure can be omitted, replaced, and added as appropriate.

[0152] (Step S101) In step S101, the control unit 11 operates as a requirements receiving unit 1101 and receives user requirements 30 for the architecture of the information system from user Z. In one example, the control unit 11 may receive the user requirements 30 via a user terminal U1 connected to the architecture proposal device 1. Upon receiving the user requirements 30, the control unit 11 proceeds to the next step S102.

[0153] (Step S102) In step S102, the control unit 11 acts as the first request unit 1102 and provides the large-scale generation model 2 with a first prompt P1 containing the input user requirements 30 and proposal instructions 61. This causes the control unit 11 to generate the proposed architecture 40 in the large-scale generation model 2. Once the proposed architecture 40 is generated, the control unit 11 proceeds to the next step S103.

[0154] (Step S103) In step S103, the control unit 11 operates as a level estimation unit 1109 and estimates the user Z's skill level 351 from the user Z's input 300. The method for estimating the skill level 351 may be appropriately selected depending on the embodiment. In one example, the control unit 11 may operate as a fifth request unit 11091 and provide the large-scale generation model 2 with a fifth prompt P5 including the user Z's input 300 and a skill level estimation instruction 65. This allows the control unit 11 to cause the large-scale generation model 2 to estimate the user Z's skill level 351. Note that the timing of executing the process in step S103 is not limited to this example. The process in step S103 may be executed at any timing. Also, the control unit 11 may execute the process in step S103 after the user Z's input 300 has accumulated to a predetermined amount. Once the skill level 351 is estimated, the control unit 11 proceeds to the next step S104.

[0155] (Step S104) In step S104, the control unit 11 acts as the eighth request unit 1112 and provides the large-scale generation model 2 with the eighth prompt P8, which includes the proposed architecture 40 and the summary instruction 68. This causes the control unit 11 to cause the large-scale generation model 2 to generate a summary 41 of the proposed architecture 40. The timing of executing the process in step S104 is as follows (in this example): It does not have to be limited. The processing in step S104 may be performed at any timing after the processing in step S102 has been performed. Once summary 41 is generated, the control unit 11 proceeds to the next step S105.

[0156] (Step S105) In step S105, the control unit 11 acts as a second request unit 1103 and provides the large-scale generation model 2 with a second prompt P2 including the generated proposed architecture 40 and verification instructions 62. As a result, the control unit 11 causes the large-scale generation model 2 to generate verification results 45 of the proposed architecture 40.

[0157] In one example, the verification instruction 62 may include a verification list 620 containing a predefined set of verification items. This ensures that if deficiencies are found during verification, at least some of the discovered deficiencies may be among the verification items included in the verification list 620 that were evaluated as deficiencies in the verification by the large-scale generation model 2. In another example, the control unit 11 may operate as an output processing unit 1105 at any timing after executing the processing in step S105 and output the generated verification result 45 to user Z. Once the verification result 45 is generated, the control unit 11 proceeds to the next step S106.

[0158] (Step S106) In step S106, the control unit 11 operates as an output processing unit 1105 and outputs the generated proposed architecture 40 and summary 41 to user Z. Note that the timing of outputting the proposed architecture 40 and summary 41 is not limited to this example. The proposed architecture 40 may be output at any time after the processing in step S102 is executed. The summary 41 may be output at any time after the processing in step S104 is executed. The proposed architecture 40 and summary 41 may be output separately. After outputting the proposed architecture 40 and summary 41, the control unit 11 proceeds to the next step S107.

[0159] (Step S107) In step S107, the control unit 11 determines the branch destination of the process based on whether or not there is one or more defects in the verification result 45. If there is one or more defects, the control unit 11 proceeds to the next step S108. On the other hand, if there are no defects, the control unit 11 terminates the processing procedure related to this example of operation.

[0160] In one example, the control unit 11 may determine whether or not there is one or more defects according to the generated verification result 45. The control unit 11 may determine the branch destination of the process according to the determination result. In another example, the control unit 11 may accept an operation from user Z (user terminal U1) to determine whether or not there is one or more defects. The control unit 11 may determine the branch destination of the process according to the result of user Z's operation. In response to user Z performing an operation to request an update of the proposed architecture 40, the control unit 11 may determine that there is one or more defects and proceed to the next step S108. On the other hand, in response to user Z performing an operation to terminate the update of the proposed architecture 40, the control unit 11 may determine that there are no defects and terminate the processing procedure related to this example of operation. The operation to request an update and the operation to terminate an update are not particularly limited and may be determined as appropriate according to the embodiment.

[0161] (Step S108) In step S108, the control unit 11 acts as the seventh request unit 1111 and provides the large-scale generation model 2 with a seventh prompt P7 which includes a generation instruction 67 for the proposed architecture 40, verification results 45, and alternative 451. As a result, the control unit 11 causes the large-scale generation model 2 to generate an alternative 451 for the proposed architecture 40.

[0162] In one example, the control unit 11 may operate as an output processing unit 1105 at any time after executing the process in step S108 and output the generated alternative 451 to user Z. The control unit 11 may also operate as a response receiving unit 1106 and receive input of an evaluation response to the alternative 451 from user Z, and update the user requirement 30 based on the input evaluation response. Note that the timing of executing the process in step S108 is not limited to this example. The process in step S108 may be executed at any time after executing the process in step S105. Once the alternative 451 is generated, the control unit 11 proceeds to the next step S109.

[0163] (Step S109) In step S109, the control unit 11 acts as a third request unit 1104 and provides the large-scale generation model 2 with a third prompt P3 including the proposed architecture 40, verification results 45, and a query generation instruction 63. As a result, the control unit 11 causes the large-scale generation model 2 to generate the query item 49.

[0164] In one example, the constructed third prompt P3 may further include the user Z's skill level 351. In another example, the third prompt P3 may further include an alternative 451. Also, in another example, the query generation instruction 63 may include at least one of the following: instruction 631 which generates a query item 49 with restricted input format for the answer 31; instruction 633 which lists the query items 49 in order of priority; and instruction 635 which generates a query item in a format that asks for the user Z's request. Once the query item 49 is generated, the control unit 11 proceeds to the next step S110.

[0165] (Step S110 - Step S111) In step S110, the control unit 11 operates as an output processing unit 1105 and outputs the generated inquiry item 49 to user Z. In one example, the control unit 11 may output the inquiry item 49 to user terminal U1. After outputting the inquiry item 49, the control unit 11 proceeds to the next step S111.

[0166] In step S111, the control unit 11 operates as a response receiving unit 1106 and receives the response 31 to the inquiry 49 from user Z. In one example, the control unit 11 may receive the response 31 via the user terminal U1. Upon receiving the response 31, the control unit 11 proceeds to the next step S112.

[0167] (Step S112) In step S112, the control unit 11 determines the branch destination of the process depending on whether the input response 31 to the inquiry 49 includes a response 315 indicating that it is impossible to answer. If the response 315 indicating that it is impossible to answer is included in the response 311, the control unit 11 proceeds to step S113. On the other hand, if the response 315 indicating that it is impossible to answer is not included in the response 311, the control unit 11 proceeds to step S114.

[0168] (Step S113) In step S113, the control unit 11 acts as the sixth request unit 1110 and provides the large-scale generation model 2 with a sixth prompt P6 including the inquiry item 49, skill level 351, and modification instruction 66. This causes the control unit 11 to cause the large-scale generation model 2 to generate the modified inquiry item 49. Once the modified inquiry item 49 is generated, the control unit 11 returns processing to step S110.

[0169] In step S110, the control unit 11 operates as an output processing unit 1105 and corrects The inquiry item 49 is output to user Z. In step S111, the control unit 11 receives the input of the response 31 to the modified inquiry item 49 from user Z. The control unit 11 may repeatedly execute the processes in steps S113, S110, and S111 until there are no more responses 315 indicating that an answer is not possible.

[0170] (Step S114) In step S114, the control unit 11 operates as a response receiving unit 1106 and updates the user requirements 30 based on the input response 31. If the processing in step S113 is being executed, the control unit 11 may update the user requirements 30 based on the response 31 to the modified inquiry 49. Once the user requirements 30 are updated, the control unit 11 proceeds to the next step S115.

[0171] (Step S115) In step S115, the control unit 11 operates as a designation receiving unit 1107 and receives input from user Z specifying at least one of the parts 318 to be kept unchanged and the parts 319 to be changed in the proposed architecture 40. In one example, the control unit 11 may receive the designation of at least one of the parts 318 to be kept unchanged and the parts 319 to be changed via user terminal U1. The control unit 11 updates the user requirement 30 based on the designation (response) of at least one of the parts 318 to be kept unchanged and the parts 319 to be changed in the proposed architecture 40. Note that the timing of executing the processes in steps S106 and S115 is not limited to this example. The process in step S115 may be executed at any time after the proposed architecture 40 has been output. The processes in steps S106 and S115 may be executed before the processes in steps S110 and S111. The processes in steps S106 and S115 may be executed at least partially in parallel with the processes in steps S110 and S111. When user requirement 30 is updated, the control unit 11 proceeds to the next step S116.

[0172] (Step S116) In step S116, the control unit 11 acts as a fourth request unit 1108 and provides the large-scale generation model 2 with a fourth prompt P4 containing the updated user requirements 30, proposed architecture 40, verification results 45, inquiry items 49, and a re-proposal instruction 64. This causes the control unit 11 to generate the updated proposed architecture 40 in the large-scale generation model 2. In one example, the re-proposal instruction 64 may include at least one of an instruction statement 648 indicating that the specified part 318 will not be changed, and an instruction statement 649 indicating that the specified part 319 will be changed. Once the updated proposed architecture 40 is generated, the control unit 11 returns to step S103 and executes the process again from step S103.

[0173] The control unit 11 may repeatedly execute the processes from steps S103 to S116 until it is determined that there are no deficiencies in the process of step S107. This allows the control unit 11 to improve the proposed architecture 40. The control unit 11 may output the improved proposed architecture 40 through the process of step S106. The control unit 11 may, at any time, operate as a history generation unit 1113 and save the user Z's input 305, the instructions 600 for the large-scale generation model 2, and the generation results 400 of the large-scale generation model 2 as a dialogue history 5, associated in chronological order. The control unit 11 may generate a history for each step, or it may generate the history of multiple steps at once. Furthermore, the control unit 11 may manage the various information obtained in each step (user requirements 30, proposed architecture 40, verification results 45, inquiries 49, etc.) as structured internal information 125.

[0174] [Features] In this embodiment, the proposed architecture device 1 is used between user Z and the large-scale generative model 2. As an alternating dialogue, the process of proposing an architecture (steps S102 and S116), verifying it (step S105), and querying user Z (steps S101 and S111) is repeatedly executed. In one example, verification and querying are automatically repeated. In this dialogue, the repeated responses 31 to the query items 49 facilitate the refinement of the user requirements 30. By refining the user requirements 30, the proposed architecture 40 obtained in the process of step S116 can be polished. As a result, according to this embodiment, it can be expected that an appropriate architecture (proposed architecture 40) will be generated.

[0175] §4 Variant While embodiments of this disclosure have been described in detail above, the above description is merely illustrative in all respects. The processes and means described in this disclosure can be freely combined and implemented as long as no technical inconsistencies arise. Furthermore, various improvements or modifications may be made to the above embodiments as appropriate. For example, the following modifications are possible. In the following, the same reference numerals are used for components similar to those in the above embodiments, and explanations of points similar to those in the above embodiments have been omitted as appropriate.

[0176] Regarding the processing procedure of the architecture proposal device 1 according to the above embodiment, at least one of the following steps may be omitted, replaced, or added. For example, the process of generating the dialogue history 5 may be omitted. Accordingly, the history generation unit 1113 may be omitted from the software configuration of the architecture proposal device 1. Also, for example, the processing of step S104 and the output of the summary 41 may be omitted. Accordingly, the eighth request unit 1112 may be omitted from the software configuration of the architecture proposal device 1. Also, for example, the processing of step S108 may be omitted. Accordingly, the seventh request unit 1111 may be omitted from the software configuration of the architecture proposal device 1. Also, for example, the processing of steps S112 and S113 may be omitted. Accordingly, the sixth request unit 1110 may be omitted from the software configuration of the architecture proposal device 1. Also, for example, the processing of step S106 may be omitted. Also, for example, the processing of step S103 may be omitted. Skill level 351 may be given in advance. Skill level 351 may be omitted. Accordingly, the level estimation unit 1109 (fifth request unit 11091) may be omitted from the software configuration of the architecture proposal device 1. Also, for example, the processing in step S115 may be omitted. Accordingly, the designation reception unit 1107 may be omitted from the software configuration of the architecture proposal device 1.

[0177] §5 Experimental Examples To verify the effectiveness of the above embodiments, the following experiments were conducted. However, the present invention is not limited to the following experimental examples.

[0178] [First Experimental Example] First, as the first experimental example, we verified that an appropriate architecture could be obtained by repeatedly proposing, verifying, querying, and re-proposing an architecture using a large-scale generative model in the same manner as the embodiment described above.

[0179] Specifically, the large-scale generative model used was GPT-4o (gpt-4o-2024-08-06). Early users The requirements given were: "I want to create a matching app on AWS that connects people based on shared hobbies. It should allow users to quickly and easily view profiles and send 'likes,' and send messages once a match is made. I also want it to be able to make restaurant reservations." Using these user requirements, a first prompt was constructed. By feeding the constructed first prompt into a large-scale generative model, the proposed architecture was generated (the initially proposed architecture was obtained).

[0180] A second prompt was constructed using the generated proposed architecture. Validation results were generated by feeding the constructed second prompt to a large-scale generative model. A third prompt was constructed using the obtained proposed architecture and validation results. Inquiries were generated by feeding the constructed third prompt to a large-scale generative model. Answers to the generated inquiries were input. A fourth prompt was constructed using the user requirements, proposed architecture, validation results, and inquiries updated with the input answers. The proposed architecture was updated by feeding the constructed fourth prompt to a large-scale generative model. That is, the second proposed architecture (the result of the first re-proposal of the architecture) was obtained.

[0181] The second prompt was reconstructed using the updated proposed architecture. The validation results were regenerated by feeding the reconstructed second prompt to the large-scale generative model. The third prompt was reconstructed using the updated proposed architecture and validation results. The queries were regenerated by feeding the reconstructed third prompt to the large-scale generative model. The answers to the generated queries were input. The fourth prompt was reconstructed using the updated user requirements, proposed architecture, validation results, and queries based on the input answers. The proposed architecture was updated by feeding the reconstructed fourth prompt to the large-scale generative model. That is, the architecture proposed for the third time (the result of the second re-proposal of the architecture) was obtained.

[0182] Figures 25 and 26 show the architecture initially proposed in the first experimental example. Figure 27 shows the verification results of the architecture initially proposed in the first experimental example. Figure 28 shows the inquiries and answers regarding the architecture initially proposed in the first experimental example. Figures 29 and 30 show the architecture proposed second in the first experimental example (results of the first revision of the architecture). Figure 31 shows the verification results of the architecture proposed second in the first experimental example. Figure 32 shows the inquiries and answers regarding the architecture proposed second in the first experimental example. Figures 33 and 34 show the architecture proposed third in the first experimental example (results of the second revision of the architecture). Comparing the initially proposed architecture (Figures 25 and 26) with the third proposed architecture (Figures 33 and 34), it was found that a suitable architecture could be obtained by refining the user requirements. From these results, it was verified that it is possible to expect the generation of a suitable architecture according to this embodiment.

[0183] [Second Experimental Example] Next, we verified that appropriate queries could be obtained by including instructions that generate queries in the form of asking for user requests in the generation instructions. Similar to the first experimental example, we used GPT-4o (gpt-4o-2024-08-06) as the large-scale generation model. Similar to the first experimental example, the initial user By providing the requirements, the process proceeded to the generation of the query. In the 2-1 experiment, the large-scale generative model was given a prompt that did not include instructions to generate the query in the form of asking for the user's requirements. On the other hand, in the 2-2 experiment, the large-scale generative model was given a prompt that included instructions to generate the query in the form of asking for the user's requirements.

[0184] Figure 35 shows the instruction given to the large-scale generation model as an instruction to generate inquiries in the form of asking for user requests. Figure 36 shows the results of generating inquiries obtained by giving a prompt that does not include an instruction to generate inquiries in the form of asking for user requests (Experimental Example 2-1). Figure 37 shows the results of generating inquiries obtained by giving a prompt that includes an instruction to generate inquiries in the form of asking for user requests (Experimental Example 2-2). When the prompt does not include an instruction to generate inquiries in the form of asking for user requests, the inquiry "Is the availability of the cloud service important?" is generated. The following was generated. In contrast, when the prompt included an instruction to generate a query in the form of asking for the user's request, the query "Do you want to improve the availability of the application?" was generated. Including other queries, the queries generated when the prompt included an instruction to generate a query in the form of asking for the user's request were more appropriate than when it did not include such an instruction. From this result, it was found that including an instruction to generate a query in the form of asking for the user's request in the generation instruction can be expected to improve the likelihood of generating appropriate queries. [Explanation of Symbols]

[0185] 1…Architecture proposal device, 2…Large-scale generative model, 30...User requirements, 61...Proposal instructions, 40...Proposed architecture, 62...Verification instructions, 45...Verification results, 63...Generation instructions, 49... Inquiries, 31...Answer, 64...Reproposal instruction, P1...First prompt, P2...Second prompt, P3...3rd prompt, P4...4th prompt, Z...User

Claims

1. Accepting user requirements for the architecture of information systems from users, By providing a first prompt that includes the input user requirements and proposed instructions, the proposed architecture of the information system is generated in a large-scale generative model. By providing a second prompt including the generated proposed architecture and verification instructions, the large-scale generative model is made to generate verification results for the proposed architecture. If any deficiencies exist in the generated verification results, a third prompt is given, including the proposed architecture, the verification results, and a query generation instruction, causing the large-scale generation model to generate queries regarding the deficiencies. Output the generated inquiry items to the user. The user provides input of answers to the aforementioned inquiries. The user requirements will be updated based on the input response. By providing a fourth prompt including the updated user requirements, the proposed architecture, the verification results, the inquiries, and the resubmission instructions, the updated proposed architecture is generated in the large-scale generative model. The proposed architecture is improved by repeatedly generating the verification results of the proposed architecture, generating the inquiries, outputting the inquiries, accepting input of answers to the inquiries, updating the user requirements, and generating the updated proposed architecture until no deficiencies remain, and To output the improved proposed architecture, A control unit configured to perform the following: Architecture proposal device.

2. The aforementioned verification instruction includes a verification list containing a predefined number of verification items, The aforementioned deficiencies are those verification items included in the verification list that were evaluated as having deficiencies in the verification by the large-scale generative model. The architecture proposed in the claim 1.

3. The third prompt further includes the user's skill level for designing the architecture, The architecture proposed in the claim 1.

4. The user's skill level is inferred from the user's input. The architecture proposed in the claim 3.

5. The control unit is further configured to cause the large-scale generative model to estimate the user's skill level by providing a fifth prompt including user input and an instruction to estimate the skill level. The architecture proposed in claim 4.

6. The control unit, If the input to the aforementioned inquiry includes a response indicating that it cannot be answered, a sixth prompt is given, including the inquiry, the user's skill level for designing the architecture, and correction instructions, thereby causing the large-scale generative model to generate a corrected inquiry. Outputting the corrected inquiry to the user, and The user provides the corrected response to the aforementioned inquiry. 、 It is configured to perform further actions, Updating the user requirements consists of updating the user requirements with the answers entered to the modified inquiries. The architecture proposed in the claim 1.

7. The control unit is further configured to associate and save the user's input, the instructions for the large-scale generative model, and the generation results of the large-scale generative model as a dialogue history in chronological order. The architecture proposed in the claim 1.

8. The control unit, Outputting the generated proposed architecture to the user, and The system accepts input from the user specifying the parts of the proposed architecture that will not be changed. It is configured to perform further actions, The aforementioned revised proposal instructions include a statement indicating that the specified portion shall not be changed. The architecture proposed in the claim 1.

9. The control unit, Outputting the generated proposed architecture to the user, and The system accepts input from the user specifying the parts to be changed within the proposed architecture. It is configured to perform further actions, The aforementioned revised proposal instructions include instructions indicating that the specified portion will be changed. The architecture proposed in the claim 1.

10. The control unit is further configured to cause the large-scale generative model to generate alternatives to the proposed architecture by providing a seventh prompt including the proposed architecture, the verification results, and an instruction to generate alternatives. The third prompt further includes the generated alternative, The architecture proposed in the claim 1.

11. The control unit, By providing an eighth prompt including the proposed architecture and summary instructions, the large-scale generative model is made to generate a summary of the proposed architecture, and Output the generated summary to the user. It is configured to perform further actions. The architecture proposed in the claim 1.

12. The aforementioned query generation instruction includes an instruction statement that generates a query item in which the input format of the answer is restricted. The architecture proposed in the claim 1.

13. The architecture proposal device according to claim 1, wherein the query generation instruction comprises an instruction statement that lists query items in order of priority.

14. The query generation instruction comprises an instruction statement that generates query items in a format that asks for the user's request. The architecture proposed in the claim 1.

15. The control unit is configured to manage the user requirements, the proposed architecture, the verification results, and the inquiries as structured internal information. The architecture proposed in the claim 1.

16. A method of information processing performed by a computer, Accepting user requirements for the architecture of information systems from users, By providing a first prompt that includes the input user requirements and proposed instructions, the proposed architecture of the information system is generated in a large-scale generative model. By providing a second prompt including the generated proposed architecture and verification instructions, the large-scale generative model is made to generate verification results for the proposed architecture. If any deficiencies exist in the generated verification results, a third prompt is given, including the proposed architecture, the verification results, and a query generation instruction, causing the large-scale generation model to generate queries regarding the deficiencies. Output the generated inquiry items to the user. The user provides input of answers to the aforementioned inquiries. The user requirements will be updated based on the input response. By providing a fourth prompt including the updated user requirements, the proposed architecture, the verification results, the inquiries, and the resubmission instructions, the updated proposed architecture is generated in the large-scale generative model. The proposed architecture is improved by repeatedly generating the verification results of the proposed architecture, generating the inquiries, outputting the inquiries, accepting input of answers to the inquiries, updating the user requirements, and generating the updated proposed architecture until no deficiencies remain, and To output the improved proposed architecture, including, Information processing methods.

17. A program that causes a computer to execute an information processing method, The aforementioned information processing method is Accepting user requirements for the architecture of information systems from users, By providing a first prompt that includes the input user requirements and proposed instructions, the proposed architecture of the information system is generated in a large-scale generative model. By providing a second prompt including the generated proposed architecture and verification instructions, the large-scale generative model is made to generate verification results for the proposed architecture. If any deficiencies exist in the generated verification results, a third prompt is given, including the proposed architecture, the verification results, and a query generation instruction, causing the large-scale generation model to generate queries regarding the deficiencies. Output the generated inquiry items to the user. The user provides input of answers to the aforementioned inquiries. The user requirements will be updated based on the input response. By providing a fourth prompt including the updated user requirements, the proposed architecture, the verification results, the inquiries, and the resubmission instructions, the updated proposed architecture is generated in the large-scale generative model. Until no deficiencies remain, generate the verification results of the proposed architecture, generate the inquiries, output the inquiries, accept input of answers to the inquiries, update the user requirements, and The proposed architecture is improved by repeatedly generating the updated proposed architecture, and To output the improved proposed architecture, including, program.

Images