Information processing system, information processing method, and information processing program

Abstract

To support providing proper content of an architectural structure to a purchasing party of the architectural structure.SOLUTION: An information processing system according to an embodiment of the present invention has at least one processor. The at least one processor executes steps of acquiring sharing-state information relating to content which an administrator of an architectural structure has shared with a purchasing party thereof, content information including content relating to the architectural structure, and first inquiry information about the architectural structure from the purchasing party, generating first extraction information obtained by extracting at least a part of the content which the administrator can share with the purchasing party from the content information based o the sharing-state information, generating second inquiry information obtained by vectorizing the acquired first inquiry information, generating, based on the second inquiry information, second extraction information obtained by extracting at least a part of the content of the first inquiry information, and generating, based on the first inquiry information and the second extraction information, a prompt for use in requesting to the purchasing party to input answer information of the purchasing party to the first inquiry information to a model for use in generating the answer information.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] One aspect of the present disclosure relates to an information processing system, an information processing method, and an information processing program. [Background technology]

[0002] Conventionally, there have been known techniques for presenting information about real estate properties to users. For example, Patent Document 1 discloses a technique for acquiring environmental information about a real estate property and environmental preference information indicating a user's preferences for the environment of a residential location, searching a database for real estate properties having environmental information that matches the user's preferences, and presenting at least a portion of the information included in the property information of the searched real estate properties to the user. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-46854 Summary of the Invention [Problem to be solved by the invention]

[0004] One aspect of the present disclosure aims to support the provision of appropriate building content to those involved in purchasing buildings. [Means for solving the problem]

[0005] An information processing system according to one embodiment of the present disclosure includes at least one processor, which acquires sharing status information regarding content shared by a building manager with parties involved in purchasing the building, content information including content related to the building, and first question information regarding the building from parties involved in purchasing, generates first extracted information based on the sharing status information, in which at least a portion of the content that the manager can share with parties involved in purchasing is extracted from the content information, generates second question information by vectorizing the acquired first question information, generates second extracted information based on the second question information, in which at least a portion of the content is extracted from the first extracted information, and generates a prompt to be input into a model for generating answer information to the first question information from parties involved in purchasing, based on the first question information and the second extracted information.

[0006] An information processing method according to one embodiment of the present disclosure is an information processing method executed by an information processing system having at least one processor, and includes the steps of acquiring sharing status information regarding content shared by a building manager with purchasing parties for the building, content information including content regarding the building, and first question information regarding the building from purchasing parties, generating first extracted information based on the sharing status information, in which at least a portion of the content that the manager can share with purchasing parties is extracted from the content information, generating second question information by vectorizing the acquired first question information, generating second extracted information based on the second question information, in which at least a portion of the content is extracted from the first extracted information, and generating a prompt to be input into a model for generating answer information to the first question information from purchasing parties, based on the first question information and the second extracted information.

[0007] An information processing program according to one embodiment of the present disclosure causes a computer to execute the following steps: acquiring sharing status information regarding content shared by a building manager with parties involved in purchasing the building, content information including content regarding the building, and first question information regarding the building from parties involved in purchasing; generating first extracted information based on the sharing status information, in which at least a portion of the content that the manager can share with parties involved in purchasing is extracted from the content information; generating second question information by vectorizing the acquired first question information; generating second extracted information based on the second question information, in which at least a portion of the content is extracted from the first extracted information; and generating a prompt to be input into a model for generating answer information to the first question information from parties involved in purchasing, based on the first question information and the second extracted information.

[0008] According to this embodiment, at least a portion of the content is extracted from the content information related to the building based on the sharing status information, thereby generating first extracted information containing content appropriate for the purchaser. Second extracted information is generated by further extracting at least a portion of the content from the first extracted information based on second question information, which is obtained by vectorizing the first question information of the purchaser. Because a prompt is output based on the first question information and the second extracted information, a prompt that can easily generate answer information in accordance with the first question information is output. Furthermore, unlike a prompt that causes a model to generate answer information based on the first question information and content information that has not been extracted, a prompt generated based on the first question information and the second extracted information reduces the processing load. This can help provide appropriate building content to the purchaser of the building. [Effects of the Invention]

[0009] According to one aspect of the present disclosure, it is possible to assist in providing appropriate building content to those involved in purchasing a building. [Brief explanation of the drawings]

[0010] [Figure 1]1 is a diagram illustrating an example of an overall configuration of an information processing system according to an embodiment. [Figure 2] FIG. 10 is a diagram illustrating an example of a prompt. [Figure 3] FIG. 10 is a diagram illustrating an example of response information. [Figure 4] FIG. 1 is a diagram illustrating an example of a hardware configuration of an information processing system according to an embodiment. [Figure 5] 1 is a flowchart illustrating an example of an information processing method according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same or equivalent elements are designated by the same reference numerals, and redundant description will be omitted.

[0012] [System Configuration] An information processing system according to an embodiment is a computer system for providing an answer to a user who has a question about a space such as a building. Specifically, the information processing system is a computer system that extracts content information about the building in response to first question information about the space such as a building, and generates a prompt to be input into a model for generating answer information based on the extracted content acquisition and additional information.

[0013] A building refers to a structure that is fixed to a specific piece of land or space. A building is, for example, a structure that has a roof and pillars or walls. The building according to this embodiment may include a civil engineering structure. A building is, for example, a detached house, an apartment building, an office building, or a public facility. A building may also be, for example, a facility that serves as an event venue, such as a baseball stadium, a sports stadium, a racetrack, an arena, or a concert hall. The type of building is not limited. The space of a building is not limited to the space within the building, but also includes the exterior of the building and the surrounding environment of the building. The space includes objects that form the space other than the building.

[0014] Questions about spaces such as buildings refer to the content and actions that people involved in purchasing a building ask the building's manager in order to clarify details about the building and objects associated with it when designing, constructing, inspecting, managing, viewing, viewing, using, etc. In the following, an example of a question about a space such as a building is asking about information about the building when considering purchasing it.

[0015] A person involved in the purchase of a building refers to an entity that directly or indirectly participates in the purchase of the building. Such a person involved in the purchase includes, for example, visitors to the building, purchasers and renters of at least a portion of the building, and persons who visit, view, rent, purchase, or consider the procedure for at least a portion of the building. Such a person involved in the purchase is not limited to an entity that intends to purchase at least a portion of the building, but may also be a user who is related to the user who has purchased the building. Such a person involved in the purchase may include, for example, residents around the building, the manager of the building, etc. Hereinafter, a person involved in the purchase of a building may be simply referred to as a "person involved in the purchase." In this embodiment, a person involved in the purchase is, for example, a person who is considering purchasing the building.

[0016] The building manager includes an entity involved in the design, construction, inspection, management, preview, viewing, use, etc. of the building that is the subject of the question. The manager may include, for example, a designer or a person in charge of a design company who designs the building, a person in charge of a real estate company or a developer who buys and sells the building, and a builder who constructs the building. The manager may also include a manager of an information processing system. Hereinafter, the building manager may be simply referred to as the "manager." In this embodiment, the manager is, for example, a person in charge of a real estate company that buys and sells the building.

[0017] 1 is a diagram illustrating an example of the overall configuration of an information processing system 1 according to an embodiment. The information processing system 1 includes a management device 4. The management device 4 can access at least one first terminal 10, at least one second terminal 15, a content database 21, a sharing status database 22, an attribute database 23, a behavior database 24, and a model server 30 via a communication network 60.

[0018] In the information processing system 1, first extraction information is generated based on first question information input by the first terminal 10 and content information acquired from the content database 21. Second question information is generated by vectorizing the first question information. Second extraction information is generated by extracting at least a portion of content from the first extraction information based on the second question information. Furthermore, a prompt to be input into a model for generating answer information is generated based on at least the first question information and the second extraction information. The prompt may be generated further based on attribute information acquired from the attribute database 23, behavior information acquired from the behavior database 24, and recommendation information acquired from the content database 21. The generated prompt is input to the model server 30, whereby answer information is generated. The answer information is displayed on the screen of the first terminal 10. The information processing system 1 will be described in detail below.

[0019] The first terminal 10 is a computer used by a purchasing party. The second terminal 15 is a computer used by an administrator. There is no limitation on the type of computer used as the first terminal 10 and the second terminal 15. For example, the first terminal 10 and the second terminal 15 may be a mobile terminal such as a high-function mobile phone (smartphone), a mobile phone, a personal digital assistant (PDA), a tablet terminal, or a wearable terminal, or a personal computer. There is no limitation on the number of first terminals 10 and second terminals 15 in the information processing system 1. The first terminal 10 and the second terminal 15 are equipped with input devices such as a keyboard, a mouse, and a touch panel, and display devices such as a monitor and a touch panel.

[0020] The first terminal 10 accepts an input operation from the purchaser and transmits information related to the input operation to the management device 4 and the attribute database 23 via the communication network 60. The first terminal 10 accepts an input operation of a first question information about the building from the purchaser. In response to the display control of the management device 4, the first terminal 10 displays a screen on which the purchaser can view the answer information. The screen may include the first question information about the building from the purchaser. The answer information on the screen is updated in response to the user's input operation.

[0021] The first question information may include the name of the building and a question about the building. The first question information may be at least one of text data, image data, and audio data. The question about the building may include the floor plan, design, layout, interior and exterior of the building, and the environment around the building. The first question information in this embodiment is, for example, text data, such as "What schools are there near building XXX?", a question sentence (text data) about the environment around building XXX.

[0022] The answer information includes an answer to the question included in the first question information, and may be at least one of text data, image data, and audio data.

[0023] The second terminal 15 accepts the administrator's input operation and transmits information related to the input operation to the content database 21, the sharing status database 22, and the behavior database 24 via the communication network 60. The second terminal 15 accepts input operations of at least one of recommendation information, sharing status information, and behavior information related to buildings from the administrator. Details of the recommendation information, sharing status information, and behavior information will be described later.

[0024] The content database 21 is a database that stores content information including content related to buildings. In one example, data items of the content information may include a content name, content content, content location, and recommendation information. The content name may include, for example, the name of the content or a content ID. The content ID is an identifier that uniquely identifies the content assigned to the content. The content content may include information such as the name of the building, the characteristics of the building, and the environment surrounding the building. The content location may include information such as the server where the content is stored, the URL, etc. The content information is stored in the content database 21 by an administrator, a building designer, or an operator of the management device 4.

[0025] The content includes the building, each component that makes up the building, and details associated with the environment related to the building. The content may refer to materials including the building, each component that makes up the building, and details associated with the environment related to the building. The materials include literature, web pages, etc. In one example, the recommendation information is information regarding the administrator's recommendation level of the content. The recommendation level of the content is the degree to which the administrator desires the content to be included in the response information presented to the purchasing party.

[0026] The sharing status database 22 is a database that stores sharing status information. The sharing status information includes information about content shared by the administrator with the purchasing parties and information about content that the administrator can share with the purchasing parties. The sharing status information may include content that the administrator shared with the purchasing parties when the purchasing parties applied for a preview of the building, when they visited the preview, or before or after the contract was signed. Shared content is content that is recognizable to the purchasing parties, and it does not necessarily mean that the purchasing parties actually recognized the content. The sharing status information may include, as sharable content, content that is published in materials that the administrator has published and distributed. The sharing status information is associated with each content and each purchasing party. Data items of the sharing status information may include the name or ID of the purchasing party and the content name of the shared content. The sharing status information is entered by the administrator on the second terminal 15 and stored in the sharing status database 22.

[0027] The attribute database 23 is a database that stores attribute information of purchase parties. The attribute information is information about the purchase party who inputs the first question information from the first terminal 10. The data items of the attribute information may include at least one piece of information indicating the characteristics of the purchase party, such as the purchase party's name or purchase party ID, the purchase party's age, gender, occupation, address, preferences, and points that the purchase party values ​​about the building. The attribute information is associated with each purchase party. The attribute information is input in advance by the purchase party at the first terminal 10 and stored in the attribute database 23. The attribute information may also be input in advance by an administrator at the second terminal 15 and stored in the attribute database 23.

[0028] The behavioral database 24 is a database that stores behavioral information. The behavioral information is information related to the purchase of a building by a purchasing party. The behavioral information is information indicating at what stage the purchasing party is in the entire process of the building purchase or lease procedure. The behavioral information may include at least one of the following information: whether the purchasing party visited the building, the number of visits, the application by the purchasing party to purchase or lease the building, or information regarding any stage of the purchasing party's purchase or lease agreement for the building. The behavioral information is associated with each building and each purchasing party. Data items of the behavioral information may include, for example, the name of the building, the name or ID of the purchasing party, the stage of the procedure, etc. The behavioral information is entered by the administrator at the second terminal 15 and stored in the behavioral database 24.

[0029] The model server 30 is a database that stores an answer generation model 31 that generates answer information based on a prompt. The answer generation model 31 is a model for generating answer information to the first question information of a purchase-related person. The answer generation model 31 generates new answer information based on a prompt generated by the management device 4. The prompt will be described in detail later. In one example, the answer generation model 31 is stored in the model server 30 by an administrator.

[0030] The answer generation model 31 is, for example, a generative artificial intelligence (AI). A generative AI is a model that can generate content in response to a prompt input according to the commands, constraints, context, question, and output format indicated by the prompt, and return the content as answer information. The answer generation model 31 may be, for example, an interactive AI configured to include a large-scale language model (LLM) and a user interface (UI) for interacting with the user, enabling text or voice chat with the user. An example of such a generative AI is ChatGPT. In this embodiment, the model server 30 may be configured to provide content provision functions using multiple answer generation models 31. The answer generation model 31 may be stored in the model server 30 or in another device connected to the model server 30 via a network, and may be configured to enable information exchange between the management device 4 and the first terminal 10 via the model server 30. Note that while FIG. 1 illustrates only one model server 30, the information processing system 1 may include multiple model servers 30.

[0031] The management device 4 is a computer that controls the content of a screen including answer information to be displayed on the first terminal 10. The management device 4 may be a Retrieval-Augmented Generation (RAG) system. The management device 4 has, as functional components, an acquisition unit 40, a conversion unit 44, an extraction unit 45, a prompt generation unit 48, an answer generation unit 51, and an output unit 52. These functional components work together to control the content of a screen including answer information on the first terminal 10.

[0032] The acquisition unit 40 has a first acquisition unit 41, a second acquisition unit 42, and a third acquisition unit 43. The first acquisition unit 41 is a functional element that acquires first question information. The first acquisition unit 41 accepts information input at the first terminal 10. Furthermore, the first acquisition unit 41 is a functional element that acquires attribute information. The first acquisition unit 41 accepts at least a portion of the attribute information input at the first terminal 10. Here, at least a portion of the attribute information includes a purchase party name or a purchase party ID. The first acquisition unit 41 acquires attribute information corresponding to the accepted purchase party name or purchase party ID from the attribute database 23. The first acquisition unit 41 outputs the acquired first question information and attribute information to each functional element of the management device 4.

[0033] The second acquisition unit 42 is a functional element that acquires content information. The second acquisition unit 42 acquires content information including a building name that matches the building name included in the first question information acquired by the first acquisition unit 41 from the content database 21. The second acquisition unit 42 outputs the acquired content information to each functional element of the management device 4.

[0034] The third acquisition unit 43 is a functional element that acquires sharing status information and behavioral information. The third acquisition unit 43 acquires sharing status information from the sharing status database 22 based on the attribute information acquired by the first acquisition unit 41. For example, the third acquisition unit 43 acquires sharing status information corresponding to the purchase party name or purchase party ID included in the attribute information from the sharing status database 22. The third acquisition unit 43 acquires behavioral information from the behavior database 24 based on the attribute information acquired by the first acquisition unit 41. For example, the third acquisition unit 43 acquires behavioral information corresponding to the purchase party name or purchase party ID included in the attribute information from the behavior database 24.

[0035] The conversion unit 44 vectorizes the first question information acquired by the first acquisition unit 41. The conversion unit 44 generates second question information by vectorizing the first question information. For example, if the first question information is in a data format other than text data, such as image data or audio data, the conversion unit 44 converts the data format to text data. The conversion unit 44 converts the first question information, which is text data, into a vector of numerical values. The second question information is vector data of numerical values ​​indicating the first question information. In the vector data, words with similar meanings are represented by close vectors, and indicators similar to the meanings of the words are expressed by the distance and similarity of the vectors.

[0036] The conversion unit 44 maps words into a dense vector space using, for example, a model (not shown) stored in the management device 4 or the model server 30, and captures the semantic similarity between words. The model is, for example, a model such as Word2Vec or GloVe that employs a technique called Word Embeddings. The model may be trained based on text data and non-text data such as paper documents and PDF files read by OCR (Optical Character Recognition). The model may vectorize the trained text data and store it in a database.

[0037] The extraction unit 45 has a first extraction unit 46 and a second extraction unit 47. The first extraction unit 46 generates first extraction information in which at least a portion of the content that the administrator can share with the purchase parties is extracted from the content information based on the sharing status information. The first extraction unit 46 extracts content that is labeled as shareable with the purchase parties in the sharing status information from the content information. The first extraction information includes content related to buildings. The first extraction information includes content that is labeled as shareable with the purchase parties from the content information.

[0038] The second extraction unit 47 generates second extraction information by extracting at least a portion of the content from the first extraction information based on the second question information. The conversion unit 44 may convert the content included in the first extraction information into a numerical vector. The second extraction unit 47 compares the vector data included in the second question information with the vector data converted from the first extraction information to calculate the similarity between the vector data. The second extraction unit 47 compares the vector data for each content included in the first extraction information to calculate the similarity. The second extraction unit 47 extracts, as second extraction information, the content included in the first extraction information that is the source of the vector data determined to have a calculated similarity higher than a predetermined threshold. In other words, the second extraction information includes content related to buildings. The second extraction information includes content related to the first question information that is part of the content information and is labeled as shareable with purchase-related parties.

[0039] The prompt generation unit 48 generates a prompt to be input to the answer generation model 31 based on at least the first question information and the second extracted information. The prompt generation unit 48 generates a prompt to be input to the answer generation model 31 based on the first question information, the second extracted information, and at least one of the attribute information, the behavior information, and the recommendation information.

[0040] Fig. 2 is a diagram showing an example of a prompt. The prompt PR shown in Fig. 2 includes constraint conditions, building information, first question information, second extracted information, recommendation information, attribute information, and behavior information. That is, the prompt generation unit 48 generates a prompt PR that includes constraint conditions, building information, first question information, second extracted information, recommendation information, attribute information, and behavior information.

[0041] The constraints indicate commands and conditions other than the input data for the answer generation model 31 when creating answer information. The building information indicates the name of the building included in the first question information and an outline of the building extracted from the content information linked to the name. The prompt PR does not need to include building information. The example shown in Figure 2 indicates that first question information about building XXX has been obtained.

[0042] The first question information included in the prompt PR is information acquired from the first terminal 10 by the first acquisition unit 41. In the example shown in FIG. 2, the first question information includes the question "What are the schools near building XXX?" as described above.

[0043] The second extracted information included in the prompt PR includes a content name, content details, and content location. In the example shown in Fig. 2, the second extracted information includes, for example, a content name, content details, and content location related to building XXX. The answer generation model 31 to which the prompt PR including the second extracted information is input grasps the reference destinations of the second extracted information in the content database 21 and the management device 4.

[0044] The recommendation information included in the prompt PR includes the recommendation level of the content. For example, based on the recommendation information, the prompt generation unit 48 generates a prompt to be input to the answer generation model 31, which generates answer information in which the display order of the content in the answer information is adjusted according to the recommendation information. In the example shown in FIG. 2, the recommendation information is written as "Summary item: 1st, Feature item: 2nd, Background item: 3rd."

[0045] The attribute information included in the prompt PR is information acquired by the first acquisition unit 41 from the first terminal 10 and the attribute database 23. For example, based on the attribute information, the prompt generation unit 48 generates a prompt to be input to the answer generation model 31 for generating answer information in which the display order of the contents in the answer information is adjusted according to the attribute information. In the example shown in FIG. 2, the attribute information is described as "gender: male, age: 35, occupation: English teacher."

[0046] The behavioral information included in the prompt PR is information acquired from the behavioral database 24 by the third acquisition unit 43. For example, based on the behavioral information, the prompt generation unit 48 generates a prompt to be input to the answer generation model 31, which is used to generate answer information in which the display order of content in the answer information is adjusted according to the behavioral information. In the example shown in FIG. 2, the behavioral information is described as "already visited, planning to apply."

[0047] The answer generation unit 51 inputs the prompt PR to the answer generation model 31 to acquire answer information. The answer generation unit 51 inputs the prompt PR generated by the prompt generation unit 48 to the answer generation model 31. The answer generation unit 51 acquires the answer information output from the answer generation model 31.

[0048] FIG. 3 is a diagram showing an example of answer information. As shown in FIG. 3, the answer information ANS may include an answer sentence indicating an answer to the first question information and reference materials related to the answer. The answer information ANS may include at least one of image data and audio data. The answer generation model 31 generates an answer sentence such as "There is an XXX elementary school in this vicinity" as an answer to the question sentence about schools in the first question information.

[0049] The answer generation model 31, to which the prompt PR including the recommendation information is input, may generate answer information ANS that is in line with at least a part of the administrator's recommendation of the content while answering the question of the purchaser. In the example shown in Fig. 3, the answer generation model 31 generates answer information ANS so that the first paragraph of the answer indicates an overview, the second paragraph indicates features, and the third paragraph indicates further features.

[0050] The answer generation model 31, to which a prompt PR including attribute information has been input, may generate answer information ANS that is in line with at least a part of the attribute information while providing an answer to the question of the purchaser. In the example shown in Fig. 2, the attribute information of the purchaser includes information about the occupation "English teacher," so in the example shown in Fig. 3, the answer generation model 31 extracts information about English from the second extracted information and generates answer information ANS that indicates the information about English.

[0051] The answer generation model 31, to which a prompt PR including behavioral information has been input, may generate answer information ANS that is in line with at least a portion of the behavioral information while providing an answer to the question of the purchase participant. In the example shown in Figure 2, the behavioral information of the purchase participant includes information regarding the procedural stage of "planned application." Therefore, in the example shown in Figure 3, the answer generation model 31 extracts information from the second extracted information that encourages the purchase participant to apply, and generates answer information ANS that indicates that information. The answer generation model 31 may generate answer information ANS based on a combination of the acquired information from recommendation information, attribute information, and behavioral information.

[0052] The output unit 52 outputs the acquired answer information ANS. For example, the output unit 52 outputs the acquired answer information ANS to the first terminal 10. The first terminal 10 displays the output answer information ANS on a screen. Furthermore, for example, the output unit 52 may store the acquired answer information ANS in the content database 21.

[0053] Fig. 4 is a diagram showing a general hardware configuration of the management device 4. As shown in Fig. 4, for example, the management device 4 includes, as physical components, a processor 101, a main memory unit 102, an auxiliary memory unit 103, a communication control unit 104, an input device 105, and an output device 106.

[0054] The processor 101 executes an operating system and application programs. The main memory 102 is composed of, for example, ROM and RAM. The auxiliary memory 103 is composed of, for example, a hard disk or flash memory, and generally stores a larger amount of data than the main memory 102. The auxiliary memory 103 stores an information processing program P for realizing each of the above-mentioned functional elements in the management device 4. The communication control unit 104 is composed of, for example, a wireless communication module. The input device 105 is composed of, for example, a keyboard, mouse, or touch panel. The output device 106 is composed of, for example, a monitor and speaker.

[0055] Each functional element of the management device 4 is realized by loading an information processing program P onto the processor 101 or the main memory 102 and having the processor 101 execute the information processing program P. The information processing program P includes code for implementing the acquisition unit 40, the conversion unit 44, the extraction unit 45, the prompt generation unit 48, the answer generation unit 51, and the output unit 52. The processor 101 operates the communication control unit 104, the input device 105, or the output device 106 in accordance with the information processing program P, and reads and writes data from and to the main memory 102 or the auxiliary memory 103. This processing realizes each functional element of the management device 4. Data or databases required for processing may be stored in the main memory 102 or the auxiliary memory 103.

[0056] The information processing program P may be provided by being fixedly recorded on a tangible recording medium such as a semiconductor memory. Alternatively, the information processing program P may be provided via a communication network as a data signal superimposed on a carrier wave. The provided information processing program P is stored in the auxiliary storage unit 103.

[0057] [System Operation] The operation of the management device 4 will be described, and an information processing method according to this embodiment will be described with reference to Fig. 5. Fig. 5 is a flowchart showing an example of the information processing method according to the embodiment. With reference to Fig. 5, a control for outputting answer information by the management device 4 when first question information is input from the first terminal 10 in the information processing system 1 will be described.

[0058] In step S1, the first acquisition unit 41 acquires first question information. The purchaser inputs the first question information by performing an input operation on the first terminal 10. The input first question information is transmitted to the management device 4 via the communication network 60.

[0059] In step S2, the first acquisition unit 41 acquires attribute information. In response to the first acquisition unit 41 receiving the first question information from the first terminal 10, the first acquisition unit 41 acquires the purchase party name or purchase party ID of the first terminal 10 that sent the first question information. The first acquisition unit 41 reads out attribute information corresponding to the purchase party name or purchase party ID from the attribute database 23, and acquires the attribute information indicated by the purchase party name or purchase party ID.

[0060] In step S3, the conversion unit 44 vectorizes the first question information to generate second question information. The conversion unit 44 may convert the data format of the first question information into text data. The conversion unit 44 converts the first question information represented as text data into numerical vector data, and sets the vector data as second question information.

[0061] In step S4, the second acquisition unit 42 acquires content information. The second acquisition unit 42 acquires, from the content database 21, content information including a building name that matches the building name included in the first question information acquired in step S1.

[0062] In step S5, the third acquisition unit 43 acquires sharing status information and behavioral information. Based on the attribute information acquired in step S2, the third acquisition unit 43 acquires sharing status information from the sharing status database 22. Based on the attribute information acquired in step S2, the third acquisition unit 43 acquires behavioral information from the behavior database 24.

[0063] In step S6, the first extraction unit 46 generates first extraction information, which is generated by extracting at least a portion of the content that the administrator can share with purchase-related parties from the content information acquired in step S4, based on the sharing status information acquired in step S5.

[0064] In step S7, the second extraction unit 47 generates second extraction information, which is generated by extracting at least a portion of the content from the first extraction information extracted in step S6, based on the second question information converted in step S3.

[0065] In step S8, the prompt generation unit 48 generates a prompt PR to be input to the answer generation model 31 based on the first question information, the second extracted information, and at least one of the attribute information, the behavior information, and the recommendation information. In this embodiment, the prompt generation unit 48 generates a prompt PR to be input to the answer generation model 31 based on the first question information, the second extracted information, the attribute information, the behavior information, and the recommendation information.

[0066] In step S9, the answer generator 51 inputs the prompt PR generated in step S8 to the answer generation model 31. The answer generator 51 obtains the answer information ANS output from the answer generation model 31.

[0067] In step S10, the output unit 52 outputs the answer information ANS acquired in step S9. For example, the output unit 52 outputs the acquired answer information ANS to the first terminal 10. The first terminal 10 displays the output answer information ANS on the screen. This ends the flowchart of an example of an information processing method shown in FIG.

[0068] [effect] As described above, an information processing system 1 according to one embodiment of the present disclosure includes at least one processor 101, which acquires sharing status information regarding content shared by the building manager with those involved in purchasing the building, content information including content regarding the building, and first question information regarding the building from those involved in purchasing, generates first extracted information based on the sharing status information, in which at least a portion of the content that the manager can share with those involved in purchasing is extracted from the content information, generates second question information by vectorizing the acquired first question information, generates second extracted information based on the second question information, in which at least a portion of the content is extracted from the first extracted information, and generates a prompt PR to be input into a model (answer generation model 31) for generating answer information ANS to the first question information of those involved in purchasing, based on the first question information and the second extracted information.

[0069] An information processing method according to one embodiment of the present disclosure is an information processing method executed by an information processing system 1 having at least one processor 101, and includes steps of acquiring sharing status information regarding content shared by a building manager with purchasers of the building, content information including content regarding the building, and first question information regarding the building from purchasers (steps S1, S4, S5); generating first extracted information (step S6) based on the sharing status information, in which at least a portion of the content that the manager can share with purchasers is extracted from the content information; generating second question information by vectorizing the acquired first question information (step S3); generating second extracted information based on the second question information, in which at least a portion of the content is extracted from the first extracted information (step S7); and generating a prompt PR to be input into a model (answer generation model 31) for generating answer information ANS to the first question information of purchasers based on the first question information and the second extracted information (step S8).

[0070] An information processing program P according to one embodiment of the present disclosure causes a computer to execute the following steps: acquiring sharing status information regarding content shared by a building manager with purchasers of the building, content information including content regarding the building, and first question information regarding the building from purchasers based on the sharing status information (steps S1, S4, S5); generating first extracted information (step S6) by extracting at least a portion of the content that the manager can share with purchasers from the content information based on the sharing status information; generating second question information by vectorizing the acquired first question information (step S3); generating second extracted information (step S7) by extracting at least a portion of the content from the first extracted information based on the second question information; and generating a prompt PR to be input into a model (answer generation model 31) for generating answer information ANS to the first question information of purchasers based on the first question information and the second extracted information (step S8).

[0071] According to this embodiment, at least a portion of the content is extracted from the content information related to the building based on the sharing status information, thereby generating first extracted information containing content appropriate for the purchaser. Second extracted information is generated by further extracting at least a portion of the content from the first extracted information based on second question information, which is obtained by vectorizing the first question information of the purchaser. Because a prompt PR is output based on the first question information and the second extracted information, a prompt PR that can easily generate appropriate answer information ANS to the first question information is output. Furthermore, unlike a prompt PR generated by answer information based on the first question information and content information that has not been extracted, a prompt PR generated based on the first question information and the second extracted information has a smaller processing load. This can help provide appropriate building content to those involved in the purchase of the building.

[0072] Information about a building can be a huge amount of data, covering a wide range of areas, including not only overall information about the building itself, information about each component that makes up the building, and information about the surrounding area of ​​the building, but also information about the design, construction, and pending matters of the building, information about the purchase or lease of the building, etc. If answer information to the first question information is generated based on the first question information and content information from which nothing has been extracted, the answer information may be generated from information that cannot be shared with the purchasing parties, or the answer information may contain information that does not match (is inappropriate for) the needs of the purchasing parties, such as the stage of the procedure.

[0073] On the other hand, in the information processing system 1, the information input to the answer generation model 31 is the second extracted information, so the answer information is generated from information that can be shared with those involved in the purchase. Also, because the second extracted information, which may have a smaller amount of data than the content information, is specified in the prompt PR, it becomes easier for the answer generation model 31 to select and reject information, and the processing load is reduced. This makes it possible to support the provision of appropriate building content to those involved in the purchase of a building.

[0074] In another embodiment of the information processing system 1, at least one processor 101 may generate first extracted information, in which content shared by the administrator with the purchase parties is extracted from the content information, based on the sharing status information. Second extracted information is generated based on the first extracted information generated in this manner. Since the information input to the answer generation model 31 is the second extracted information, answer information is generated from the information shared with the purchase parties. This prevents the purchase parties from obtaining answer information ANS from content not intended by the administrator, and makes it easier for the administrator to appropriately control the content provided to the purchase parties.

[0075] In another embodiment of the information processing system 1, at least one processor 101 may input a prompt PR to the answer generation model 31 to obtain answer information ANS, and output the answer information ANS. In this way, the answer generation model 31 generates answer information from information that can be shared with purchase-related parties based on the prompt PR. Furthermore, because the prompt PR specifies second extracted information that may have a smaller amount of data than the content information, the answer generation model 31 can easily generate answer information. Therefore, answer information containing content appropriate for purchase-related parties of the building is generated.

[0076] In another embodiment of the information processing system, at least one processor 101 may generate second extracted information by extracting from the first extracted information content whose similarity to the second question information is higher than a predetermined threshold. According to this embodiment, the second extracted information can include content that is highly similar to the second question information, which is a vectorized version of the first question information of the purchase-related person. This allows the prompt PR to include content that is highly similar to the first question information, making it possible to generate a prompt PR that can generate appropriate answer information ANS to the first question information.

[0077] In another embodiment of the information processing system, the at least one processor 101 may further acquire attribute information of the purchase participants and generate a prompt PR based on the first question information, the second extracted information, and the attribute information. By generating the prompt PR based on the attribute information in this manner, it is possible to prompt the purchase participants to output answer information according to their attributes.

[0078] In another embodiment of the information processing system, the at least one processor 101 may further acquire behavioral information of the purchasers regarding the purchase of the building, and generate a prompt PR based on the first question information, the second extracted information, and the behavioral information. By generating the prompt PR based on the behavioral information in this way, it is possible to prompt the purchasers to output answer information corresponding to the behavioral information, such as the stage of the purchase procedure.

[0079] In another embodiment of the information processing system, at least one processor 101 may further acquire recommendation information related to the recommendation level of the administrator's content, and generate a prompt PR based on the first question information, the second extracted information, and the recommendation information. By generating the prompt PR based on the recommendation information in this manner, it is possible to prompt the administrator to output answer information according to the recommendation level of the administrator's content.

[0080] [Variations] The present disclosure has been described in detail above based on the embodiments. However, the present disclosure is not limited to the above embodiments. Various modifications of the present disclosure are possible without departing from the spirit and scope of the present disclosure. For example, the second acquisition unit 42 may acquire first extracted information from the content database 21, including content that the administrator can share with purchase parties, based on the sharing status information. In this case, the information processing system according to the modified example may not acquire content information in step S4, but may acquire first extracted information from the content database 21, including content that the administrator can share with purchase parties, based on the sharing status information, in step S6. The second acquisition unit 42 acquires content information from the content database 21 as first extracted information, including a building name that matches the building name included in the first question information acquired in step S1 and including content labeled as shareable with purchase parties in the sharing status information. In this case, the information processing system according to the modified example achieves the same effects as the information processing system 1 according to the above embodiment. Furthermore, the amount of data acquired by the second acquisition unit 42 from the content database 21 can be reduced, thereby reducing communication costs and traffic and enabling high-speed processing.

[0081] The third acquisition unit 43 may also update the sharing status information based on the behavioral information. If the administrator has not input or updated the sharing status information via the second terminal 15, the sharing status information stored in the sharing status database 22 may not exist or may remain outdated. The third acquisition unit 43 extracts content that can be shared by the administrator according to the stage of the purchase or rental procedure included in the behavioral information. If the content that can be shared by the administrator is available and the shared content is not included in the sharing status information acquired from the sharing status database 22, the third acquisition unit 43 updates the sharing status information acquired from the sharing status database 22. The third acquisition unit 43 also compares the content that can be shared by the administrator with the sharing status information acquired from the sharing status database 22, and if the sharing status information acquired from the sharing status database 22 contains only a small amount of content, updates the sharing status information. Updating the sharing status information includes storing information about the content that can be shared by the administrator, extracted based on the behavioral information, in the sharing status information.

[0082] The third acquisition unit 43 may also update the behavioral information based on the shared status information. If the administrator has not input or updated the behavioral information via the second terminal 15, the behavioral information stored in the behavior database 24 may not exist or may remain outdated. The third acquisition unit 43 estimates the behavioral information of the purchase participants based on the content included in the shared status information. The third acquisition unit 43 compares the behavioral information of the purchase participants estimated based on the shared status information with the behavioral information acquired from the behavior database 24, and if the behavioral information acquired from the behavior database 24 is outdated, updates the behavioral information. Updating the behavioral information includes replacing the behavioral information stored in the behavior database 24 with behavioral information estimated based on the shared status information. In this way, the third acquisition unit 43 has at least one of the functions of updating the shared status information based on the behavioral information and updating the behavioral information based on the shared status information, thereby appropriately assisting the administrator in inputting information via the second terminal 15.

[0083] Furthermore, even if either the behavioral information or the sharing status information is not the latest information, by using the third acquisition unit 43 in the above-described form, the prompt generation unit 48 can generate a prompt PR to be input to the answer generation model 31 based on the latest information about the purchase parties. In other words, a prompt PR that can easily generate appropriate answer information ANS to the first question information is output. This can help provide more appropriate building content to those involved in the purchase of a building.

[0084] Here, various exemplary embodiments included in the present disclosure are described in the following [Embodiment 1] to [Embodiment 10].

[0085] [Form 1] at least one processor; The at least one processor: Acquire sharing status information regarding content shared by a building manager with a person involved in the purchase of the building, content information including the content regarding the building, and first question information regarding the building from the person involved in the purchase; generating first extracted information in which at least a part of the content that can be shared by the administrator with the purchasing parties is extracted from the content information based on the sharing status information; generating second question information by vectorizing the acquired first question information; generating second extraction information in which at least a part of the content is extracted from the first extraction information based on the second question information; generating a prompt to be input into a model for generating answer information to the first question information of the purchasing party based on the first question information and the second extracted information; Information processing system.

[0086] [Form 2] the at least one processor generates the first extracted information, in which the content shared by the administrator with the purchasing parties is extracted from the content information, based on the sharing status information; The information processing system according to [Mode 1].

[0087] [Form 3] The information processing system according to [Mode 1] or [Mode 2], wherein the at least one processor inputs the prompt into the model to obtain the answer information, and outputs the answer information.

[0088] [Form 4] the at least one processor generates second extracted information by extracting content from the first extracted information that has a similarity to the second question information higher than a predetermined threshold. The information processing system according to any one of [Mode 1] to [Mode 3].

[0089] [Form 5] The at least one processor: Further acquiring attribute information of the purchasing party; generating the prompt based on the first question information, the second extracted information, and the attribute information; The information processing system according to any one of [Mode 1] to [Mode 4].

[0090] [Form 6] The at least one processor: Further acquiring behavioral information regarding the purchase of the building by the purchaser; generating the prompt based on the first question information, the second extracted information, and the behavioral information; The information processing system according to any one of [Mode 1] to [Mode 5].

[0091] [Form 7] The at least one processor: further acquiring recommendation information regarding the manager's recommendation level of the content; generating the prompt based on the first question information, the second extracted information, and the recommendation information; The information processing system according to any one of [Mode 1] to [Mode 6].

[0092] [Form 8] 1. An information processing method executed by an information processing system including at least one processor, comprising: A step of acquiring sharing status information regarding content shared by a building manager with a person involved in purchasing the building, content information including content regarding the building, and first question information regarding the building from the person involved in purchasing; generating first extracted information in which at least a part of the content that can be shared by the administrator with the purchasing parties is extracted from the content information based on the sharing status information; generating second question information by vectorizing the acquired first question information; generating second extraction information in which at least a part of the content is extracted from the first extraction information based on the second question information; generating a prompt to be input into a model for generating answer information to the first question information of the purchasing party based on the first question information and the second extracted information; An information processing method, including:

[0093] [Form 9] A step of acquiring sharing status information regarding content shared by a building manager with a person involved in purchasing the building, content information including content regarding the building, and first question information regarding the building from the person involved in purchasing; generating first extracted information in which at least a part of the content that can be shared by the administrator with the purchasing parties is extracted from the content information based on the sharing status information; generating second question information by vectorizing the acquired first question information; generating second extraction information in which at least a part of the content is extracted from the first extraction information based on the second question information; generating a prompt to be input into a model for generating answer information to the first question information of the purchasing party based on the first question information and the second extracted information; An information processing program that causes a computer to execute the above.

[0094] [Form 10] at least one processor; The at least one processor: Acquire sharing status information regarding content related to the building that has been shared by a building manager with a person involved in purchasing the building, and first question information regarding the building from the person involved in purchasing the building; The administrator acquires first extracted information including the content that can be shared with the purchasing parties based on the sharing status information; generating second question information by vectorizing the acquired first question information; generating second extraction information in which at least a part of the content is extracted from the first extraction information based on the second question information; generating a prompt to be input into a model for generating answer information to the first question information of the purchasing party based on the first question information and the second extracted information; Information processing system. [Explanation of symbols]

[0095] 1...information processing system, 4...management device, 10...first terminal, 15...second terminal, 21...content database, 22...sharing status database, 23...attribute database, 24...behavioral database, 30...model server, 31...answer generation model, 40...acquisition unit, 41...first acquisition unit, 42...second acquisition unit, 43...third acquisition unit, 44...conversion unit, 45...extraction unit, 46...first extraction unit, 47...second extraction unit, 48...prompt generation unit, 51...answer generation unit, 52...output unit, 60...communication network, ANS...answer information, P...information processing program, PR...prompt.

Claims

1. at least one processor; The at least one processor Acquire sharing status information regarding content shared by a building manager with a person involved in the purchase of the building, content information including the content regarding the building, and first question information regarding the building from the person involved in the purchase; generating first extracted information in which at least a part of the content that can be shared by the administrator with the purchasing parties is extracted from the content information based on the sharing status information; generating second question information by vectorizing the acquired first question information; generating second extraction information in which at least a part of the content is extracted from the first extraction information based on the second question information; generating a prompt to be input into a model for generating answer information to the first question information of the purchasing party based on the first question information and the second extracted information; Information processing system.

2. the at least one processor generates the first extracted information in which the content shared by the administrator with the purchasing parties is extracted from the content information based on the sharing status information; The information processing system according to claim 1 .

3. The information processing system according to claim 1 , wherein the at least one processor inputs the prompt into the model to obtain the answer information, and outputs the answer information.

4. the at least one processor generates second extracted information in which content having a similarity to the second question information higher than a predetermined threshold is extracted from the first extracted information.

3. The information processing system according to claim 1.

5. The at least one processor Further acquiring attribute information of the purchasing party; generating the prompt based on the first question information, the second extracted information, and the attribute information; 3. The information processing system according to claim 1.

6. The at least one processor Further acquiring behavioral information regarding the purchase of the building by the purchaser; generating the prompt based on the first question information, the second extracted information, and the behavioral information; 3. The information processing system according to claim 1.

7. The at least one processor further acquiring recommendation information regarding the manager's recommendation level of the content; generating the prompt based on the first question information, the second extracted information, and the recommendation information; 3. The information processing system according to claim 1.

8. 1. An information processing method executed by an information processing system including at least one processor, comprising: A step of acquiring sharing status information regarding content shared by a building manager with a person involved in purchasing the building, content information including content regarding the building, and first question information regarding the building from the person involved in purchasing; generating first extracted information in which at least a part of the content that can be shared by the administrator with the purchasing parties is extracted from the content information based on the sharing status information; generating second question information by vectorizing the acquired first question information; generating second extraction information in which at least a part of the content is extracted from the first extraction information based on the second question information; generating a prompt to be input into a model for generating answer information to the first question information of the purchasing party based on the first question information and the second extracted information; An information processing method, including:

9. A step of acquiring sharing status information regarding content shared by a building manager with a person involved in purchasing the building, content information including content regarding the building, and first question information regarding the building from the person involved in purchasing; generating first extracted information in which at least a part of the content that can be shared by the administrator with the purchasing parties is extracted from the content information based on the sharing status information; generating second question information by vectorizing the acquired first question information; generating second extraction information in which at least a part of the content is extracted from the first extraction information based on the second question information; generating a prompt to be input into a model for generating answer information to the first question information of the purchasing party based on the first question information and the second extracted information; An information processing program that causes a computer to execute the above.

10. at least one processor; The at least one processor Acquire sharing status information regarding the content related to the building that has been shared by the building manager with the purchasers of the building, and first question information regarding the building from the purchasers; The administrator acquires first extracted information including the content that can be shared with the purchasing parties based on the sharing status information; generating second question information by vectorizing the acquired first question information; generating second extraction information in which at least a part of the content is extracted from the first extraction information based on the second question information; generating a prompt to be input into a model for generating answer information to the first question information of the purchasing party based on the first question information and the second extracted information; Information processing system.

Images