Processing device, processing program and processing method

Abstract

To provide a processing device, a processing program, and a processing method that can more efficiently processing questions from users who use a property.SOLUTION: A processing device is provided with at least one processor. The at least one processor is configured to: receive question information indicating a question related to a property from a terminal device available to a user who uses the property; acquire answer information indicating an answer to the question from a procedure-trained model obtained by machine learning of procedure information indicating procedures related to the property or articles installed in the property, based on the received question information; and execute processing to output the acquired answer information to the user terminal device.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a processing device, a processing program, and a processing method for processing questions from a user using a property. [Background technology]

[0002] The number of rental properties that have introduced a system called smart lock that digitizes the handover of keys has been increasing. In such rental properties, procedures are simplified for both the management company and the tenants who use the property. However, the instruction manuals for the rental property itself and the various items installed in the rental property that show how to use them are still provided in paper form, so the management company must visit the rental property from time to time to provide explanations.

[0003] In addition, there are many different items installed in the rental property, and the number of instruction manuals is huge. In such a situation, it is a hassle for tenants to search for the desired page among the huge number of paper instruction manuals. On the other hand, the operator also faces various challenges, such as operational costs for responding to inquiries from tenants and securing personnel.

[0004] Furthermore, in the case of instruction manuals provided in paper form, even if the usage procedures or the like are updated, once they have been distributed, it is difficult to update them.

[0005] Incidentally, as described in Patent Document 1, "a user assist system comprising: an acquisition means for acquiring text data input by a user; a selection means for selecting one or more pieces of output data including functional information for input data including a word group including one or more words included in the text data acquired by the acquisition means, with reference to a selection model showing the relationship between the input data including a word group including one or more words acquired in advance and output data including functional information related to the functions of the software; and a presentation means for presenting the functional information included in the output data selected by the selection means to the user," in recent years, there has been an increase in the introduction of online chat responses by businesses or responses by so-called chatbots to various inquiries from users. However, even when responses are made by online chat, it is necessary to secure personnel, and even when responses are made by chatbots, various preparations such as information required for responses, questions, and responses, etc., must be made in advance. Therefore, there are various issues with the introduction of such systems, such as time and cost. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent Publication No. 2023-55405 Summary of the Invention [Problem to be solved by the invention]

[0007] Therefore, in consideration of the above-mentioned technology, the present disclosure aims to provide a processing device, processing program, and processing method that can process questions from users who use a property more efficiently through various embodiments. [Means for solving the problem]

[0008] According to one aspect of the present disclosure, there is provided a processing device having at least one processor, the at least one processor being configured to execute a process for receiving question information indicating a question related to a property from a terminal device available to a user using the property, acquiring answer information indicating an answer to the question from a procedure-learned model obtained by machine learning procedure information indicating procedures related to the property or an item installed in the property based on the received question information, and outputting the acquired answer information to the user terminal device.

[0009] According to one aspect of the present disclosure, there is provided a processing program that causes a computer having at least one processor to function as follows: receive question information indicating a question related to a property from a terminal device available to a user using the property, acquire answer information indicating an answer to the question from a procedure-learned model obtained by machine learning procedure information indicating procedures related to the property or an item installed in the property based on the received question information, and execute a process to output the acquired answer information to the user terminal device.

[0010] According to one aspect of the present disclosure, there is provided a "processing method executed by at least one processor in a computer having the at least one processor, the processing method including the steps of receiving question information indicating a question related to the property from a terminal device available to a user who uses the property, acquiring answer information indicating an answer to the question from a procedure-learned model obtained by machine learning procedure information indicating procedures related to the property or an item installed in the property based on the received question information, and outputting the acquired answer information to the user terminal device." Effect of the Invention

[0011] According to the present disclosure, it is possible to provide a processing device, a processing program, and a processing method capable of more efficiently processing questions from users who use a property.

[0012] It should be noted that the above effects are merely illustrative for the convenience of explanation and are not limiting. In addition to or instead of the above effects, any effect described in this disclosure or any effect obvious to a person skilled in the art may be achieved. [Brief description of the drawings]

[0013] [Figure 1] FIG. 1 is a block diagram showing a configuration of a processing system 1 according to an embodiment of the present disclosure. [Figure 2A] FIG. 2A is a block diagram showing a configuration of the management device 100 according to an embodiment of the present disclosure. [Figure 2B] FIG. 2B is a block diagram showing a configuration of a terminal device 200 according to an embodiment of the present disclosure. [Figure 3A] FIG. 3A is a diagram conceptually illustrating a property management table stored in management device 100 according to an embodiment of the present disclosure. [Figure 3B] FIG. 3B is a diagram conceptually illustrating a procedure management table stored in management device 100 according to an embodiment of the present disclosure. [Figure 3C] FIG. 3C is a diagram conceptually illustrating a knowledge management table stored in the management device 100 according to an embodiment of the present disclosure. [Figure 3D] FIG. 3D is a diagram conceptually illustrating a feedback management table stored in management device 100 according to an embodiment of the present disclosure. [Figure 4A] FIG. 4A is a diagram showing a processing sequence executed by the processing system 1 according to an embodiment of the present disclosure. [Figure 4B] FIG. 4B is a diagram showing a processing sequence executed by the processing system 1 according to an embodiment of the present disclosure. [Figure 5A] FIG. 5A is a diagram showing a process flow executed in the management device 100 according to an embodiment of the present disclosure. [Figure 5B] FIG. 5B is a diagram showing a process flow executed in the management device 100 according to an embodiment of the present disclosure. [Figure 6A] FIG. 6A is a diagram conceptually illustrating an example of a screen output on the user terminal device 200-1 according to an embodiment of the present disclosure. [Figure 6B] FIG. 6B is a diagram conceptually illustrating an example of a screen output on the user terminal device 200-1 according to an embodiment of the present disclosure. [Figure 6C] FIG. 6C is a diagram conceptually illustrating an example of a screen output on the user terminal device 200-1 according to an embodiment of the present disclosure. [Figure 6D] FIG. 6D is a diagram conceptually illustrating an example of a screen output on the user terminal device 200-1 according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] 1. Overview of Processing System 1 The processing system 1 according to the present disclosure is used, for example, to receive questions from a user who uses a property and to answer the questions. Specifically, when the processing system 1 receives a question related to the property from a user terminal device available to the user who uses the property, the processing system 1 is used to obtain an answer using a procedure-learned model that has previously learned procedures related to the property, and to output the obtained answer to the user terminal device.

[0015] In this disclosure, "property" refers to movable or real property that can be used by a user. In other words, property includes both real property, which is land and its fixtures, and movable property, which is other than the real property. Examples of such property include real property, which is land, office buildings, apartment buildings, factories, warehouses, houses, parking lots, bicycle parking lots, and individual plots divided therein, as well as movable property, such as automobiles, ships, money and valuables, jewelry, watches, electrical appliances, and furniture. Among these properties, in this disclosure, real property provided to users by businesses that manage these properties is preferably used, and real property that is expected to be managed by the business is further used publicly.

[0016] In addition, in this disclosure, "goods" refers to goods other than real estate that are installed in a property. Examples of such goods include automobiles, ships, money and valuables, jewelry, watches, electrical appliances, toilets, kitchen equipment, public facilities such as gas, electricity, and water, locking and unlocking devices such as smart keys, security equipment, shared facilities of the property, etc. Among these goods, electrical appliances, toilets, kitchen equipment, public facilities such as gas, electricity, and water, locking and unlocking devices such as smart keys, and security equipment are preferably used in this disclosure.

[0017] In addition, in the present disclosure, "procedure information" is information indicating procedures related to a property or an item installed in the property. The procedure information is not limited to information directly describing the procedure for using the property or the item installed in the property, but may be any information used to understand the procedure for using the property. Examples of such procedure information include documents that guide each business such as gas, electricity, and water when moving into the property, property management regulations documents, insurance-related documents such as earthquake insurance and fire insurance for the property, contract documents for telephone and Internet, contract documents for the property such as a rental contract exchanged with the property manager, a copy of the important information explanation for the property, a notice of cancellation of the property, an inquiry desk for the business managing the property, the check results of the move-in checklist for the property created by the user who is the resident at the time of moving in, instructions for using shared facilities such as a parking lot, bicycle parking lot, and guest room, documents informing the user of the garbage sorting method and collection date in the area where the property is located, documents informing the evacuation site and evacuation procedure in the event of a disaster, an instruction manual for the property indicating the procedure for using the property when moving into the property, a list of parts and maintenance products used in the property, and procedure information related to the property that is a combination of these. Other examples of such procedure information include various instruction manuals for items installed in the property, their warranties, lists of parts and maintenance supplies used in these, and procedure information related to items that is a combination of these. At least a portion of the procedure information related to such items and procedure information related to items includes information indicating procedures related to the items or items, and is mainly used as first procedure information for machine learning in the first procedure trained model.

[0018] Further, another example of the procedure information in the present disclosure includes questions that have been asked by various users in the past to the manager of the property or the provider of the item installed in the property, and answers to the questions, in relation to the property or the item installed in the property. Such questions and answers may be created by the manager of the property, or may be created by the installer of the property or the manufacturer of the item installed in the property. Furthermore, the questions and answers do not need to be created by a person, and may be created, for example, by a system used for answering. These past questions and answers are mainly used for machine learning in the second trained model as second procedure information.

[0019] In addition, in this disclosure, the terms "first", "second", etc. may be used, but unless otherwise specified, these are merely names used for the convenience of explanation of both. Therefore, even if the terms "first", "second", etc. are used, it does not mean that the elements are limited to those to which they are attached. Naturally, the processing system 1 may include at least one of the elements "first" and "second", or may include multiple elements such as "third", "fourth", and more.

[0020] Fig. 1 is a block diagram showing a configuration of a processing system 1 according to an embodiment of the present disclosure. According to Fig. 1, the processing system 1 includes at least a management device 100 for managing questions from users in the processing system 1, a user terminal device 200-1 available to a user who uses a property or an item installed in the property, and a business operator terminal device 200-2 available to a business operator who manages the property or the item installed in the property, which are communicatively connected via a communication network.

[0021] Furthermore, the processing system 1 may use a procedure-learned model in its processing, and a model generation device 300 for generating the procedure-learned model is communicably connected to the management device 100 and the like via a communication network. Note that the model generation device 300 may be used independently of the management device 100 as shown in FIG. 1, or the management device 100 may function as the model generation device 300.

[0022] In FIG. 1, the user terminal device 200-1 and the provider terminal device 200-2 are each shown as a single device, but of course they are not limited to a single device and may include a plurality of user terminal devices or a plurality of provider terminal devices.

[0023] 1, the processing device is not shown, but may be any of the management device 100, the user terminal device 200-1, and the business operator terminal device 200-2. That is, the following describes a case where the management device 100 executes a process of acquiring answer information from a procedure-learned model based on question information from a user who uses a property, but the process may be executed by the user terminal device 200-1 or the business operator terminal device 200-2.

[0024] 2. Configuration of management device 100 Fig. 2A is a block diagram showing a configuration of a management device 100 according to an embodiment of the present disclosure. The management device 100 does not need to include all of the components shown in Fig. 2A, and it is possible to adopt a configuration in which some components are omitted, or other components can be added. The management device 100 does not need to include the components shown in Fig. 2A in a single housing, and each component and process of the management device 100 can be distributed to multiple devices such as on-premise server devices and cloud server devices.

[0025] 2A, the management device 100 includes a processor 111 including a CPU or the like, a memory 112 including a RAM, a ROM, a non-volatile memory, an HDD, or the like, and a communication interface 113. These components are electrically connected to each other via control lines and data lines.

[0026] The processor 111 is composed of a CPU (microcomputer) and functions as a control unit for controlling other components connected thereto based on various programs stored in the memory 112. The processor 111 reads out from the memory 112 a program for executing an application according to the present disclosure and a program for executing an OS, and executes the program. Specifically, the processor 111 executes the following based on the programs stored in the memory 112: "receiving question information indicating a question related to a property from a terminal device available to a user who uses the property," "acquiring answer information indicating an answer to the question from a procedure-learned model obtained by machine learning procedure information indicating a procedure related to the property or an item installed in the property, based on the received question information," "converting the received question information into input information to be input to the procedure-learned model," "outputting the acquired answer information to a user terminal device," and "acquiring feedback information by a user on the answer information from a user terminal device." The processor 111 is mainly composed of one or more CPUs, but may be appropriately combined with a GPU, an FPGA, or the like.

[0027] The memory 112 includes a RAM, a ROM, a non-volatile memory, and a HDD, and functions as a storage unit. The ROM stores instructions for executing the application and the OS according to the present disclosure as a program. Such a program is loaded and executed by the processor 111. The RAM is used to write and read data while the program stored in the ROM is being processed by the processor 111. The non-volatile memory is a memory into which data is written and read by the execution of the program, and the data written therein is saved even after the execution of the program is terminated. Specifically, the memory 112 stores a program for the processor 111 to execute the above-mentioned processes, etc. The memory 112 also stores a procedure-learned model. The memory 112 further stores various information such as a property management table (FIG. 3A), a procedure management table (FIG. 3B), a knowledge management table (FIG. 3C), and a feedback management table (FIG. 3D). Note that these pieces of information may be stored in a database device communicably connected to the management device 100 as necessary (in this case, the database may also be included as one of the memories 112).

[0028] The communication interface 113 functions as a communication unit that transmits and receives information to and from other devices, such as the user terminal device 200-1 and the operator terminal device 200-2, which are installed remotely, via a communication processing circuit and an antenna. The communication processing circuit processes programs and various information used in the processing system 1 to transmit and receive information according to the progress of processing. The communication processing circuit processes based on a wideband wireless communication method such as the LTE method, but it is also possible to process based on a method related to narrowband wireless communication such as wireless LANs such as IEEE802.11 and Bluetooth (registered trademark) or a method related to non-contact wireless communication. It is also possible to use wired communication instead of or in addition to wireless communication.

[0029] In Figure 2A, the procedure-learned model is described as being stored in memory 112, but it may also be stored in the memory of a server device communicatively connected to the management device 100, and output information may be obtained by inputting input information received from the management device 100 into the procedure-learned model, and the output information may be transmitted to the management device 100.

[0030] 3. Configuration of Terminal Device 200 Fig. 2B is a block diagram showing a configuration of a terminal device 200 according to an embodiment of the present disclosure. The terminal device 200 does not need to have all of the components shown in Fig. 2B, and may have a configuration in which some components are omitted, or may have other components added. The terminal device 200 is used as a user terminal device 200-1 or a provider terminal device 200-2, but both do not need to have the same configuration, and each terminal device may have a different configuration.

[0031] 2B, the terminal device 200 includes a processor 211 including a CPU or the like, a memory 212 including a RAM, a ROM, a non-volatile memory, an HDD, and the like, a communication interface 213, an input interface 214, and an output interface 215. These components are electrically connected to each other via control lines and data lines.

[0032] The processor 211 is composed of a CPU (microcomputer) and functions as a control unit for controlling other connected components based on various programs stored in the memory 212. The processor 211 reads out from the memory 212 a program for executing an application according to the present disclosure and a program for executing an OS, and executes the program. Specifically, when the processor 211 functions as the user terminal device 200-1, the processor 211 executes the following processes based on the programs stored in the memory 212: "accepting an operation input from a user using a property via the input interface 214 and inputting question information indicating a question related to the property", "outputting question information input via the communication interface 213 to the management device 100", "receiving answer information received from the management device via the communication interface 213", "outputting answer information received via an output interface", and "accepting an operation input from a user via an input interface and inputting feedback information for the received answer information". The processor 211 is mainly composed of one or more CPUs, but may be appropriately combined with a GPU, an FPGA, or the like.

[0033] The memory 212 includes a RAM, a ROM, or a non-volatile memory, and functions as a storage unit. The ROM stores instructions for executing the application and the OS according to the present disclosure as a program. Such a program is loaded and executed by the processor 211. The RAM is used to write and read data while the program stored in the ROM is being processed by the processor 211. The non-volatile memory is a memory into which data is written and read by the execution of the program, and the data written therein is saved even after the execution of the program is terminated. Specifically, the memory 212 stores a program for the processor 211 to execute the above-mentioned processes, etc.

[0034] The communication interface 213 functions as a communication unit that transmits and receives information between the management device 100 and the business operator terminal device 200-2 that are electrically connected via a communication processing circuit. The communication processing circuit processes the programs and various information used in the processing system 1 to transmit and receive information according to the progress of processing. The communication processing circuit processes based on a wideband wireless communication method such as the LTE method, but it is also possible to process based on a method related to narrowband wireless communication such as wireless LANs such as IEEE802.11 and Bluetooth (registered trademark) or a method related to non-contact wireless communication. It is also possible to use wired communication instead of or in addition to wireless communication.

[0035] The input interface 214 functions as an input unit that accepts operation inputs from a user or a business operator to the terminal device 200. Examples of the input interface 214 include various hard keys such as a keyboard and a mouse, and a touch panel that is superimposed on the display of a display device and has an input coordinate system corresponding to the display coordinate system of the display. In the case of a touch panel, icons corresponding to commands to be input are displayed on the display, and the user or business operator performs operation input via the touch panel to select each icon. The detection method of the operation input by the touch panel may be any method such as a capacitive type or a resistive film type. The input interface 214 does not always need to be physically provided in the terminal device 200, and may be connected as necessary via a wired or wireless network.

[0036] The output interface 215 functions as an output unit for outputting various information. An example of the output interface 215 is an interface for connecting to an external device or equipment such as a display device configured with a liquid crystal panel, an organic EL display, a plasma display, or the like. However, when the terminal device 200 itself has a display, the display can function as the output interface. In addition, when the terminal device 200 is connected to a display device or the like via a communication interface 213, the communication interface 213 can also function as the output interface 215.

[0037] 4. Configuration of the model generating device 300 Although the specific configuration of the model generating device 300 is not specifically shown in the figure, it includes components necessary for executing each process, such as a processor, memory, and a communication interface, and each component is electrically connected to each other via a control line and a data line.

[0038] Although the model generating device 300 is configured as a separate entity from the management device 100 in this embodiment, the functions executed by the model generating device 300 may be executed by the management device 100 instead.

[0039] 5. Information stored in the management device 100 3A is a diagram conceptually illustrating a property management table stored in management device 100 according to an embodiment of the present disclosure. Information stored in the property management table is updated and stored as needed in accordance with the progress of processing by processor 111 of management device 100.

[0040] 3A, the property management table stores equipment information, home appliance information, location information, document information, property information, user information, etc., in association with property ID information. "Property ID information" is information unique to each property and is information for identifying each property. The property ID information is generated by the management device 100, for example, every time a request for registering a new property is received from the business operator terminal device 200-2 to the management device 100 by a business operator.

[0041] "Equipment information" is information for identifying equipment installed in each property, such as kitchen equipment and shared equipment. Specifically, equipment ID information for identifying each piece of equipment is given to each piece of equipment, and the equipment information stores the equipment ID information for one or more pieces of equipment installed in each property. "Electrical appliance information" is information for identifying electrical appliances installed in each property, such as televisions, air conditioners, and washing machines. Specifically, electrical appliance ID information for identifying each electrical appliance is given to each piece of equipment, and the electrical appliance information stores the electrical appliance ID information for one or more electrical appliances installed in each property.

[0042] "Location information" is information that identifies the address or location of each property. The location information may be any information that can identify a location, such as coordinate information or an address. The location information does not need to be information that can identify a property, such as an address, and may be information that can identify an approximate location of a municipality or district, or information that indicates the time required from a landmark such as the nearest station. "Document information" is information that identifies documents created in relation to a property, such as documents that guide businesses such as gas, electricity, and water, property management regulations documents, insurance-related documents such as earthquake insurance and fire insurance for the property, contract documents for telephone and internet, contract documents for the property such as a rental contract exchanged with the property business, copies of important information explanation documents for the property, notice of cancellation of the property, contact point for the business that manages the property, check results of a checklist for moving into the property created by a user who is a resident at the time of moving in, instructions for using shared facilities such as a parking lot, bicycle parking lot, and guest room, documents informing the user of the method of separating garbage and collection dates for the area where the property is located, and documents informing the evacuation site and evacuation procedures in the event of a disaster. Specifically, document ID information for identifying each piece of document information is given to each piece of document information, and one or more pieces of document ID information related to each property are stored in the document information.

[0043] "Property information" is information indicating the attributes of each property. Examples of the property information include the name of the property, the current status of each property (for example, occupied, scheduled to be vacated, accepting tenants, and no tenants), structural information of the property, the date of completion of the property, the contract date of the property, and the contract period of the property. "User information" is information indicating the attributes of the user who uses each property. Examples of the user information include user ID information for identifying the user who uses the property, the user's name, contact information, gender, age or age group, family structure, and place of employment. Note that a user management table for managing user information may be provided separately, and the user information stored may be only user ID information used when referring to the user management table.

[0044] In addition, various types of information, such as facility information, electrical appliance information, location information, document information, property information, and user information, are stored in correspondence with each property ID information by the management device 100 receiving a registration request for the various types of information from the business operator terminal device 200-2 that has accepted operational input by the business operator.

[0045] 3B is a diagram conceptually illustrating a procedure management table stored in the management device 100 according to an embodiment of the present disclosure. The procedure management table is prepared for each property in association with property ID information. In other words, the information managed in the procedure management table may differ depending on each property.

[0046] The information stored in the procedure management table is information stored as first procedure information, which is procedure data including procedures related to the property or items installed in the property among the procedure information. The information is updated and stored as needed according to the progress of the processing of the processor 111 of the management device 100. The procedure management table stores the procedure information exemplified above in association with each ID information. Note that FIG. 3B shows a case where the procedure information is managed separately into an equipment management table in which the procedure information of each facility is stored, an electrical appliance management table in which the procedure information of each home appliance is stored, a rule management table in which rule information of each municipality or region is stored, and a document management table in which the contents of each document are stored, but it is natural that the procedure management table may be managed in a single table.

[0047] According to FIG. 3B, the equipment management table stores instruction manual information in association with equipment ID information. "Equipment ID information" is information for identifying equipment installed in a property. "Instruction manual information" is information indicating the contents of an instruction manual that includes at least a part of the usage procedure of each equipment installed in a property. The instruction manual information is typically information stored by the management device 100 receiving image data, text data, or a combination of these, which is obtained by reading an instruction manual prepared in paper form in the business operator terminal device 200-2 of the business operator with a camera or scanner, from the business operator terminal device 200-2 in an arbitrary file format. Note that the instruction manual information is not limited to the above, and may be information obtained in an arbitrary file format from, for example, the homepage of the provider of the equipment.

[0048] The electrical appliance management table stores instruction manual information in association with electrical appliance ID information. "Electrical appliance ID information" is information for identifying an electrical appliance installed in a property. "Instruction manual information" is information indicating the contents of an instruction manual, at least a part of which includes the usage procedures of each electrical appliance installed in a property. The instruction manual information is typically information stored by the management device 100 receiving image data, text data, or a combination of these, which is obtained by reading instruction manuals prepared in paper form in the business operator terminal device 200-2 of the business operator using a camera or scanner, from the business operator terminal device 200-2 in an arbitrary file format. Note that the instruction manual information is not limited to the above, and may be information obtained in an arbitrary file format from, for example, the homepage of the provider of the electrical appliance.

[0049] The rule management table stores rule content information in association with rule ID information. The "rule ID information" is information for identifying rules determined in a local government or area specified by the location information of the property. The "rule content information" is information indicating the content of rules determined in a local government or area specified by the location information of the property, and is information including at least a part of the procedure. Examples of such rules include a garbage sorting method, a collection date, and an evacuation site and evacuation procedure in the event of a disaster. The rule content information is typically information stored by the management device 100 receiving image data, text data, or a combination of these data from the business terminal device 200-2 of the business operator, which is read by a camera or scanner from a document in which the rule content is described (for example, a guide document on a garbage sorting method and a collection date, and a guide document on an evacuation site and evacuation procedure in the event of a disaster, etc.) in an arbitrary file format. Note that the rule content information is not limited to the above, and may be information obtained in an arbitrary file format from, for example, a website of a local government or the like that determines the rule.

[0050] The document management table stores content information in association with document ID information. "Document ID information" is information for identifying a document created in relation to a property. "Document content information" is information indicating the content of a document created in relation to a property, and is information that includes at least a procedure in part. Examples of such documents include documents that introduce gas, electricity, water, and other businesses, property management regulations documents, insurance-related documents such as earthquake insurance and fire insurance for the property, contract documents for telephone and Internet, contract documents for the property such as a rental contract exchanged with the property business, a copy of the important information explanation document for the property, a notice of cancellation of the property, an inquiry desk for the business that manages the property, the check results of a checklist created by a user who is a resident at the time of moving in, and instructions for using shared facilities such as a parking lot, a bicycle parking lot, and a guest room. Typically, the document content information is information stored by the management device 100 receiving image data, text data, or a combination of these, obtained by scanning a document (e.g., a rental contract) prepared on paper at the business operator's business terminal device 200-2 by a camera or scanner, from the business operator's terminal device 200-2 in an arbitrary file format. Note that the rule content information is not limited to the above, and may be information obtained in an arbitrary file format from a server device that manages these documents.

[0051] Note that the various information listed in FIG. 3B is at least a part of the procedure information, and other procedure information may naturally be stored.

[0052] 3C is a diagram conceptually illustrating a knowledge management table stored in the management device 100 according to an embodiment of the present disclosure. Information stored in the knowledge management table is updated and stored as needed in accordance with the progress of processing by the processor 111 of the management device 100.

[0053] The information stored in the knowledge management table is information stored as second procedure information, which is past questions and answers to procedures related to the property or items installed in the property, among the procedure information. The information is updated and stored as needed according to the progress of the processing of the processor 111 of the management device 100.

[0054] 3C, the knowledge management table stores question information, property ID information, answer information, etc. in association with knowledge ID information. "Knowledge ID information" is information specific to each question and is information for identifying each question. The knowledge ID information is generated by the management device 100 every time a new question is received from the user terminal device 200-1.

[0055] The "question information" is information indicating the contents of past questions about procedures related to a property or an item installed in the property. Such question information may be the content of the user's question input in the user terminal device 200-1 itself, or may be information after being converted into input information in the procedure-learned model. Therefore, the question information is stored by receiving the question information from the user terminal device 200-1 or by being converted into input information based on the question information. The "property ID information" is information unique to each property and is information for identifying each property. It is possible to identify property information and user information of the property for which a question and answer have been asked based on the property ID information, and to grasp the tendency of questions and answers. The "answer information" is information indicating the contents of answers to past questions about procedures related to a property or an item installed in the property. As such answer information, the contents of the answer acquired as output information from the procedure-learned model are stored. In addition, when the answer information is corrected to a correct answer in the business operator terminal device 200-2 in response to a user's feedback on the answer acquired as output information from the procedure-learned model, the contents of the corrected answer are stored.

[0056] The question information and answer information stored in the knowledge management table are stored by the management device 100 receiving a large number of combinations of question information and answer information received from users from the business operator terminal device 200-2, which are generated in advance by the business operator. The information is also added by the business operator generating answer information when the business operator cannot answer the newly received question information from the user using the procedure-learned model.

[0057] According to FIG. 3D, the feedback management table stores target information and feedback information in association with answer ID information. The "answer ID information" is information specific to the answer information indicating the content of the answer acquired from the procedure-learned model, and is information for identifying each answer information. The "target information" is information indicating information referenced when acquiring each answer information. Specifically, one or more pieces of ID information for identifying information stored in the property management table, the procedure management table, and the knowledge management table are stored. The "feedback information" is information indicating user feedback such as "whether or not the answer information solved" input after the answer information is output in the user terminal device 200-1. For example, when "solved" is input, it means that the information referred to in generating the answer information was correct. Also, when "(because it was not solved) ask a different question" is selected, it means that the information referred to in generating the answer information may have been incorrect. By managing these pieces of information, it is possible to correct the answer information generated in the business operator terminal device 200-2.

[0058] Although not specifically shown in FIGS. 3A to 3D, tables for managing other information, such as a user management table for managing user information, may also be provided.

[0059] 6. Processing sequence performed by processing system 1 4A and 4B are diagrams showing processing sequences executed in the processing system 1 according to an embodiment of the present disclosure. Specifically, FIG. 4A shows a processing sequence executed for processing related to registration of property information and generation of a procedure-learned model, and processing related to new user registration, and FIG. 4B shows a processing sequence executed for processing related to generation of answer information. The processing in each device is executed by a processor processing a program stored in the memory of each device.

[0060] (A) Processing related to registration of property information, etc. and generation of procedure-trained models FIG. 4A shows a process for registering property information and generating a procedure-learned model. The process is performed upon receiving a request from the business operator terminal device 200-2 each time a property to be managed by the management device 100 is newly registered. Specifically, the processor 211 of the business operator terminal device 200-2 receives an operation input from the business operator via the input interface 214 and inputs property information of the property to be newly managed by the management device 100 (S11). As shown in FIG. 3A, the input property information includes equipment information that identifies the equipment to be installed in the property, electrical appliance information that identifies the electrical appliances to be installed in the property, location information that identifies the address and location of the property, document information created in relation to the property, and property information that indicates the attributes of the property. The input may be performed by receiving an operation input for selecting the equipment and electrical appliances to be installed in the property one by one from a list of equipment and electrical appliances prepared in advance via the input interface 214. The input may be performed by, for example, reading document information prepared on a paper medium with a camera or scanner provided in the business operator terminal device 200-2 and converting the information into image data, text data, or a combination of these by the processor 211. The input may also be performed by accepting operational input of property information or location information by the business operator via the input interface 214.

[0061] When the property information is input, the processor 211 of the business operator terminal device 200-2 transmits the input property information (T11) to the management device 100 via the communication interface 213.

[0062] When the processor 111 of the management device 100 receives the property information via the communication interface 113, it generates new property ID information and stores the received property information in the property management table in association with the property ID information (S13). The processor 111 also generates a procedure management table for each property ID information and stores procedure information such as instruction manual information, rule content information, and document content information managed in the procedure management table based on the received property information (S14).

[0063] Specifically, memory 112 of management device 100 stores instruction manual information indicating the contents of the instruction manuals in association with each ID information for a large number of facilities and electrical appliances that may be installed in a previously prepared property as a database. Processor 111 refers to the database based on the facility ID information and electrical appliance ID information stored in the facility information and electrical appliance information of the received property information, and reads out the corresponding instruction manual information. Processor 111 then stores each of the read instruction manual information in association with the facility ID information or electrical appliance ID information in the procedure management table (equipment management table and electrical appliance management table) generated in association with the property ID information of the corresponding property.

[0064] Furthermore, processor 111 generates document ID information for each piece of document information in the received property information, and stores the document ID information in association with the document ID information in the procedure management table (document management table).

[0065] Rule content information such as the garbage sorting method and collection date for each municipality or region, and evacuation sites and procedures in the event of a disaster is stored in advance as a database. Processor 111 refers to the location information in the received property information and reads out rule content information and rule ID information associated with the relevant municipality or region. Processor 111 then stores the read rule content information in a procedure management table (rule management table) generated in association with the property ID information of the relevant property, in association with the rule ID information.

[0066] When the procedure information is stored in the procedure management table, the processor 111 transmits the stored procedure information to the model generation device 300 via the communication interface 113. Although not particularly illustrated, the processor of the model generation device 300 generates a procedure-learned model by having a learner machine-learn the received procedure information. Specifically, the model generation device 300 is a model such as a Generative Pretrained Transformer (GPT) including GPT-1, GPT-2, GPT-3, and GPT-4, a Bidirectional Encoder Representations from Transformers (BERT), a Bidirectional and Auto-regressive Transformer (BART), or a Recurrent Neural Network, and includes a so-called generative AI. Note that, although a language model is given as an example here, it is also possible to use a generative AI model for images, videos, and audio.

[0067] For example, the model generating device 300 constructs a procedure-learned model by a learning method such as supervised learning, unsupervised learning, semi-supervised learning combining supervised learning and unsupervised learning, and reinforcement learning for machine learning algorithms such as nearest neighbor method, naive Bayes method, decision tree, support vector machine, and deep learning using neural network. For example, in supervised learning, the model generating device 300 may obtain a procedure-learned model by providing answer information to procedure information and question information from a user as label information and performing machine learning. In addition, the model generating device 300 may include a procedure-learned model of general-purpose natural language processing such as a large-scale language model that has learned a huge amount of data.

[0068] In this embodiment, the model generating device 300 inputs, as procedure information, the instruction manual information, the rule content information, and the document content information stored in the procedure management table, as well as the information associated with the question information, the answer information, and the property ID information stored in the knowledge management table (property information, user information, etc.) as a huge amount of data into a machine learning algorithm to obtain a procedure learned model. The generated procedure learned model is capable of outputting an appropriate answer as answer information based on the contents of the data set used for learning by inputting input information (e.g., a prompt) generated based on the question information received from the user terminal device 200-1.

[0069] In this embodiment, the model generating device 300 may generate a plurality of procedure-trained models using different data sets to be trained, instead of generating a single procedure-trained model. For example, the model generating device 300 may generate a first procedure-trained model by inputting instruction manual information, rule content information, and document content information (examples of first procedure information) stored in the procedure management table as data sets to be trained, and performing machine learning. In addition, the model generating device 300 may generate a second procedure-trained model by inputting information (property information and user information, etc.) (examples of second procedure information) associated with question information, answer information, and property ID information stored in the knowledge management table as data sets to be trained, and performing machine learning.

[0070] Furthermore, in this embodiment, the processor 111 of the management device 100 provides procedure information to the model generation device 300 via the communication interface 113, and the model generation device 300 generates a procedure-learned model. However, instead of this, the management device 100 itself may generate a procedure-learned model by performing machine learning based on the procedure information.

[0071] In addition, in this embodiment, the model generation device 300 generates a procedure-learned model by machine learning the procedure information provided by the management device 100, but machine learning can also be performed at any time. For example, feedback information for the generated answer information is stored in the feedback management table, and the procedure information used for learning is updated based on the feedback information. The model generation device 300 may perform machine learning again using this updated procedure information to generate a procedure-learned model. By performing this, it becomes possible to generate more accurate answer information.

[0072] This completes the process of registering property information, etc. and generating a procedure-learned model.

[0073] (B) Processing related to new user registration FIG. 4A shows a process for new user registration. The process is performed upon receiving a request from the user terminal device 200-1 each time a user to be managed by the management device 100 is newly registered. Specifically, the processor 211 of the user terminal device 200-1 accepts a user's operation input via the input interface 214 and selects new registration for the service provided via the processing system 1. The user terminal device 200-1 inputs user information required for initial registration via the input interface 214 (S21). Examples of the user information include the user's name, contact information, sex, age or age group, family structure, and place of employment. The processor 211 also accepts a user's operation input via the input interface 214 and inputs information that identifies the property that the user will use by moving in, etc., in addition to the user information. For such information, for example, property ID information and property information stored in advance in the property management table of the management device 100 are used. When the user information is input, the processor 211 transmits to the management device 100 the input user information (T21) via the communication interface 213, information specifying the property, and a new registration request.

[0074] When the processor 111 of the management device 100 receives user information and the like via the communication interface 113, it generates new user ID information and stores the user information in association with the user ID information (S22). Specifically, the processor 111 refers to the property management table and stores the user information in association with information that identifies the property (for example, property ID information). In this way, the association between the property and the user who uses the property is registered.

[0075] This completes the process for new user registration.

[0076] (C) Processing related to generation of response information 4 shows a series of processes related to the generation of answer information. Specifically, when the processor 211 of the user terminal device 200-1 accepts an operational input from the user via the input interface 214, it starts an application program used for the service related to the processing system 1 (S31). When the processor 211 further accepts an operational input from the user via the input interface 214, it inputs preregistered user ID information, a password, etc., and transmits the user ID information (T31), the password, and a login request to the management device 100 via the communication interface 213.

[0077] When the processor 111 of the management device 100 receives the login request, it refers to the user management table to authenticate the user and reads out the user information associated with the user ID information (S32). The processor 111 transmits the read out user information, etc. (T32) via the communication interface 113 to the user terminal device 200-1 that transmitted the login request.

[0078] Next, the processor 211 of the user terminal device 200-1 accepts an operational input from the user via the input interface 214, and inputs question information indicating a question related to the property (S33).

[0079] Here, FIG. 6A is a diagram conceptually illustrating an example of a screen output by the user terminal device 200-1 according to an embodiment of the present disclosure. Specifically, FIG. 6A is a diagram illustrating an example of a screen 10 displayed when question information is input in S33 of FIG. 4B. According to FIG. 6A, the screen 10 includes at least an information display area 13 and a question input area 12 together with a display of "Application A", which is the name of an application program used to provide a service related to the processing system 1. The information display area 13 is an area in which various questions and answers are exchanged between the user terminal device 200-1 and the management device 100, and the exchange of questions and answers is displayed vertically in a chat format so that it can be referenced. In addition, the information display area 13 displays, for example, items 11 that are expected to be highly likely to be input as questions in advance. The processor 211 can input question information by accepting a user's operation input for selecting a desired item from the items 11. In addition, the screen 10 includes the question input area 12 as described above. When processor 211 accepts an operation input to question input area 12, it displays a numeric keypad screen for inputting characters, and allows the user to input question information using any character string.

[0080] In this way, question information is input by selecting a desired item from items 11 on screen 10 or by inputting an arbitrary character string in question input area 12. In the following, a case where "How to use electricity, water, and gas" is input as question information will be described.

[0081] Returning to FIG. 4B, when question information (e.g., "How to use electricity, water, and gas") is input as shown in FIG. 6A, processor 211 of user terminal device 200-1 transmits the input question information (T33) together with user ID information to the management device via communication interface 213.

[0082] When the processor 111 of the management device 100 receives question information or the like, it executes a process of generating input information to be input to the procedure-learned model (S34). Specifically, the processor 111 refers to user information and property information associated with the user ID information based on the received user ID information. Then, the processor 111 adds the user information and property information to the received question information and converts it into input information. For example, when the processor 111 receives question information from a user with user ID information "G1", it refers to the user management table and reads out the age group and family structure as user information. In addition, the processor 111 refers to the property information and reads out the name of the property. The processor 111 adds each piece of information that has been read out to the character string "How to use electricity, water, and gas" received as the question information, and generates input information (prompt) of "I have two 30-year-old dependents and live in property A. Please tell me how to use electricity, water, and gas." Note that the processor 111 does not simply connect each piece of information together, but converts it into a plain text prompt using a language analysis technique.

[0083] When the input information is generated, the processor 111 of the management device 100 inputs the generated input information into the procedure learned model (S35), and acquires answer information from the procedure learned model (S36).

[0084] When the processor 111 of the management device 100 acquires the answer information, it transmits the acquired answer information (T34) via the communication interface 113 to the user terminal device 200-1 that has transmitted the question information.

[0085] When processor 211 of user terminal device 200-1 receives the response information, it outputs the received response information via output interface 215 (S38).

[0086] Here, FIG. 6B is a diagram conceptually illustrating an example of a screen output by the user terminal device 200-1 according to an embodiment of the present disclosure. Specifically, FIG. 6B is a diagram illustrating an example of a screen 20 displayed when answer information is output in S38 of FIG. 4B. According to FIG. 6B, the screen 20 includes at least an information display area 13 and a question input area 12, together with a display of "Application A," which is the name of an application program used to provide a service related to the processing system 1. The information display area 13 includes a display 15 of "How to use electricity, water, and gas," which is question information input by the user as the first question, as well as a display 16 and a display 17 based on the answer information. The display 16 displays the answer generated as the answer information as plain text. The display 17 displays information related to the answer of the display 16 from among the pieces of information stored in an arbitrary format in the procedure management table. In other words, in this embodiment, the response information does not simply include text responses, but also includes image data of the instruction manual for the relevant property or item, excerpts from various information stored as procedural information, and the like, generated in any data format such as CSV or PDF.

[0087] Information display area 13 includes items 18 for selecting feedback information for the answer information. Here, as feedback information, an item "solved" indicating that the generated answer information was correct, and an item "ask a different question" indicating that the generated answer information did not completely solve the problem are displayed. Processor 211 can input feedback information by accepting a user's operation input for selecting a desired item from items 18.

[0088] As described above, screen 20 includes question input area 12. When processor 211 accepts an operational input to question input area 12, it displays a numeric keypad screen for inputting characters, and allows input of additional question information using any character string. In the following, a case will be described in which the option "Ask a different question" is selected as feedback information, and further, "What is the contact information for the water supplier?" is input in question input area 12.

[0089] Returning to FIG. 4 again, the processor 211 accepts an operation input by a user who has referred to the answer information output via the input interface 214, and inputs feedback information (S39). The feedback information is information indicating whether or not the question has been resolved by the generated answer information. That is, the feedback information is used to feed back to the business operator whether or not the generated answer information was correct. For example, as shown in FIG. 6B, when an item "Ask a different question" is selected for the answer information and additional question information "What is the contact information of the water business operator?" is input, the processor 211 of the user terminal device 200-1 stores, as feedback information, that "Ask a different question" has been selected and that "What is the contact information of the water business operator?" has been input as additional question information. Then, the processor 211 transmits the feedback information (T35) to the management device 100 via the communication interface 213 together with answer ID information that identifies the answer information.

[0090] When the processor 111 of the management device 100 receives the feedback information, it stores the received feedback information in the feedback management table in association with the answer ID information received together (S40). At this time, since the feedback information also includes additional question information such as "What is the contact information of the waterworks company?", the processor 111 repeats the process of S34 to S37 again and transmits the generated answer information (T34) to the user terminal device 200-1 via the communication interface 113. The processor 211 of the user terminal device 200-1 repeats the process of S38 to S39 and transmits the feedback information, etc., to the management device 100 again via the communication interface 213 in response to the answer information to the additional question. The processor 111 of the management device 100 executes the process of S40 again for the received feedback information. That is, the processes of S34 to S40 in the management device 100 and the user terminal device 200-1 are repeated every time question information is asked.

[0091] Here, FIG. 6C is a diagram conceptually illustrating an example of a screen output by the user terminal device 200-1 according to an embodiment of the present disclosure. Specifically, FIG. 6C is a diagram illustrating an example of a screen 30 displayed when answer information is output again in S38 of FIG. 4B after additional question information is asked in the feedback information. According to FIG. 6C, the screen 30 includes at least an information display area 13 and a question input area 12 together with a display of "Application A", which is the name of an application program used to provide a service related to the processing system 1. The information display area 13 includes a display 15 of "What is the contact information of the water utility?" which is additional question information, in response to the selection of "Ask a different question" as feedback information in FIG. 6B and the input of additional question information of "What is the contact information of the water utility?". In addition, the information display area 13 includes a display 22 based on answer information generated in response to the additional question information. The display 22 includes answer information in plain text, "The contact information of the water utility is the customer center's '03-xxxx-xxxx'." in response to the additional question information.

[0092] The information display area 13 includes an item 23 for selecting feedback information for the answer information. Here, as feedback information, an item "solved" indicating that the generated answer information was correct and an item "ask a different question" indicating that the generated answer information did not completely solve the problem are displayed. The processor 211 can input feedback information by accepting an operational input from the user for selecting a desired item from the items 23. The screen 30 also includes the question input area 12 as described above. When the processor 211 accepts an operational input to the question input area 12, it displays a numeric keypad screen for character input and inputs additional question information using any character string. In the following, a case where the item "solved" is selected as feedback information will be described.

[0093] 6D is a diagram conceptually illustrating an example of a screen output by the user terminal device 200-1 according to an embodiment of the present disclosure. Specifically, FIG. 6D is a diagram illustrating an example of a screen 40 displayed after feedback information is input. According to FIG. 6D, the screen 30 includes at least an information display area 13 and a question input area 12, together with a display of "Application A," which is the name of an application program used to provide a service related to the processing system 1. The information display area 13 includes a display 25 of "Resolved" in response to "Resolved" being selected as feedback information in FIG. 6C. This marks the end of a session related to a series of questions and answers.

[0094] Returning to FIG. 4B again, when the processor 111 of the management device 100 receives the feedback information as shown in FIG. 6B and FIG. 6C, the processor 111 stores in the feedback management table the feedback information that "Ask a different question" has been selected and the feedback information "What is the contact information of the waterworks company?" as additional question information and the feedback information "Resolved" in association with the answer ID information that identifies each answer information. Then, the processor 111 generates a report to be provided to the business entity that manages the property based on these feedback information (S41). Although not shown, the report includes information for outputting the question information, answer information, target information, user information, property information, equipment information, feedback information, etc., for each answer ID information, based on each information stored in the feedback management table and the knowledge ID information, so that the information can be understood. In addition, even if the input information is generated in the procedure-learned model in S35, there are cases where highly accurate answer information cannot be generated. The report includes information that an answer could not be generated in association with the question information. When the report is generated, the processor 111 transmits the generated report via the communication interface 113 to the operator terminal device 200-2 of the operator as report information (T36).

[0095] When the processor 211 of the business operator terminal device 200-2 receives the report information, it outputs the received report information via the output interface 215 (S42). As described above, the report information includes a report in which the question information, answer information, target information, user information, property information, equipment information, and feedback information are compiled so that they can be understood for each answer ID information, and a report that an answer could not be generated in association with the question information. Therefore, by referring to the report information, the business operator can understand that, for example, when question information about "how to use electricity, water, and gas" is received from a "user who has two dependents aged 30 and lives in property A," it would have been better to provide not only the answer information generated in S36 but also the contact information of the water supplier. Also, when question information of "what is the contact information of the water supplier?" is received, it can be understood that the answer information generated in S36 has been used to solve the problem. Furthermore, the business operator can create correct answer information by referring to question information for which answer information could not be generated. Therefore, processor 211 accepts the business operator's operational input via input interface 214, and corrects the answer information associated with the answer ID information in the knowledge management table to answer information including the water business operator's contact information (S43). Also, processor 211 accepts the business operator's operational input via the input interface, and if the water business operator's contact information is stored in the property management table, stores that the information was correct. Furthermore, processor 211 accepts the business operator's operational input via the input interface, and inputs answer information created by the business operator himself, in association with question information for which answer information could not be generated.

[0096] When the processor 211 of the business operator terminal device 200-2 corrects the information stored in each table such as the response information of the knowledge management table, the processor 211 transmits each piece of updated information (T37) to the management device 100 via the communication interface 213.

[0097] When the processor 111 of the management device 100 receives each piece of corrected information, it updates and stores the property information (S44), and also updates and stores the procedure information in the procedure management table and the knowledge management table (S45). Then, the processor 111 transmits the updated procedure information to the model generation device 300 via the communication interface 113, and re-learns the procedure-learned model based on the updated information. Details of these processes are the same as those described in S13 and S14 of FIG. 4A. This completes the process related to the generation of answer information.

[0098] In this way, the procedure information etc. received from the business operator terminal device 200-2 in S14 of FIG. 4A is provided to the model generation device 300, and a procedure-learned model is generated based on the information. Also, the corrected procedure information etc. received from the business operator terminal device 200-2 in S45 of FIG. 4B is provided to the model generation device 300, and the procedure-learned model is re-learned based on the information. In this way, by using the information received from the business operator terminal device 200-2 that can be used by the business operator as information to be learned, it is possible to use information with a certain level of accuracy, unlike the case of using information input from an unspecified number of people. This makes it possible to further improve the accuracy of the answer information obtained. Also, as described above, the procedure information used for learning is corrected based on the report information including the feedback information, and the procedure-learned model is re-learned using the information. This makes it possible to further improve the accuracy of the answer information obtained.

[0099] In addition, by providing information linked to property ID information, such as user information, which is a user's attribute, and property information, as feedback information, and re-training the procedure-trained model based on this information, it is possible to generate answer information that also reflects the information linked to the property ID information.

[0100] In the above, we have described the case where procedure information, etc. is provided from the business operator terminal device 200-2, but it may also be received from the user terminal device 200-1 or a provider terminal device of a provider that provides goods, etc. related to the property.

[0101] 7. Processing flow of the management device 100 Fig. 5A is a diagram showing a process flow executed in management device 100 according to an embodiment of the present disclosure. Specifically, Fig. 5A is a diagram showing a process flow related to acquisition of answer information executed by management device 100 in S34 to S37 in Fig. 4. This process flow is mainly executed by processor 111 of management device 100 reading and executing a program stored in memory 112.

[0102] As shown in FIG. 5A, the processor 111 receives question information from the user terminal device 200-1 via the communication interface 113 (S111).

[0103] Next, the processor 111 refers to the user management table and the property management table based on the user ID information received together with the question information, and reads out user information and property information. Then, the processor 111 generates input information (prompt) to be input to the procedure-trained model using the read user information and property information and the received question information (S112). As an example of generating input information, when the processor 111 receives question information from a user with user ID information "G1", it refers to the user management table and reads out the age group and family structure as user information. In addition, the processor 111 refers to the property information and reads out the name of the property. The processor 111 adds each piece of information that has been read out to the character string "How to use electricity, water, and gas" received as the question information, and generates input information (prompt) of "I have two dependents aged 30 and live in property A. Please tell me how to use electricity, water, and gas." At this time, the processor 111 does not simply connect each piece of information together, but converts it into a plain text prompt using a language analysis technique. In this example, the age group and family structure are read out as the user information, and the name of the property is read out as the property information, but of course other information may be used.

[0104] Next, the processor 111 transmits the generated input information to the model generating device 300 via the communication interface 113 and inputs it to the procedure learned model. Here, the procedure learned models include a first procedure learned model generated by machine learning the instruction manual information, rule content information, and document content information (examples of first procedure information) stored in the procedure management table, and a second procedure learned model generated by machine learning the information (property information and user information, etc.) (examples of second procedure information) associated with the question information, answer information, and property ID information stored in the knowledge management table. Therefore, the processor 111 first instructs the model generating device 300 to input the input information to the first procedure learned model (S113).

[0105] When the model generation device 300 processes the first procedure trained model to generate output information, the processor 111 acquires the generated output information from the model generation device 300 via the communication interface 113 (S114). At this time, if no procedure information necessary for answering the question information of "how to use electricity, water, and gas" has been learned, for example, the output information may not be generated. In such a case, the processor 111 instructs the model generation device 300 to input the input information to the second procedure trained model (S115). When the model generation device 300 processes the second procedure trained model to generate output information, the processor 111 acquires the generated output information from the model generation device 300 via the communication interface 113 (S116).

[0106] When processor 111 is able to obtain output information from the first procedure trained model as described above, processor 111 generates answer information (first answer information) based on the output information (S117). When processor 111 is unable to obtain output information from the first procedure trained model but is able to obtain output information from the second procedure trained model, processor 111 generates answer information (second answer information) based on the output information. When processor 111 is unable to obtain output information from either the first procedure trained model or the second procedure trained model, processor 111 generates answer information indicating that "No answer was obtained."

[0107] Here, when the output information is acquired from any of the procedure-learned models, the processor 111 may naturally use the acquired output information as it is as the answer information. The processor 111 may also generate answer information by converting the acquired output information. An example of such a conversion process is a process in which the processor 111 appropriately reads out various information such as user information and property information, and adds the read out various information to the output information. Another example of the conversion process is a process in which the processor 111 translates the output information into an arbitrary language. Specifically, when the processor 111 receives question information, it analyzes the language of the question stored as the question information. The processor 111 translates the question information in the process of generating input information. Then, when the processor 111 receives output information from the procedure-learned model, it executes a process of translating the output information into the language obtained by analyzing the question information. Then, the processor 111 generates answer information in which the output information is translated into the language. In this way, by translating the output information to match the language when the question information is input, it is possible to eliminate inconvenience due to language.

[0108] When the answer information is generated as described above, the processor 111 outputs the acquired answer information to the user terminal device 200-1 that has transmitted the question information via the communication interface 113 (S118). This ends the processing flow.

[0109] In this way, in the example of FIG. 5A, answer information is generated using the second step trained model in addition to the first step trained model, but the first answer information based on the output information obtained from the first step trained model is processed preferentially over the second answer information based on the output information obtained from the second step trained model. In other words, when the first answer information is acquired, the second answer information is not acquired. Therefore, it is possible to more efficiently execute the process related to the generation of answer information.

[0110] In the example of Figure 5A, the first answer information based on output information obtained from the first step trained model is processed with priority over the second answer information based on output information obtained from the second step trained model, but it is of course also possible to process the second answer information based on output information obtained from the second step trained model with priority over the first answer information based on output information obtained from the first step trained model.

[0111] Fig. 5B is a diagram showing another process flow executed in the management device 100 according to an embodiment of the present disclosure. Specifically, Fig. 5B is a diagram showing another process flow related to acquisition of answer information performed by the management device 100 in S34 to S37 in Fig. 4, and is a process flow executed in place of Fig. 5A. This process flow is mainly performed by the processor 111 of the management device 100 reading and executing a program stored in the memory 112.

[0112] In Fig. 5A, when the first answer information is obtained from the first trained model, the processing by the second trained model is omitted and the first answer information is preferentially used. In Fig. 5B, the case where the output information from both the trained models is used is described.

[0113] As shown in FIG. 5B, the processor 111 receives question information from the user terminal device 200-1 via the communication interface 113 (S211).

[0114] Next, the processor 111 generates input information by the same process as S212 in FIG. 5A (S212). Details of this process will be omitted. Then, the processor 111 transmits the generated input information to the model generation device 300 via the communication interface 113, and instructs the model generation device 300 to input the input information to the first procedure trained model (S213). When the model generation device 300 processes the first procedure trained model to generate output information, the processor 111 acquires the generated output information from the model generation device 300 via the communication interface 113 (S214).

[0115] Next, the processor 111 instructs the model generation device 300 to input the input information to the second procedure trained model (S215). When the output information is generated by processing the second procedure trained model in the model generation device 300, the processor 111 acquires the generated output information from the model generation device 300 via the communication interface 113 (S216).

[0116] Next, the processor 111 executes a weighting process on the output information acquired from the first procedure trained model and the output information acquired from the second procedure trained model (S217). The weighting process may be performed based on at least one of the contents of the question information, the user information of the user who inputted the question information, the property information of the property used by the user, and a combination thereof. For example, when the question information is a question about an electrical appliance or equipment, the processor 111 weights the accuracy (numerical value) acquired together with the output information of the first procedure trained model by multiplying it by 1.5, and weights the accuracy (numerical value) acquired together with the output information of the second procedure trained model by multiplying it by 0.8. Then, the processor 111 adopts the output information with the higher accuracy after weighting.

[0117] The weighting described above is merely an example, and the coefficients used for weighting may be changed based on at least one of the contents of the question information, the user information of the user who input the question information, the property information of the property used by the user, and a combination of these. Also, the coefficients used for weighting may be always constant regardless of the contents of the question information, etc. (for example, the accuracy of the first step trained model is always weighted 1.2 times).

[0118] When the output information from any of the procedure-learned models is selected by the weighting process as described above, the processor 111 generates either the first answer information or the second answer information based on the selected output information (S218). The process related to the generation of the answer information is similar to S117 in FIG. 5A. Then, when the answer information is generated, the processor 111 outputs the acquired answer information to the user terminal device 200-1 that has transmitted the question information via the communication interface 113 (S219). This ends the processing flow.

[0119] In this way, in the example of Fig. 5B, answer information is generated using the second procedure trained model in addition to the first procedure trained model, and the output information to be adopted is selected by weighting the output information from both procedure trained models. This makes it possible to generate answer information that is more appropriate depending on the situation, such as question information and user information.

[0120] Also, without weighting as in S217, both the first answer information and the second answer information may be generated based on both pieces of output information, and each may be output as answer information to the user terminal device 200-1. Also, without weighting as in S217, answer information may be generated by simply adopting output information with a high degree of accuracy, and output to the user terminal device 200-1.

[0121] As described above, in the present embodiment, it is possible to provide a processing device, a processing program, and a processing method that are capable of more efficiently processing questions from users who use a property.

[0122] 8. Variations In the above, it has been described that the information stored in the various tables of the management device 100 is transmitted from the business operator terminal device 200-2 of the business operator. In this way, it is possible to ensure the accuracy of the stored information. However, in addition to this, various information may be stored in response to an instruction from the user terminal device 200-1. Specifically, the processor 211 of the user terminal device 200-1 transmits an information addition request to the business operator terminal device 200-2 via the management device 100 and the communication interface 213. When the business operator terminal device 200-2 receives the addition request, it acquires instruction manual information and the like based on the electric appliance ID information and facility ID information specified in the addition request, and transmits it to the management device 100. In this way, by storing various information based on a request from a user, it is possible to flexibly respond to requests from users while ensuring the accuracy of the information.

[0123] In the above, a case where information such as contact information of a waterworks company or the like is generated as the answer information has been described. In addition, when the processor 111 of the management device 100 determines that the received question information is a question concerning human life or a question of high urgency, such as "A disaster has occurred, please tell me where to evacuate," the processor 111 of the management device 100 may automatically notify the contact information generated as the answer information of an emergency support request. In such a case, a display indicating a high level of urgency may be output to the user terminal device 200-1.

[0124] In addition, when the answer information includes information such as the contact information of a water supplier, even if the question is not related to human life or is not of high urgency as described above, the processor 111 may generate answer information including link information for sending to the contact information.

[0125] In the above, in order to use the first procedure trained model and the second procedure trained model stored in the model generation device 300, input information is transmitted from the management device 100 and output information is received by the management device 100. However, instead of this, a program for processing the first procedure trained model and the second procedure trained model may be stored in the management device 100, and both processes may be executed in the management device 100.

[0126] In the above, the answer information generated by the first procedure learned model and the second procedure learned model is simply output to the user terminal device 200-1 (for example, S117 and S118 in FIG. 5A, or S218 and S219 in FIG. 5B). However, in addition to this, the generated answer information may be stored in the knowledge management table. Specifically, when the processor 111 receives question information in S111 in FIG. 5A, the processor 111 generates new knowledge ID information and stores the received question information in association with the knowledge ID information. Then, when answer information to the question information is generated in S117 in FIG. 5A, the processor 111 stores the answer information in association with the knowledge ID information. Note that the same processing is possible in the case of FIG. 5B. In this way, the answer information generated by the procedure learned model is further used for learning, so that the data set used for learning can be further enriched. Also, as described in S39 in FIG. 4B, feedback information is provided to the business operator for the generated answer information. Therefore, even if answer information generated by a procedure-trained model is used for learning, it will be confirmed by the operator, and accuracy can be guaranteed.

[0127] The processes and procedures described in this specification can be realized not only by those explicitly described in the embodiments, but also by software, hardware, or a combination of these. Specifically, the processes and procedures described in this specification can be realized by implementing logic corresponding to the processes in a medium such as an integrated circuit, a volatile memory, a non-volatile memory, a magnetic disk, or an optical storage. In addition, the processes and procedures described in this specification can be implemented as computer programs and executed by various computers including a processing device and a server device.

[0128] Although the processes and procedures described herein are described as being executed by a single device, software, component, or module, such processes or procedures may be executed by multiple devices, multiple software, multiple components, and / or multiple modules. Furthermore, although the various information described herein is described as being stored in a single memory or storage unit, such information may be stored in multiple memories provided in a single device or multiple memories distributed across multiple devices. Furthermore, the software and hardware elements described herein may be realized by integrating them into fewer components or breaking them down into more components. [Explanation of symbols]

[0129] 1 Processing System 100 Management device 200 Terminal Equipment 200-1 User terminal device 200-2 Operator terminal device 300 Model Generator

Claims

1. A processing device comprising at least one processor, wherein: the at least one processor: receives question information indicating a question related to the property from a user terminal device available to a user using the property; obtains answer information indicating an answer to the question from a procedure learned model obtained by machine learning a procedure information indicating a procedure related to the property or an article installed on the property based on the received question information; outputs the obtained answer information to the user terminal device; A processing device configured to execute processing therefor.

2. The processing device according to claim 1, wherein the at least one processor is configured to execute processing for converting the received question information into input information to be input to the procedure learned model.

3. The processing device according to claim 2, wherein the question information is converted into the input information based on user information indicating an attribute of the user.

4. The procedure learned model includes a first procedure learned model obtained by machine learning first procedure information which is procedure data including a procedure related to the property or an article installed on the property, and a second procedure learned model obtained by machine learning second procedure information which is past questions and answers for a procedure related to the property or an article installed on the property. The processing device according to claim 1.

5. The processing device according to claim 4, wherein the answer information is such that either the first answer information obtained from the first procedure learned model or the second answer information obtained from the second procedure learned model is preferentially processed.

6. The processing device according to claim 5, wherein when the first answer information is obtained from the first procedure learned model, the first answer information is processed as the answer information.

7. The processing device according to claim 5, wherein either the first answer information or the second answer information is preferentially processed by weighting the output from the first procedure learned model and the output from the second procedure learned model.

8. The processing device according to claim 1, wherein the at least one processor is configured to execute processing for obtaining feedback information from the user terminal device regarding the answer information by the user.

9. The processing device according to claim 8, wherein the learned procedure model is updated by performing machine learning based on the feedback information.

10. In a computer including at least one processor, cause the at least one processor to receive question information indicating a question related to the property from a user terminal device available to a user who uses the property, obtain answer information indicating an answer to the question from a learned procedure model obtained by performing machine learning on procedure information indicating a procedure related to the property or an article installed in the property based on the received question information, output the obtained answer information to the user terminal device, A processing program that functions to execute processing for the purpose.

11. In a computer including at least one processor, a processing method executed by the at least one processor, receiving question information indicating a question related to the property from a user terminal device available to a user who uses the property; obtaining answer information indicating an answer to the question from a learned procedure model obtained by performing machine learning on procedure information indicating a procedure related to the property or an article installed in the property based on the received question information; outputting the obtained answer information to the user terminal device; A processing method including.

Images