Information management system, information management method, and information management device

Abstract

We provide an information management system, information management server, and information management method that effectively utilize the business information held by each location. [Solution] In the information management system 1, the information management server 40 stores attribute data of each site 10 where a predetermined task is performed and multiple answer data corresponding to question data related to the predetermined task, and outputs a related answer extraction model in which the portion of the multiple answer data related to that site is input. The server receives answer data corresponding to question data common to each information management device related to the predetermined task from each information management device 30 associated with each of the multiple sites where the predetermined task is performed, inputs the attribute data of the specified site and each of the received answer data into the related answer extraction model, outputs the portion of the answer data from the multiple sites related to the specified site, and transmits data created based on the output data to the information management device related to the specified site.

Description

Technical Field

[0001] The present invention relates to an information management system, an information management method, and an information management apparatus.

Background Art

[0002] Due to changes in the social situation and the evolution of information processing technology, the amount of business data that companies need to manage is increasing day by day, and the number of stakeholders who share business information is also increasing.

[0003] As a technology for sharing and managing such business information among stakeholders, for example, in Patent Document 1, account information of companies and employees is acquired, and first management object information regarding a management object that exists for classification purposes is acquired. At the same time, an information acquisition unit that acquires second management object information regarding a management object in which there are a plurality of classification items when classifying the management object by an arbitrary classification item, and account information and the first management object information are associated with each of a plurality of classification items constituting the second management object information. A security information management system including a security management unit that generates security management information to be registered in a database unit is disclosed. Desired information can be searched based on the database created in this way.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] On the other hand, with the recent rapid globalization, business is often conducted at a large number of business bases (factories, sales offices, offices, etc.), and information sharing between these bases has become important.

[0006] However, even for similar tasks, operational policies, know-how, and specific work details are often established independently at each location, making it cumbersome to integrate and share information between locations. Furthermore, the sheer volume of information to be integrated is enormous, meaning that searching for data from other locations at one location could result in a large amount of irrelevant information being included in the search results, potentially increasing workload. Conversely, there was also a risk of overlooking useful information.

[0007] This invention has been made in view of these circumstances, and its purpose is to provide an information management system, an information management method, and an information management device that can effectively utilize the business information held by each location. [Means for solving the problem]

[0008] One aspect of the present invention for solving the above problems is an information management server comprising: a storage device that stores a related answer extraction model, which is a trained model that takes attribute data of a location where a predetermined task is performed and a plurality of answer data corresponding to question data relating to the predetermined task as inputs, and outputs the portion of the plurality of answer data related to the location; an information management server that performs an answer data reception process that receives answer data corresponding to question data common to each of the information management devices relating to the predetermined task from each of the plurality of locations where the predetermined task is performed; an answer data aggregation process that outputs the portion of the answer data of the plurality of locations related to the designated location by inputting the attribute data of a designated location and the received answer data into the related answer extraction model; and a know-how data transmission process that transmits data created based on the output data to the information management device relating to the designated location.

[0009] Furthermore, another aspect of the present invention for solving the above problems is an information management system comprising a plurality of information management devices, each associated with a plurality of locations where a predetermined task is performed, and an information management server that is communicably connected to the plurality of information management devices, wherein each of the information management devices includes a storage device that stores a question-answer model, which is a trained model into which question data relating to the predetermined task at the location associated with the information management device is input and answer data corresponding to the question data is output, and a computing device that performs a question-answer processing that inputs question data common to each of the information management devices relating to the predetermined task at the location associated with the information management device into the question-answer model, outputs answer data corresponding to the question data, and transmits the output answer data to the information management server, The information management server is an information management system comprising: a storage device that stores a related answer extraction model, which is a trained model that receives attribute data of the location where the predetermined business is performed and multiple answer data corresponding to question data related to the predetermined business, and outputs the portion of the multiple answer data related to the location; and a computing device that performs an answer data reception process that receives the outputted answer data from each of the information management devices, an answer data aggregation process that inputs the attribute data of a specified location and each of the received answer data into the related answer extraction model to output the portion of the answer data of the multiple locations related to the specified location; and a know-how data transmission process that transmits data created based on the outputted data to the information management device related to the specified location. [Effects of the Invention]

[0010] According to the present invention, business information held by each location can be effectively utilized.

[0011] Other configurations and effects will be clarified by the following description of the embodiments. [Brief explanation of the drawing]

[0012] [Figure 1]This figure shows an example of the configuration of the information management system according to this embodiment. [Figure 2] This diagram illustrates the information stored by each information management device and an example of the functional components of each information management device. [Figure 3] This figure shows an example of the information stored by the information management server and the functional units that the information management server possesses. [Figure 4] This figure shows an example of the hardware configuration of each information processing device in the information management device and the information management server. [Figure 5] This is a sequence diagram illustrating an example of a process performed in an information management system. [Modes for carrying out the invention]

[0013] Hereinafter, one embodiment of the present invention will be described in detail with reference to the drawings. However, the present invention is not to be construed as being limited to the embodiments described below. It will be easily understood by those skilled in the art that the specific configuration can be modified without departing from the spirit or intent of the present invention. Furthermore, in the configurations described below, the same reference numerals are used in common between different drawings for the same parts or parts having similar functions, and redundant explanations may be omitted.

[0014] Figure 1 shows an example of the configuration of the information management system 1 according to this embodiment. The information management system 1 is applied to predetermined operations (hereinafter referred to as "the operations"; in this embodiment, these are assumed to be product manufacturing operations, but this is not intended to limit them) performed at multiple locations 10 (for example, areas such as sales offices, business offices, factories, etc., or areas managed by these areas). Each location 10 is equipped with equipment 20 used for the operations (for example, equipment for manufacturing products, a camera for photographing equipment or products, and a server for storing various data used in the operations) and an information management device 30.

[0015] Furthermore, the information management system 1 includes an information management server 40 that manages the data confidentiality policies (details described later) at each of the 10 locations.

[0016] The information management device 30 manages data regarding the operations that are being or have been performed at its base 10 (hereinafter referred to as operation data). Specifically, the information management device 30 acquires data from the equipment 20 at a predetermined timing (for example, a predetermined time, a predetermined time interval, etc.), and stores the acquired data as operation data. The equipment 20 is, for example, manufacturing equipment for products, sensors provided in the manufacturing equipment, or an operation management system. Details of the operation data will be described later.

[0017] Between the equipment 20 and the information management device 30 at each base 10, between the information management devices 30 between the bases 10, and between the information management device 30 at each base 10 and the information management server 40, they are communicably connected by a wired or wireless communication network 5 such as the Internet, a LAN (Local Area Network), a WAN (Wide Area Network), or a dedicated line.

[0018] FIG. 2 is a diagram for explaining an example of the information stored in each information management device 30 and the functional units possessed by each information management device 30. First, each information management device 30 stores attribute data 31, a question-and-answer model 32, and operation data 34.

[0019] The attribute data 31 is data representing various attributes or characteristics related to the base 10 to which the corresponding information management device 30 belongs. The attribute data 31 is, for example, an identifier of the base 10, the name of the base, the equipment 20 possessed by the base 10, the products manufactured by the equipment 20, the manufacturing processes required for manufacturing the products, or the manufacturing methods of the products, but is not limited to these. The attribute data 31 may vary greatly depending on the base. For example, even when manufacturing the same product, the components constituting the product and the manufacturing methods employed may differ depending on the base 10.

[0020] The question-and-answer model 32 is a trained model into which data of questions regarding the main business in the base 10 related to the corresponding information management device 30 is input, and data of answers corresponding to the question data is output. In the present embodiment, the question-and-answer model 32 is a trained model into which data of questions regarding the situation in the main business at the base (for example, data of questions asking about the method for resolving problems occurring in the main business at the base) is input, and data of countermeasures for the situation (for example, data of the specific content of the method for resolving problems) is output. The question-and-answer model 32 may be an existing publicly available natural language processing model (such as a large language model (LLM)), or may be a newly created trained model.

[0021] In the present embodiment, it is assumed that the question data of the question-and-answer model 32 is data of natural language, image, or video, while the answer data is data of natural language. That is, for the question-and-answer model 32, natural language data of questions asking about the main business in the base 10 is input, and natural language data of answers corresponding to the question data is output. Alternatively, for the question-and-answer model 32, data of questions asking about an image or video acquired in the main business in the base 10 may be input, and natural language data of answers to the questions (improvement points for the business problems indicated by the image or video) may be output.

[0022] In this embodiment, the question answering model 32 is a neural network having an input layer into which question data is input, one or more hidden layers that extract and output features from the question data, and an output layer that outputs one or answer data from the features. However, it is not limited to this. Examples of neural networks include CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), BERT (Bidirectional Encoder Representations from Transformers), XLNet, and GPT (Generative Pre-trained Transformer).

[0023] The business data 34 is used as training data for the question answer model 32. Specifically, the business data 34 includes data on past situations in the business at the corresponding site 10 (training data for question data) and data on countermeasures for each past situation (training data for answer data), which are obtained from the equipment 20. The business data 34 may be, for example, natural language data, image data, or video data.

[0024] If equipment 20 is a product manufacturing facility, the business data 34 is data representing the attributes of the manufacturing facility, the operation of the manufacturing facility (manufacturing methods such as forging, casting, and assembly methods), the parts used in product manufacturing, or the manufactured product (its attributes, type, size, quality, etc.) acquired from predetermined sensors installed on equipment 20. If equipment 20 is a camera, the business data 34 is data of images or videos of the manufacturing facility, product, parts, or worker captured by the camera.

[0025] Furthermore, if equipment 20 is a business management system (server), the business data 34 consists of data obtained from the business management system regarding the progress or process of the business in question, data on defect cases or accident cases that occurred in the product or the manufacturing equipment for the product, data on customer complaint cases, data on the risks that may arise from each case, or data on the causes or countermeasures (improvement measures) for each case.

[0026] Next, each information management device 30 is equipped with a learning unit 35, an inquiry unit 36, and a question answering unit 37.

[0027] The learning unit 35 trains the question answering model 32 using business data 34. Since the training data for the question answering model 32 differs at each location 10, the content of the trained question answering model 32 also differs at each location 10.

[0028] The inquiry unit 36 ​​transmits data (natural language data, image data, or video data) regarding the business in question to each of the other information management devices 30 via the information management server 40. Note that this question is common to all information management devices 30.

[0029] When the question answering unit 37 receives the above-mentioned question data from another information management device 30, it inputs the above-mentioned question data relating to the business at the location associated with the information management device 30 related to the question answering unit 37 (hereinafter referred to as "the local location") into the question answering model 32, outputs answer data corresponding to that question data, and transmits the outputted answer data to the information management server 40.

[0030] Next, Figure 3 shows an example of the information stored by the information management server 40 and the functional units provided by the information management server 40. First, the information management server 40 stores the related answer extraction model 42 and the confidentiality determination model 43.

[0031] The related response extraction model 42 is a trained model that takes as input the attribute data of the local site where the business is performed and multiple response data from multiple information management devices 30 corresponding to the above-mentioned question data regarding the business, and outputs the portion of the multiple response data that is related to the local site (hereinafter referred to as local site-related response data).

[0032] In this embodiment, the related response extraction model 42 is a neural network having an input layer into which attribute data and response data are input, one or more intermediate layers (hidden layers) that extract and output features from the attribute data and response data, and an output layer that outputs local location-related response data from the features, but it is not intended to be limited to this. Examples of neural networks include CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), BERT (Bidirectional Encoder Representations from Transformers), XLNet, and GPT (Generative Pre-trained Transformer).

[0033] The related response extraction model 42 may be an existing natural language processing model (such as an LLM) or a newly created pre-trained model. If the related response extraction model 42 is a newly created pre-trained model, it is created, for example, as follows: The information management server 40 acquires attribute data 31 or business data 34 from each site 10, and performs machine learning using the acquired attribute data 31 or business data 34 to create a pre-trained model that takes the attribute data of its own site and multiple response data as input and outputs the part of the multiple response data that is relevant to its own site, which is the site-related response data.

[0034] The confidentiality determination model 43 is a trained model that receives local location-related response data as input and outputs data from the local location-related response data that should be kept confidential at a predetermined location 10. While each information management device 30 may store the confidentiality determination model 43, this embodiment will primarily describe the case where the information management server 30 stores the confidentiality determination model 43.

[0035] In this embodiment, the confidentiality determination model 43 is a neural network having an input layer into which local-site related response data is input, one or more intermediate layers (hidden layers) that extract and output features from the local-site related response data, and an output layer that outputs data to be kept confidential from the features, but it is not intended to be limited to this. Examples of neural networks include CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), BERT (Bidirectional Encoder Representations from Transformers), XLNet, and GPT (Generative Pre-trained Transformer).

[0036] The confidentiality determination model 43 may be an existing natural language processing model (such as an LLM) or a newly created pre-trained model. If the confidentiality determination model 43 is a newly created pre-trained model, it is created, for example, as follows: The information management server 40 acquires confidentiality policy data for each site 10, which is the standard (policy) for confidential information that is prohibited from being provided to any party other than that site 10, as agreed upon by each site 10. The information management server 40 uses this confidentiality policy data to perform machine learning and creates a pre-trained model that takes attribute data and site-related response data for each site 10 as input and outputs data from the site-related response data that should be kept confidential at a given site 10.

[0037] Furthermore, the information management server 40 includes a response data receiving unit 45, a response data aggregation unit 46, a confidentiality determination unit 48, and a know-how data transmission unit 47.

[0038] The response data receiving unit 45 receives response data from each information management device 30 associated with each of the multiple locations 10 where the business is performed, corresponding to the question data common to each information management device 30 related to the business.

[0039] The response data aggregation unit 46 inputs the attribute data of its own location and the response data received from each information management device 30 into the related response extraction model 42, and outputs local location-related response data, which is the portion of the response data from multiple locations that is relevant to its own location.

[0040] The confidentiality determination unit 48 identifies data that should be kept confidential at a predetermined location 10 (specifically, another location 10 that provided the original response data for the know-how data) from the local location-related response data output by the response data aggregation unit 46 using a predetermined algorithm, and creates data (hereinafter referred to as know-how data) by deleting or processing the identified data to be kept confidential from the data output by the response data aggregation unit 46.

[0041] In this embodiment, the confidentiality determination unit 48 identifies data that should be kept confidential at a predetermined location 10 by inputting the local location-related response data output by the response data aggregation unit 46 into the confidentiality determination model 43.

[0042] The know-how data transmission unit 47 transmits data created based on the local-site-related response data output by the response data aggregation unit 46 (specifically, the know-how data created by the confidentiality determination unit 48) to the information management device 30 related to its own site.

[0043] Next, Figure 4 shows an example of the hardware configuration of each information processing device in the information management device 30 and the information management server 40. Each information processing device includes a computing device 51 such as a CPU (Central Processing Unit), a main memory 52 such as RAM (Random Access Memory) or ROM (Read Only Memory), an auxiliary storage device 53 such as an HDD (Hard Disk Drive) or SSD (Solid State Drive), an input device 54 such as a keyboard, mouse, or touch panel, an output device 55 such as a display or touch panel, and a communication device 56 consisting of a NIC (Network Interface Card), wireless communication module, USB (Universal Serial Interface) module, or serial communication module.

[0044] The functions of the functional units of each information processing device described above are realized by the arithmetic unit 51 of each information processing device reading programs from the main memory 52 or auxiliary memory 53. Each program can also be distributed, for example, by recording it on a portable or fixed recording medium. Furthermore, each program in each information processing device may be implemented, in whole or in part, using virtual information processing resources provided using virtualization technology, process space isolation technology, etc., such as a virtual server provided by a cloud system. Also, all or part of these programs may be implemented by services provided by a cloud system via an API (Application Programming Interface), etc.

[0045] Next, we will explain the processes performed in the information management system 1.

[0046] Figure 5 is a sequence diagram illustrating an example of a process performed in the information management system 1.

[0047] First, the information management device 30 at each location 10 receives data input from the administrator or other person at each location 10 regarding the operations being carried out at each location 10 (s1).

[0048] For example, the information management device 30 at each location 10 accepts input of text describing the content of the business (parts, products, equipment) currently being performed or planned to be performed at each location 10 (its own location). Also, for example, the information management device 30 accepts specification of image or video data acquired from the equipment 20.

[0049] The information management device 30 (the local information management device 30) that received the data input in s1 inputs the question data, including the data input in s1, into the local question answer model 32, thereby obtaining the corresponding answer data (local answer data) (s3).

[0050] For example, the information management device 30 generates a prompt using RAG (Retrieval-Augmented Generation) that includes the text entered in s1 and a text asking whether there are any problems or areas for improvement in the business at its own location that the text represents. By inputting the generated prompt into the question answer model 32, it obtains text data representing the problems and areas for improvement at its own location.

[0051] Furthermore, for example, the information management device 30 generates a prompt using RAG that includes an image or video specified in s1 and text asking whether there are any problems or areas for improvement in the business at its own location represented by the image or video, and inputs this into the question answer model 32 to obtain text data representing the problems and areas for improvement at its own location.

[0052] Furthermore, the local information management device 30 acquires attribute data 31 for the local site (s4).

[0053] Furthermore, the local information management device 30 transmits the question data generated in s1 to the information management server 40 (s5). The information management server 40 remembers that the local information management device 30 was specified. In s5, the local information management device 30 may also specify an information management device 30 at another location (a location other than its own, 10) to which the question data will be sent.

[0054] The information management server 40 transmits the question data received from the information management device 30 at its own location to the information management devices 30 at other locations (s7).

[0055] The information management devices 30 at each other location that receive the question data perform the following processing (s9-s13).

[0056] In other words, the information management device 30 at the other location, similar to s3, inputs the received question data into the question answer model 32 at that other location, thereby obtaining the corresponding answer data (data representing the problems and areas for improvement at its own location) (s9).

[0057] Furthermore, the information management device 30 at the other location acquires attribute data 31 of that other location (s11).

[0058] Then, the information management device 30 at the other location transmits the response data and attribute data 31 acquired in s9 and s11 to the information management server 40 along with an identifier indicating the other location (s13).

[0059] Meanwhile, the local information management device 30 also transmits the attribute data 31 of the local site acquired in s4 to the information management server 40 (s14).

[0060] The information management server 40 uses a related response extraction model 42 to extract response data from other locations that are relevant to (useful to) the local location (the local location) from the response data of other locations, based on the question data and attribute data 31 of the local location received from the local location's information management device 30, and the attribute data 31 and response data of other locations received from the other locations' information management devices 30, and aggregates the response data to output local location-related response data (s15).

[0061] For example, the information management server 40 generates a natural language prompt instructing the system to extract response data from other locations that are useful or relevant to the local location, based on the attribute data of the local location and each other location, and inputs the generated prompt into the relevant response extraction model 42.

[0062] Next, the information management server 40 identifies the data related to information that should be kept confidential at other locations from the response data aggregated in s15, that is, the response data related to its own location (s17).

[0063] For example, the information management server 40 identifies data containing information that should be kept confidential at each other location, which is included in the response data related to its own location, by referring to a predetermined database. Alternatively, for example, the information management server 40 transmits the response data related to its own location to the information management device 30 at each other location. The information management device 30 at each other location identifies data containing information that should be kept confidential at that other location, which is included in the response data related to its own location, by referring to a predetermined database, and transmits the identified data to the information management server 40.

[0064] Alternatively, the information management server 40 or the information management devices 30 at each other location may identify the data of information that should be kept confidential at each other location, which is included in the response data related to their own location, by inputting a prompt to the confidentiality determination model 43 instructing it to identify the portion of the response data related to their own location that should be kept confidential. Or, the information management server 40 or the information management devices 30 at each other location may display a screen that accepts the designation of information to be kept confidential and accept the designation input from an administrator or the like.

[0065] Next, the information management server 40 creates know-how data (s19) by deleting the data related to the confidential information identified in s16 from the local-related response data output in s15 or by performing a predetermined masking process.

[0066] Then, the information management server 40 sends the know-how data created in s17 to the information management device 30 at its own site (s21). The information management device 30 at its own site displays the received know-how data and the contents of the response data acquired at its own site in s3 on the screen (s23).

[0067] Furthermore, the local information management device 30 may limit the other locations targeted for local location-related response data to those similar to local location. For example, the local information management device 30 identifies other locations with similar attribute data to local location and sends the identified locations to the information management server 40. The information management server 40 generates a prompt instructing the server to output local location-related response data using the response data of the other locations and inputs the generated prompt into the related response extraction model 42. Here, the method for identifying other locations with similar attribute data to local location may, for example, be to use a predetermined model (a trained model or database that outputs other locations with similar attribute data to local location when local location attribute data is input), or the local information management device 30 may display a screen for selecting other locations with similar attribute data 31 to local location and accept input for the selection of other locations from the local location's administrator or the like.

[0068] As described above, the information management server 40 in this embodiment receives answer data corresponding to question data common to each information management device 30 related to the business in question, from each of the information management devices 30 associated with each of the multiple locations 10 where the business in question is performed. The information management server 40 then inputs the attribute data of its own location and the received answer data into the related answer extraction model 42, outputs the portion of the answer data from the multiple locations 10 that is relevant to its own location (own location-related answer data), and transmits know-how data created based on the own location-related answer data to the information management device 30 related to its own location.

[0069] In other words, the information management server 40 of this embodiment can aggregate response data from each of the other locations 10 (other locations) into data relevant to its own location, output response data related to its own location, and provide know-how data based on this data to its own location.

[0070] As described above, the information management server 40 of this embodiment makes it possible to effectively utilize the business information held by each location 10. In other words, even without centrally managing the business information held by each location 10, each location 10 can obtain appropriate business information (business know-how, etc.) from other locations related to its own location 10, and can find defects and areas for improvement in the business that would be difficult to discover through manual work by humans at its own location.

[0071] For example, in the case of an automotive parts manufacturer, the same parts are produced under different conditions (climate, processing machinery, raw material suppliers, and human skill levels) at each factory around the world. Furthermore, even in the case of the same quality problem, such as a hole appearing during product processing, the cause may differ depending on the manufacturing method (forging, casting, injection molding, etc.). In this way, even when the preconditions differ between one's own factory and other factories, the information management server 40 collects such information as well, compares it with the conditions of one's own factory, and narrows down the vast amount of data from numerous factories and businesses to the data necessary for one's own factory, thereby providing the most suitable answer for one's own factory.

[0072] Furthermore, the information management server 40 in this embodiment outputs related response data for its own location by inputting the natural language attribute data of its own location and the natural language response data received from the information management devices 30 of other locations into the related response extraction model 42.

[0073] By utilizing such pre-trained models based on natural language, it is possible to accurately extract response data from other locations that are related to one's own location.

[0074] Furthermore, the information management server 40 of this embodiment outputs local-site related response data by inputting data on at least one of the equipment owned by the local site, the products manufactured by that equipment, or the manufacturing method of those products, along with the natural language response data received from the information management devices 30 at each other site, into the related response extraction model 42.

[0075] This makes it possible to accurately extract response data from other locations that are relevant to the content of the work at each of the 10 locations.

[0076] Furthermore, the information management server 40 of this embodiment identifies data that should be kept confidential at other locations from among the response data related to its own location using a predetermined algorithm, creates know-how data by deleting or processing the identified confidential data from the response data related to its own location, and transmits the created know-how data to the information management device 30 related to its own location.

[0077] This prevents other locations from providing information that should not be provided to one's own location (such as confidential data). This prevents each location 10 from being reluctant to provide information to other locations 10 on the grounds that it possesses confidential information, and enables the active sharing and utilization of business information to an appropriate extent.

[0078] Specifically, the information management server 40 of this embodiment identifies data that should be kept confidential at its own location by inputting the response data related to its own location into the confidentiality determination model 43.

[0079] This allows for the accurate identification of confidential data for each other location.

[0080] For example, within a company's own factory, the sharing of certain information (e.g., information on processing accuracy) between factories may be restricted by contracts with client companies, even within the same company. In this case, information that could potentially contain such information could not be shared between factories. However, if a confidentiality determination model 43 determines whether or not information can be shared, humans only need to check the results of the confidentiality determination model 43, and each factory can share information more broadly and appropriately. In particular, when multiple companies (factories) collaborate to manufacture a product, the restrictions on information sharing become even stricter. For example, a parts manufacturer that has been contracted to perform work will not disclose information such as quality or the operating status of the production line to the finished product manufacturer that commissioned the work. Therefore, if there is a downtime in the production line of the parts manufacturer, even if the client ordering the parts wants to increase their finished inventory at that time, their needs will not match unless both parties have a forum for discussion. Even in such cases, the information management server 40 processes confidential information using the confidentiality determination model 43, etc., which makes it possible to share information when an agreement is reached between the ordering party and the parts manufacturer, thereby enabling the utilization of business opportunities that were previously missed.

[0081] Furthermore, in the information management system 1 of this embodiment, each information management device 30 inputs question data common to each information management device related to the business at the base 10 associated with the information management device 30 into the question answer model 32, outputs answer data corresponding to the question data, and transmits the outputted answer data to the information management server 40. The information management server 40 then receives the answer data from each information management device 30 and creates know-how data as described above based on the received answer data.

[0082] In this way, each location 10 is equipped with a question answering model 32, allowing each location 10 to obtain answer data based on its own operational know-how.

[0083] In this embodiment of the information management system 1, the question-answer model 32 is a model in which natural language question data related to the business in question is input, and natural language answer data corresponding to the question data is output.

[0084] In this way, because the question-answering model 32 uses natural language data as input and output data, administrators and others can ask precise questions and obtain useful answers to them.

[0085] Furthermore, in the information management system 1 of this embodiment, the question answer model 32 is a model in which question data of questions about images or videos acquired in the business is input, and natural language answer data corresponding to the questions is output.

[0086] Thus, because the question-answering model 32 uses image or video data of the work in question as input data and natural language data as output data, it can accurately reflect the situation of the work in question in the questions and obtain accurate answers that correspond to them.

[0087] Traditionally, information sharing between the 10 locations involved personnel from other factories or offices visiting other factories to observe the site and identify areas for production improvement. Therefore, even with remote monitoring using cameras, it was impossible to ensure that problems were not constantly overlooked. This applies not only to productivity improvements but also to safety inspections. Potential hazards occur even when monitoring is not in progress. Furthermore, in overseas factories with high employee turnover, it is difficult to consistently provide and enforce safety training. However, according to the configuration of the above-mentioned question-and-answer model 32, effective monitoring can be conducted at each of the 10 locations from the perspectives of product productivity and safety.

[0088] Specifically, the information management device 30 at each base 10 in this embodiment generates a question-answer model 32 that receives question data regarding the situation in the business at the base 10 and outputs data for countermeasures for that situation, based on data representing past situations in the business at the base 10 and data for countermeasures for those situations.

[0089] In this way, the question-answering model 32 is trained using past business data for each location 10, so that each location 10 can generate a question-answering model 32 that is relevant to its actual business performance.

[0090] The present invention is not limited to the embodiments described above, and can be implemented using any components without departing from its essence. The embodiments and modifications described above are merely examples, and the present invention is not limited to these as long as the features of the invention are not impaired. Furthermore, although various embodiments and modifications have been described above, the present invention is not limited to these. Other embodiments conceivable within the scope of the technical idea of ​​the present invention are also included within the scope of the present invention.

[0091] For example, some of the hardware provided in each device of each embodiment may be provided in other devices.

[0092] Furthermore, each program of each device may be provided in other devices, a program may consist of multiple programs, or multiple programs may be integrated into a single program.

[0093] Furthermore, each information management device 30 may input all or any part of the natural language data, video data, and image data into the question answer model 32 to obtain various types of answer data.

[0094] Furthermore, while the confidentiality determination model 43 in this embodiment is a model that identifies confidential portions from the local-site related response data, the confidentiality determination model 43 may also be a model that not only identifies confidential portions but also deletes or processes the confidential portions from the local-site related response data.

[0095] Furthermore, the question answering model 32, the related answer extraction model 42, and the privacy determination model 43 described in this embodiment may be other types of models, such as databases, instead of pre-trained models. [Explanation of Symbols]

[0096] 30 Information management device, 40 Information management server, 35 Learning unit, 36 Inquiry unit, 37 Question answering unit, 45 Answer data receiving unit, 46 Answer data aggregation unit, 47 Know-how data transmission unit, 48 Confidentiality determination unit

Claims

1. A storage device that stores a related answer extraction model, which is a trained model that takes attribute data of a location where a prescribed task is performed and multiple answer data corresponding to question data related to the prescribed task as input, and outputs the portion of the multiple answer data related to the location, and A response data reception process that receives response data corresponding to question data common to each of the information management devices associated with each of the multiple locations where the aforementioned prescribed operations are performed, A response data aggregation process that outputs data related to the specified location from among the response data of multiple locations by inputting the attribute data of the specified location and each of the received response data into the related response extraction model, A computing device that performs a know-how data transmission process, which involves sending data created based on the output data to the information management device at the designated location. An information management server equipped with the following features.

2. The storage device stores attribute data of a location where a predetermined task is performed, and multiple natural language response data corresponding to natural language question data related to the predetermined task, and stores a related response extraction model in which the natural language data of the portion of the multiple response data related to the location is output. In the response data aggregation process, the computing device inputs the natural language attribute data of the specified location and the received natural language response data into the related response extraction model, thereby outputting the portion of the natural language response data from the multiple locations that is related to the specified location. The information management server according to claim 1.

3. The storage device stores a related answer extraction model, which is a trained model, into which data of at least one of the following are inputs: equipment owned by the site where the predetermined business is performed, products manufactured by the equipment, or methods for manufacturing the products, and a plurality of answer data corresponding to question data relating to the predetermined business, and which outputs the portion of the plurality of answer data related to the site. The computing device, in the response data aggregation process, inputs data on at least one of the equipment owned by the designated location, the products manufactured by the equipment, or the manufacturing method of the products, along with the received response data, into the related response extraction model, thereby outputting the portion of the plurality of response data related to the location. The information management server according to claim 1.

4. The aforementioned computing device is From the output data, a predetermined algorithm identifies data that should be kept confidential at a predetermined location, and a confidentiality determination process is executed to delete or modify the identified confidential data from the output data to create new data. In the know-how data transmission process described above, the created data is transmitted to the information management device relating to the designated location. The information management server according to claim 1.

5. The storage device stores a confidentiality determination model, which is a trained model that takes the output data as input and outputs data from the output data that should be kept confidential at the predetermined location. The computing device, in the confidentiality determination process, identifies data to be kept confidential at the predetermined location by inputting the output data into the confidentiality determination model. The information management server according to claim 4.

6. An information management system comprising multiple information management devices, each associated with a multiple location where a prescribed operation is performed, and an information management server that is communicatively connected to the multiple information management devices, Each of the aforementioned information management devices is, A storage device that stores a question-answer model, which is a trained model, into which question data relating to the prescribed tasks at the base associated with the information management device is input, and into which answer data corresponding to the question data is output, and The system includes a computing device that performs question-answering processing by inputting question data common to each of the information management devices related to the predetermined tasks at the locations associated with the information management device into the question-answering model, outputting answer data corresponding to the question data, and transmitting the outputted answer data to the information management server. The aforementioned information management server is A storage device that stores a related answer extraction model, which is a trained model that takes attribute data of the location where the prescribed work is performed and multiple answer data corresponding to question data related to the prescribed work as input, and outputs the portion of the multiple answer data related to the location, and The response data reception process receives the output response data from each of the aforementioned information management devices, A response data aggregation process that outputs data related to the specified location from among the response data of multiple locations by inputting the attribute data of the specified location and each of the received response data into the related response extraction model, The information management device relating to the designated location is equipped with a computing device that performs a know-how data transmission process that transmits data created based on the output data. Information management system.

7. Each of the information management devices stores a question-answer model in which natural language question data relating to the predetermined task is input and natural language answer data corresponding to the question data is output. Each information management device's storage device outputs natural language response data corresponding to the question data by inputting natural language question data from a location associated with the information management device into the question response model during the question answer processing. The information management system according to claim 6.

8. Each information management device's storage device stores a question-answer model into which question data of questions about images or videos acquired in the predetermined business is input, and natural language answer data corresponding to the questions is output. Each information management device's storage device, in the question answering process, inputs question data of a question about an image or video acquired in the predetermined business at a location associated with the information management device into the question answering model, and outputs natural language answer data corresponding to the question data. The information management system according to claim 6.

9. Each information management device's storage device stores data representing past situations in the prescribed operations at the location associated with the information management device, and data on countermeasures for each of the past situations. Each of the aforementioned information management devices' computing units executes a learning process to generate a question-answer model based on the stored data, in which question data regarding the situation in the predetermined work at the location associated with the information management device is input, and data on countermeasures for the situation is output. The information management system according to claim 6.

10. An information management method using an information management server comprising a storage device and a computing device, wherein the storage device stores a related answer extraction model, which is a trained model that stores attribute data of a location where a prescribed task is performed and a plurality of answer data corresponding to question data related to the prescribed task, and outputs the portion of the plurality of answer data related to the location, the storage device and the computing device, The aforementioned computing device A response data reception process that receives response data corresponding to question data common to each of the information management devices associated with each of the multiple locations where the aforementioned prescribed operations are performed, A response data aggregation process that outputs data related to the specified location from among the response data of multiple locations by inputting the attribute data of the specified location and each of the received response data into the related response extraction model, The system performs a know-how data transmission process, which involves sending data created based on the output data to the information management device of the designated location. Information management method.

Images