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

Abstract

This information processing device comprises: a control unit having a retrieval-augmented generation module that generates an inference request to be input to a machine trained large language model, such generation performed on the basis of input information including entity identification information with which it is possible to identify individual entities; and a transmission / reception unit that transmits the inference request to the large language model and receives inference results from the large language model. The retrieval-augmented generation module generates an inference request that uses information which does not directly include the entity identification information but is associated with the entity identification information, such generation performed on the basis of: first acquisition results which are output from a first information processing module and which include first information that is bound to an individual entity identified by the entity identification information; and second acquisition results obtained by searching a second information repository in which second information including information other than the first information is accumulated. The retrieval-augmented generation module executes processing in the first information processing module before searching the second information repository, and sets the search conditions for the second information repository on the basis of the first acquisition results which were obtained.

Description

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

[0001] This invention relates to an information processing device, information processing program, information processing system, and information processing method that belong to the fields of medical information technology and artificial intelligence technology, and more particularly to an information processing device, information processing program, information processing system, and information processing method that utilize personal health checkup data, lifestyle data, genetic information, etc., and provide personalized health management and disease prevention advice using a large-scale language model (LLM) and Retrieval-Augmented Generation (RAG).

[0002] As an information processing technology, Patent Document 1 describes a similar document search device comprising: a search target document storage means for storing a plurality of search target documents; a search key document storage means for storing a plurality of search key documents that serve as keys for deriving the search target documents; a search means for searching for a predetermined document based on an input search string; an output means for outputting the predetermined document found by the search means; a search example storage means for storing a plurality of search example sentences; and a registration means for registering a new search example sentence in the search example storage means, wherein the search means calculates the similarity of a combination of the search key document and the search target document based on the similarity between the input search string and each search key document stored in the search key document storage means, and the calculated combination of the search key document and the search target document A similar document search device is disclosed, characterized in that it searches for a combination of a predetermined search key document and a predetermined search target document based on the similarity to the previous document, the output means outputs the searched predetermined search target document, the registration means registers the predetermined search key document as a new search example in the search example storage means based on user instructions if the output predetermined search target document is desired by the user, the search means searches for search examples in descending order of similarity based on the similarity between the input search string and each search example stored in the search example storage means based on user instructions if the output predetermined search target document is not desired by the user, the output means outputs the search examples in descending order of similarity, and the registration means registers a search sentence edited based on the output search examples as a new search example in the search example storage means.

[0003] As an artificial intelligence technology, Non-Patent Document 1 shows the effects of large language models in few-shot learning. From the perspective of obtaining desirable inference results, it is preferable for a large language model to include appropriate information in an inference request (prompt). For this purpose, a technique called Retrieval-Augmented Generation may be introduced, which is a method of searching for relevant information from an external database and using that information to generate an answer. Non-Patent Document 2 is known as technical information related to Retrieval-Augmented Generation.

[0004] The integration of medical information technology and artificial intelligence technology has been rapidly advancing in recent years. For example, in Patent Document 1, for a large language model that has undergone pre-training with 500 GB or more of text, a machine learning unit that performs machine learning for generating the advice using, as learning data, a combination of text indicating the patient's condition and advice utilized for the patient's treatment, an acquisition unit that acquires text (first text) indicating the patient's condition, and a generation unit that inputs the first text into the large language model and causes the large language model to generate advice utilized for the patient's treatment are provided. The machine learning unit performs machine learning using, as learning data, a combination of the medical records of the subjects in a clinical trial and the number of subjects required for the clinical trial. The acquisition unit acquires text (sixth text) of a plurality of the medical records. The generation unit inputs the sixth text and generates advice indicating the number of subjects required for the clinical trial related to the medical record. A medical support device has been proposed.

[0005] In addition, Non-Patent Document 3 discloses a technique for providing health risk prediction information as a tool for health management using an AI system that has learned university medical record information.

[0006] Japanese Patent No. 7300237 Japanese Patent No. 7454090

[0007] Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, Scott Gray, Benjamin Chess, Jack Clark, Christopher Berner, Sam McCandlish, Alec Radford, Ilya Sutskever, Dario Amodei, Language Models are Few-Shot Learners, https: / / arxiv.org / pdf / 2005.14165, arXiv:2005.14165, 22 Jul 2020. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Kuttler, Mike Lewis, Wen-tau Yih, Tim Rocktaschel, Sebastian Riedel, Douwe Kiela, Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks, https: / / arxiv.org / pdf / 2005.11401, arXiv:2005.11401, 12 Apr 2021. Shinji Ueno, "Specific Results of AI-Based Lifestyle-Related Disease Risk Prediction Expanded to Six Diseases for Health Management and Health Business," Shin Iryo (New Medical Journal), October 2023 issue, p. 90.

[0008] Traditional health management systems fail to fully utilize individual health data and are limited to general health advice. As a result, they cannot adequately address individual health conditions and risks, and the early prediction of diseases and the provision of specific preventive measures are limited. Furthermore, data privacy protection and information security issues remain challenges.

[0009] Therefore, the present invention aims to provide an information processing device and information processing system that can provide personalized health advice, so to speak, that appropriately reflects an individual's health data, while appropriately addressing issues such as data privacy protection and information security, in an information processing device that provides health advice using a large-scale language model.

[0010] The present invention, provided to solve the above problems, includes the following aspects: (1) An information processing device comprising: a control unit having a search extension generation module that generates an inference request to be input to a machine-learned large-scale language model based on input information including personally identifiable information that can identify an individual; and a transmitting / receiving unit that transmits the inference request to the large-scale language model and receives an inference result from the large-scale language model, wherein the search extension generation module generates the inference request influenced by the personally identifiable information based on a first search result obtained by searching a first database in which first information linked to the individual identified by the personally identifiable information is stored, and a second search result obtained by searching a second database in which second information consisting of non-personal information is stored.

[0011] (2) The inference request created by the search extension generation module does not include the personally identifiable information, and is an information processing device as described in (1) above.

[0012] (3) The information processing device according to (1) above, wherein the search extension generation module performs a search of the first database before searching the second database, and sets search conditions for the second database based on the first search results.

[0013] (4) The search conditions for the second database are the information processing device described in (3) above, which does not include the personally identifiable information.

[0014] (5) The information processing device according to (1) above, further comprising a personal information database in which personal information useful for searching the first information linked to the personal identification information is stored, wherein the search extension generation module sets first search conditions for obtaining the first search result based on the personal information obtained by searching the personal information database.

[0015] (6) The information processing device described in (5) above, wherein the first search result does not include the personally identifiable information and does not include the personally related information.

[0016] (7) The information processing device described in (5) above, wherein the personal information includes at least one of health checkup information and social insurance medical fee statement information.

[0017] (8) The information processing device described in (1) above, wherein the first information includes one or more selected from the group consisting of health risk data, stress visualization data, and health risk fluctuation data of the individual.

[0018] (9) The information processing device described in (1) above, wherein the second information includes one or more selected from the group consisting of medical papers, health management-related information, nutrition and diet-related information, exercise function-related information, and physician's findings-related information.

[0019] (10) The information processing device described in (1) above, wherein the large-scale language model is managed in a closed environment with restricted access.

[0020] (11) The first database is managed in a closed environment with restricted access, as described in (1) above.

[0021] (12) The information processing apparatus according to (1) above, wherein the control unit further comprises a display signal generation unit that receives the inference result from the large-scale language model and generates a display signal that includes at least a part of the inference result.

[0022] (1) The information processing apparatus according to (1) above, wherein the control unit further comprises a training data creation unit that creates training data based on the inference request, the inference result, and any modifications to the inference result.

[0023] (14) An information processing system comprising: a first device having input / output functions; a second device having data storage functions; and a third device having a control unit, wherein the first device is capable of communicating with the second and third devices, the second device having a model storage device storing a large-scale language model that has been trained on machine learning; a first database storing first information linked to an individual identified by personally identifiable information input to the first device; and a second database storing second information consisting of non-personal information, the control unit of the third device having a search extension generation module that generates inference requests to be input to the large-scale language model based on information input from the first device, the first device having a transmitting / receiving unit that transmits the inference requests input from the third device to the second device having the large-scale language model and receives inference results from the second device, and the search extension generation module having a search extension generation module that generates inference requests influenced by the first information included in the first search results based on a first search result including the first information obtained by searching the first database and a second search result including the second information obtained by searching the second database.

[0024] (15) The information processing system according to (14) above, wherein at least two of the first device, the second device, and the third device are integrated.

[0025] (16) An information processing system having a plurality of computer devices constituting a peer-to-peer network, wherein at least one of the plurality of computer devices has a search extension generation module that generates an inference request to be input to a machine-learned large-scale language model based on input information, at least one of the plurality of computer devices has a transmitting and receiving unit that transmits the inference request to the large-scale language model and receives the inference result from the large-scale language model, at least one of the plurality of computer devices has an input information generation unit that generates information to be input to the search extension generation module, at least one of the plurality of computer devices has a display signal generation unit that generates a display signal including at least a part of the inference result, and the search extension generation module generates an inference request influenced by the first information included in the first search result based on a first search result including the first information obtained by searching a first database in which first information linked to an individual identified by personally identifiable information is stored, and a second search result including the second information obtained by searching a second database in which second information consisting of non-personal information is stored.

[0026] According to the present invention, the following effects can be obtained: (1) Personalized health management and disease prevention advice can be provided in real time by utilizing diverse personal health data. (2) Users can be provided with highly accurate health predictions and advice by utilizing large-scale language models and search extension generation. (3) Secure management and analysis of data can be performed while protecting user privacy.

[0027] This is a block diagram illustrating an information processing system having an information processing device according to one embodiment of the present invention. This is a flowchart (part 1) illustrating the information processing performed by the information processing device according to one embodiment of the present invention. This is a flowchart (part 2) illustrating the information processing performed by the information processing device according to one embodiment of the present invention. This is a block diagram illustrating a modified example (first modified example) of an information processing system having an information processing device according to one embodiment of the present invention. This is a block diagram illustrating another modified example (second modified example) of an information processing system having an information processing device according to one embodiment of the present invention. This is a block diagram illustrating an information processing system having an information processing device according to a third modified example, which is a further generalization of the second modified example shown in Figure 5. This is a diagram showing the relative positioning of the support tools realized by the information processing system according to one embodiment of the present invention.

[0028] Embodiments of the present invention will be described below with reference to the drawings.

[0029] Figure 1 is a block diagram illustrating the functions of an information processing system having information processing according to one embodiment of the present invention.

[0030] As shown in Figure 1, an information processing system 1000 according to one embodiment of the present invention comprises an information processing device 100, a large-scale language model 200 that has been trained on machine learning, a personal information database (in Figure 1, "database" is abbreviated as "DB") 300, a first database 400 for storing personal information, and a second database 500 for storing non-personal information.

[0031] The information processing device 100 includes an input / output unit 110 for exchanging data with a user interface, a control unit 100C for performing information processing, and a transmission / reception unit 130 for exchanging data with a large-scale language model 200. The control unit 100C includes a search extension generation module 120, a display signal generation unit 140, and a training data creation unit 150.

[0032] The input / output unit 110 receives information from users of the information processing system 1000, which is input into user interfaces such as wearable devices represented by smartphones and personal computers installed in medical institutions. It also outputs the results of information processing performed by the information processing system 1000 to the user interface as data for display to the users of this system. In this specification, the term "display" includes showing images or videos on an image display device, emitting sound from audio output devices such as speakers and earphones, and printing text information and image information using a printing device. The input / output unit 110 may also have user interface functions, in which case it may include input devices such as keyboards and microphones, and output devices such as image display devices, speakers, and printing devices.

[0033] The search extension generation module 120 receives input data, including personally identifiable information that can identify an individual, from the input / output unit 110, and generates inference requests to be input to the machine-learned large-scale language model 200 based on this input. The specific processing performed by the search extension generation module 120 will be described later.

[0034] The transmitting / receiving unit 130 transmits inference requests to the large-scale language model 200 and receives inference results from the large-scale language model 200. It can also transmit training data created by the training data creation unit 150 to the large-scale language model 200. The communication environment between the transmitting / receiving unit 130 and the large-scale language model 200 is not limited. They may be located in a common device and connected by internal wiring, or they may be located in different devices but connected by a dedicated line. An internet connection may exist between the transmitting / receiving unit 130 and the large-scale language model 200. In that case, it is preferable that processing to enhance the level of information security, such as encrypting the communication data, is implemented.

[0035] The display signal generation unit 140 takes the inference result received by the transmission / reception unit 130 as input, creates data for displaying the answer / report to be presented to the user of this system based on the information described in the inference result, and outputs the said data to the input / output unit 110.

[0036] When a user of this system confirms the answer / report generated by the display signal generation unit 140 and presented via the input / output unit 110, and modifies the content of the answer / report, and the modifications are input via the input / output unit 110, the training data creation unit 150 creates training data including the modifications and sends that training data to the large-scale language model 200 via the transmission / reception unit 130.

[0037] The large-scale language model 200 is pre-trained, and when an inference request is input from the transmitting / receiving unit 130, it performs inference processing, generates an inference result, and transmits it to the transmitting / receiving unit 130. The large-scale language model 200 may be provided as open source or as closed source. If it is provided as open source, it is possible to manage the large-scale language model 200 in a closed environment with restricted access, and by managing it in this way, the users of this system can exclusively use the learned content of the large-scale language model 200. As a result, third parties cannot enjoy the results of fine-tuning the large-scale language model 200, and operating it in this way is preferable from the standpoint of enhancing the level of information security.

[0038] The personal information database 300 is a database that stores related information linked to an individual indicated by personally identifiable information entered by the user of this system into the search extension generation module 120 via the input / output unit 110. The type of related information is not particularly limited, but examples include an individual's health checkup information (individual health checkup information 310) and social insurance medical fee statement information issued by a medical institution (receipt information 320), and may include social insurance medical fee information and information on medical history. Because the personal information database 300 contains personal information, and in specific cases an individual's health information, it is managed in a closed environment with restricted access.

[0039] The first database 400 is a database that stores first information linked to an individual as indicated by personally identifiable information. The first database 400 is searched using first search conditions created by the search extension generation module 120, and outputs data including information about the health of the individual indicated by personally identifiable information as first search results. Examples of data included in the first search results include data including health risk prediction information (health risk data 410), data including information that visualizes stress (stress visualization data 420), and data showing the degree of change in health risk (health risk change data 430).

[0040] In one example, the data included in these first search results consists of data accumulated from external systems. Specifically, the health risk data 410 is the calculation result of an externally established health risk prediction system, the stress visualization data 420 is the calculation result of an externally established stress (fatigue level) calculation system, and the health risk fluctuation data 430 is the calculation result of an externally established health risk fluctuation prediction system. The first database 400 may acquire and store the calculation result data from each external system, for example, via a Web API.

[0041] Specific examples of health risk data 410 include results predicting the likelihood of developing diabetes in three, five, and ten years. Specific examples of stress visualization data 420 include data showing which of multiple stages a stress level or mood falls into, data showing cortisol levels which increase when stressed, and quantitative data on fatigue levels evaluated by integrating this data. Specific examples of health risk fluctuation data 430 include data showing the current trend of health risk fluctuations based on health risk data from different time points in the past.

[0042] Since the first database 400 contains first information which is personal information, in one example, it is managed in a closed environment with restricted access. Note that the first information stored in the first database 400 may be personal-related information. In that case, the first database 400 will also have the function of the personal-related information database 300.

[0043] The second database 500 stores second information which is non-personal information. The second database 500 is searched by the second search condition created by the search extension generation module 120, and outputs data including non-personal information as the second search result. Examples of the data included in the second search result are information indicated by a doctor for medical papers 510 of each medical department, information related to health management (health management-related information 520), information related to nutritional diet (nutritional diet-related information 530), information related to motor function (motor function-related information 540), health risk data 410, etc. (examples of doctor's findings, treatment plans, measures to prevent exacerbation, etc. are given, and in this specification, it is referred to as "doctor's findings-related information 550").

[0044] These second information are not information for identifying an individual, but general information related to medical care and health, and most of them can be collected from the Internet. Also, for academic papers such as medical papers 510 of each medical department, if a contract with the organization publishing the academic paper is properly made, the desired papers can be easily collected.

[0045] For example, physician opinion-related information 550 includes comments made by a physician who is a user of a health risk prediction engine 610 (see Figure 4) that outputs health risk data 410 as shown in Non-Patent Document 3. These comments are based on information indicated by the health risk data 410 obtained by inputting the medical and health data of a specific individual (for example, that the current risk of diabetes is the second of three levels) and are used to advise the individual on how to improve their health. Such comments included in physician opinion-related information 550 are linked to specific first information (for example, that the current risk of diabetes is the second of three levels) stored in the first database 400, such as the health risk data 410, and are stored in the second database 500. However, since physician opinion-related information 550 is not linked to information that identifies an individual, it is considered non-personal information.

[0046] Figure 2 is a flowchart (part 1) illustrating the information processing performed by an information processing device according to one embodiment of the present invention, and Figure 3 is the same flowchart (part 2).

[0047] First, the user interface of the user outputs input information data indicating a question containing personally identifiable information to the information processing device 100, and this input information data containing personally identifiable information is input to the input / output unit 110 of the information processing device 100 (step S101). The input / output unit 110 outputs the above input information data to the search extension generation module 120 of the control unit 100C, and upon receiving this input information data, the search extension generation module 120 creates search conditions including personally identifiable information and searches the personal information database 300 using these search conditions. As a result, the personal information database 300 outputs personal information such as individual health checkup information 310 and medical claim information 320 to the search extension generation module 120 (step S102).

[0048] The search expansion generation module 120 creates a first search condition for searching the first database 400 by using the input personal-related information, and searches the first database 400 (step S103). When the data stored in the first database 400 is managed in terms of information for each facility where a health examination or treatment has been received, the search expansion generation module 120 identifies the facility where the target individual has received a health examination or treatment based on the input personal-related information, and adds information regarding the authority to access the information within the facility to the first search condition. The first database 400 outputs, as a first search result, the health information of individuals who satisfy this first search condition (step S104).

[0049] When the first database 400 not only stores data but also has an information processing unit 600 (see FIG. 4) that executes a health risk prediction engine 610 or the like, the search expansion generation module 120 sets the first search condition in step S103 so as to include the health-related information (height, weight, blood glucose level, etc.) of a specific individual reflected by the personal-related information obtained from the personal-related information database 300. In this specification, "setting a search condition" means converting a specific attribute (in the above example, health-related information such as height) into a search query. The first database 400 executes the health risk prediction engine 610 with the health-related information of a specific individual included in the first search condition as input to generate first information such as health risk data 410, and outputs a first search result including the first information of the specific individual thus generated (step S104). The stress visualization data 420 and health risk fluctuation data 430 included in the first search result may also be results of information processing executed by the first database 400 (results of executing a stress visualization engine 620, results of executing a health risk fluctuation prediction engine 630).

[0050] An information processing unit 600 is provided separately from the first database 400, and it has a health risk prediction engine 610 that generates health risk data 410, a stress visualization engine 620 that generates stress visualization data 420, and a health risk fluctuation prediction engine 630 that generates health risk fluctuation data 430. At least a portion of the first search conditions may be input to this information processing unit 600. In this case, the information processing unit 600 performs predetermined information processing, and the data resulting from that processing (health risk data 410, stress visualization data 420, health risk fluctuation data 430) becomes part of the first information, and the first search results including the first information may be output from the information processing unit 600. Such a configuration is shown in Figure 4 as a modified example of the information processing system 1000 according to this embodiment. As mentioned above, the portion enclosed by the dashed line in Figure 4 may be the first database 400.

[0051] The search extension generation module 120 takes the first search result from the first database 400 as input to generate a second search condition and searches the second database 500 (step S105). The first search result includes information that identifies an individual's health status, such as health risk data 410, and the second search condition is set to include such specific health statuses. Therefore, the second search result, which is retrieved using the second search condition and output from the second database 500 (step S106), is more likely to contain information that is useful for a specific health status.

[0052] For example, if the second search criterion includes information indicating that the current risk of diabetes is at the second level of a three-level assessment, the second database 500 will output physician's findings-related information 550, which is effective when the risk of diabetes is at the second level of the three-level assessment, as the second search result. As a result, information that would be effective when the risk of diabetes is at the third level of the three-level assessment, or conversely, information that would be effective when it is at the first level of the three-level assessment, will not be included in the second search result. If the second search result includes information that is not appropriate for the degree of diabetes risk of a particular individual, the content of the inference request generated by the search extension generation module 120 afterward will differ from what it should be. In such cases, the accuracy (degree of effectiveness) of the information included in the inference result will naturally decrease.

[0053] Furthermore, if the second search criterion includes information indicating a moderate risk of diabetes, the physician's findings related information 550 included in the second search result will not consist solely of general information such as "get moderate exercise and get enough sleep" to prevent or worsen diabetes. Such general information is actually detrimental from the perspective of improving the accuracy of the information included in the inference results, as it generalizes the content of the inference request that the search extension generation module 120 generates afterward.

[0054] Thus, although the second information accumulated in the second database 500 is non-personal information, the second search results obtained by searching the second database 500 will contain information that is significant to a specific individual, because the second search conditions are based on the first search results which reflect personal information. In other words, the non-personal information included in the second search results output in step S106 will be biased in content to correspond to a specific individual.

[0055] Using the second search results, which include non-personal information and are biased in content, and the first search results, which include personal health information, as input, the search extension generation module 120 generates an inference request (step S107). The inference request generated by the search extension generation module 120 is output to the transmission / reception unit 130, and the transmission / reception unit 130 transmits the data including the inference request to the large-scale language model 200 (step S108).

[0056] The large-scale language model 200 to which the transmitting / receiving unit 130 transmits is not limited to one type, but may be used to transmit to multiple language models. Furthermore, if the inference request includes information for determining which language model to send it to, the transmitting / receiving unit 130 will transmit the inference request to a predetermined language model identified based on that information.

[0057] The large-scale language model 200 receives an inference request from the transmission / reception unit 130, performs inference processing, and generates an inference result (step S109). The inference result generated by the large-scale language model 200 is transmitted to the transmission / reception unit 130 of the information processing device 100, and the transmission / reception unit 130 receives the inference result (step S110).

[0058] The inference results received by the transmitting / receiving unit 130 may be displayed directly on the user interface, or they may be converted into information that is easier for the user to understand as a primary answer. Specifically, the inference results received by the transmitting / receiving unit 130 are input to the display signal generation unit 140, and the display signal generation unit 140 creates display signals such as text and charts based on the information contained in the inference results, which are used as secondary answers such as answers and reports (step S111). Data including this secondary answer is output to the input / output unit 110, and the user interface receives this data and, based on that data, displays the secondary answer such as answers and reports on an image display device or prints it on a printer (step S112).

[0059] The user, having recognized the inference result output by the information processing device 100 through the above information processing, can input any modifications to the result, such as adding comments, into the user interface (step S113). The input modifications are then input from the user interface to the input / output unit 110 of the information processing device 100 (step S114). These modifications, along with the inference request generated by the search extension generation module 120 and the secondary response from the display signal generation unit 140, are then input to the training data creation unit 150. Based on the input data, the training data creation unit 150 generates training data including the inference request and the inference result reflecting the modifications (step S115), and outputs it to the transmission / reception unit 130. The training data creation unit 150 may also receive a primary response in addition to the secondary response, or it may receive a primary response instead of a secondary response.

[0060] The transmitting / receiving unit 130 outputs training data to the large-scale language model 200 (step S116), and the large-scale language model 200 performs fine tuning using the input training data (step S117).

[0061] If the large-scale language model 200 is located in a separate location from the information processing device 100, inference requests, inference results, and training data are transmitted from the information processing device 100 to the large-scale language model 200 via an open communication environment such as the internet. However, the large-scale language model 200 may be located in the same physical location as the information processing device 100, for example, within the same device. Localizing the large-scale language model 200 in this way can reduce the resources (such as power) required for communication and improve the operational efficiency of the information processing system 1000. Furthermore, it becomes possible to manage the large-scale language model 200 in a closed environment with restricted access. In this case, it becomes difficult for third parties to access the large-scale language model 200, thus reducing the possibility of information leakage compared to using an open communication environment. This is advantageous from the perspective of enhancing the level of information security.

[0062] On the other hand, the localized large-scale language model 200 can only obtain training data from the training data creation unit 150, and therefore its level of learning can only be improved by operating the information processing system 1000. From the perspective of efficiently improving the level of learning, the large-scale language model 200 may have a federated learning function. Specifically, the localized large-scale language model 200 provided by the information processing system 1000 transmits the improvements and local language models generated by the model to a large-scale language model (central language model) that has a common basic configuration and belongs to a central integrated environment.

[0063] The central language model integrates improvements received from the information processing system 1000, local language models, and other large-scale language models (local language models) that are not included in the information processing system 1000 but share a basic structure with the central language model, to generate comprehensive improvements and a global language model. The generated comprehensive improvements and global language model are transmitted from the central integration environment to localized large-scale language models, including the large-scale language model 200, for sharing (updating).

[0064] In this type of associative learning, specific information about the inference requests of individual localized large-scale language models is abstracted away, and only abstracted information is sent to the central integrated environment, thus ensuring the independence of individual large-scale language models (non-sharing of inference requests and inference results). It should be noted that in integrated learning, there may be no central integrated environment, only multiple localized large-scale language models, which share information with each other to improve their learning level.

[0065] The embodiments described above are provided to facilitate understanding of the present invention and are not intended to limit it. Accordingly, each element disclosed in the above embodiments is intended to include all design modifications and equivalents that fall within the technical scope of the present invention.

[0066] Figure 5 is a block diagram illustrating another modified example (second modified example) of an information processing system having an information processing device according to one embodiment of the present invention, which is a revised version of the block diagram illustrating the first modified example (first modified example) of an information processing system having an information processing device according to one embodiment of the present invention shown in Figure 4, including changes in terminology.

[0067] The differences between the second modified version and the first modified version are: (1) it targets information of individual entities, including individuals; (2) the part that outputs information about individual entities, including individuals, is defined as the first information processing module 700, a concept that includes the first database 400; and (3) the location where the information is accumulated is defined as a repository, a broader concept that includes databases. The repository includes the concept of a network from which information can be obtained, such as the internet, as a place where information is accumulated.

[0068] Specifically, the signal output from the input / output unit 110 was personally identifiable information in the first modified example, but in the second modified example, it is individual entity identifiable information as a concept that includes personally identifiable information. In the concept of individual entities, "individual" refers to a "natural person," and individual entities include the concept of "legal entities." Individual entities also include living organisms other than humans (Biological Entities), with specific examples including animals, plants, fungi, and viruses. Individual entities include non-living entities as long as they are identifiable from others, with specific examples including built structures such as buildings and houses, mobile equipment such as vehicles, ships, and aircraft, and artificial satellites. Individual entities can be classified according to their purpose of existence, with specific examples including entities that participate in competitions and entities that learn. Individual entities can also be classified according to their value, with specific examples including mining resources, copyrighted works, intellectual property, and trading cards.

[0069] In the second modification, as described above, the target is information of individual entities, including individuals. Therefore, the block that was the personal-related information database 300 in the first modification becomes the entity-related information repository 300A in the second modification. The entity-related information repository 300A receives entity identification information from the search extension generation module 120 and outputs entity-related information.

[0070] Furthermore, the block that was the second database 500 in the first modified example becomes the second information repository 500A in the second modified example, and it is clarified that it can take the form of an information aggregate containing multiple databases, such as the Internet. The second information repository 500A receives a second information processing command including the second search condition from the search extension generation module 120 and outputs a second acquisition result including the second search result.

[0071] Furthermore, in the second modification, the block enclosed by the dashed line, which includes the information processing unit 600 and the first database 400 in the first modification, is configured as the first information processing module 700. Specifically, the first information processing module 700 receives a first information processing command from the search extension generation module 120, which may include search conditions as an example, and outputs a first acquisition result, which may include the first search result as an example.

[0072] The first information processing module 700 comprises a generation processing unit 600A and a search processing unit 450A. The generation processing unit 600A takes information including entity identification information as input and generates entity metrics information, which is an indicator of the performance of individual entities, as first information. The first information processing instruction includes a control signal for generating the first information when it causes the generation processing unit 600A to execute.

[0073] The generation processing unit 600A may generate first information by performing information processing using a machine learning model. The health risk prediction engine 610 and the health risk fluctuation prediction engine 630 shown in Figure 5 are engines that perform information processing using a machine learning model. The entity metrics information included in the first information includes at least one of the following: information regarding the evaluation result of the current performance of individual entities and information regarding the prediction of the future performance of individual entities. If the first information includes information regarding predictions, it may be preferable for the generation processing unit 600A to perform information processing using a machine learning model.

[0074] The search processing unit 450A searches the stored information in the first information repository 400A using a first search condition generated from information including entity identification information as input, and outputs search results that include the first information. The first information processing command includes a control signal to execute a search using the first search condition when the search processing unit 450A is to be executed.

[0075] The first information repository 400A stores the first information generated by the generation processing unit 600A. The first information processing command may cause the generation processing unit 600A to generate the first information, store the generated first information in the first information repository 400A, and then perform a search on the first information repository 400A containing the generated first information using the first search conditions. In this case, the search results may include first information previously generated by the generation processing unit 600A, which may provide information about the changes in the state of individual entities over time.

[0076] In the first modified example, the block that was the second database 500 becomes the second information repository 500A in the second modified example, and it is clarified that it includes information aggregation sites such as the internet as part of its concept.

[0077] Figure 6 is a block diagram illustrating an information processing system having an information processing device related to a third modified example, which is a further generalization of the second modified example shown in Figure 5.

[0078] An embodiment of the information processing device 100 included in the information processing system shown in Figure 6 is described below.

[0079] (1) The information processing device 100 according to this embodiment includes a control unit 100C having a search extension generation module 120 that generates inference requests to be input to a machine-learned large-scale language model 200 based on input information including entity identification information that can identify individual entities, and a transmitting / receiving unit 130 that transmits inference requests to the large-scale language model 200 and receives inference results from the large-scale language model 200. The search extension generation module 120 generates inference requests influenced by entity identification information based on a first acquisition result output from the first information processing module 700 that includes first information linked to individual entities identified by entity identification information, and a second acquisition result obtained by searching a second information repository 500A (Information Repository) that stores second information including information other than the first information. Specifically, it generates inference requests that do not directly include entity identification information but have a relation to entity identification information. Here, the search extension generation module 120 executes processing in the first information processing module 700 before searching the second information repository 500A, and sets the search conditions for the second information repository 500A based on the obtained first acquisition results.

[0080] This enables optimization of the search order by utilizing entity identification information to control the stages of the search. Furthermore, by inputting the second acquisition results, including the search results from the second information repository 500A, into the large-scale language model 200 to generate meaningful information, the inference accuracy is improved through the use of the large-scale language model 200.

[0081] (2) In the information processing device 100 described above, the first information processing module 700 may include at least one of the following: a search processing unit 450A that searches the stored information in the first information repository 400A using first search conditions generated with information including entity identification information as input, and outputs search results including the first information; and a generation processing unit 600A that generates entity metrics information, which is an indicator of the performance of individual entities, as the first information, with information including entity identification information as input. This enables the evaluation of individual entities and future predictions, thereby realizing the dynamic generation of performance indicators.

[0082] (3) The first information repository 400A may store the first information generated by the generation processing unit 600A. This enables data reuse. Furthermore, by utilizing past search results, more advanced inference becomes possible, thus optimizing the search.

[0083] (4) In the information processing device 100 described above, the generation processing unit 600A may perform information processing using the machine learning model 610A. By performing information generation using the machine learning model 610A, dynamic data generation is realized based on the integration of search and data generation. Unlike static database search, dynamic data generation enables real-time information generation using the machine learning model 610A. Furthermore, by accumulating the information generated by the machine learning model 610A, it can be used for further learning of the machine learning model 610A, thus building a foundation for continuous learning. This is expected to improve the accuracy of the information generated by the machine learning model 610A.

[0084] (5) In the information processing device 100 described above, the entity metrics information may include at least one of the following: information regarding the evaluation results of the current performance of an individual entity and information regarding the prediction of the future performance of an individual entity.

[0085] This enables the provision of information that includes future predictions, rather than simply searching past data, thus achieving dynamic generation of performance information (generation of real-time information). Furthermore, by incorporating prediction data into the search criteria, it becomes possible to obtain more relevant information, thereby improving the accuracy of search results in the second information repository 500A.

[0086] (6) The information processing device 100 may further include an entity-related information repository 300A in which entity-related information useful for obtaining the first acquisition result is stored. In this case, the search extension generation module 120 executes a search process on the entity-related information repository 300A with search conditions including entity identification information to obtain entity-related information related to individual entities. The first information processing module 700 takes information including entity identification information and entity-related information as input and outputs the first acquisition result.

[0087] This expands the searchable information by utilizing related information rather than just entity information. Furthermore, adding related data enables more detailed analysis, allowing for multifaceted analysis of individual entities.

[0088] (7) In the information processing device 100 described above, entity-related information may consist of multiple types of information, in which case entity identification information consists of a part of the multiple types of information. This improves the contextual understanding of the search because entity-related information is considered, rather than just a simple string search. Furthermore, by dynamically configuring entity identification information, the accuracy of the search can be improved, enabling the design of multidimensional search conditions. This contributes to improving the accuracy of the first information.

[0089] (8) In the information processing device 100 described above, it is preferable that one or more selected from the group consisting of the first acquisition result, the search conditions for the second information repository, and the inference request does not contain information that can identify individual entities.

[0090] A specific example of information processing performed by the information processing system shown in Figure 6 is given below. The information processing system 1000 shown in Figure 6 can be used as the following support tool.

[0091] (1) A support tool for becoming the person you want to be (first support tool) An information processing system 1000 according to one embodiment of the present invention functions as a support tool (first support tool) that outputs a plan and advice for correcting the gap between a predicted future (including the present) profile and a target value.

[0092] First, entity identification information and entity-related information are input into the first information processing module 700, which includes a machine learning model 610A, to obtain first information regarding the prediction of the future state of individual entities (Step 1).

[0093] Next, we identify the discrepancy (gap) between the prediction and the target value based on user input (Step 2). The degree of this discrepancy (gap) can also be considered a type of metric information.

[0094] Next, the second information processing instruction is appropriately set to obtain a second acquisition result from the second information repository 500A that includes the elements necessary to bridge this gap (step 3).

[0095] Furthermore, the search extension generation module 120 is used to generate an inference request that includes the second acquisition result, and the large-scale language model 200 outputs a plan (inference result) to bridge the discrepancy (gap) (step 4).

[0096] Because entity identification information changes through practice and over time, users will repeat steps 1 through 4 multiple times. This is equivalent to going through the so-called PDCA (Plan-Do-Check-Action) cycle.

[0097] Specific application areas for the first support tool include health (prevention of physical and mental illness), education (learning, qualifications, skills), skills (exercise, music, art, performing arts), and management (management skills, leadership skills).

[0098] (2) A support tool to avoid potential value impairment (second support tool) The information processing system 1000 according to one embodiment of the present invention functions as a support tool (second support tool) that outputs plans and advice to avoid predicted future (including present) risks.

[0099] First, entity identification information and entity-related information are input to the first information processing module 700, which includes a machine learning model 610A, to obtain first information regarding the future state of individual entities (Step 1).

[0100] Next, we identify the minimum loss for each individual entity (Step 2). This loss information can also be considered a type of metric information.

[0101] Next, the second information processing instruction is appropriately set to obtain a second acquisition result from the second information repository 500A that includes the elements necessary to achieve the minimum value of this loss (step 3).

[0102] Furthermore, the search extension generation module 120 is used to generate an inference request including the second acquisition result, and the large-scale language model 200 outputs an action plan (inference result) to minimize the loss (step 4).

[0103] Because entity identification information changes through practice and over time, users will repeat steps 1 through 4 multiple times. This is equivalent to going through the so-called PDCA (Plan-Do-Check-Action) cycle.

[0104] Specific application areas for the second support tool include the creation of maintenance plans, reports, and insurance contracts for vehicles, buildings, satellites, and assets subject to depreciation and wear and tear.

[0105] (3) Brand support tool (third support tool) for raising the value of products or harvests to a target value An information processing system 1000 according to one embodiment of the present invention functions as a support tool (third support tool) that outputs a value-added strategy and implementation plan that bridges the gap between the predicted future (including present) value and the target value of the product.

[0106] First, entity identification information and entity-related information are input to the first information processing module 700, which includes a machine learning model 610A, to obtain first information regarding the future state of individual entities (Step 1).

[0107] Next, we identify the discrepancy (gap) between the prediction and the target value based on user input (Step 2). The degree of this discrepancy (gap) can also be considered a type of metric information.

[0108] Next, the second information processing instruction is appropriately set to obtain a second acquisition result from the second information repository 500A that includes the elements necessary to bridge this gap (step 3).

[0109] Furthermore, the search extension generation module 120 is used to generate an inference request that includes the second acquisition result, and the large-scale language model 200 outputs a brand strategy (inference result) to bridge the discrepancy (gap) (step 4).

[0110] As the implementation of a brand strategy changes entity identification information, users will repeat steps 1 through 4 multiple times. This is equivalent to going through the so-called PDCA (Plan-Do-Check-Action) cycle.

[0111] Examples of specific application areas for the third support tool include agriculture, fisheries, forestry, and mining (products).

[0112] (4) Support tool for increasing the value of companies, organizations, and portfolios to target value (fourth support tool) An information processing system 1000 according to one embodiment of the present invention functions as a support tool (fourth support tool) that outputs a strategy formulation and implementation plan to bridge the gap between predicted future (including present) value and target value.

[0113] First, entity identification information and entity-related information are input to the first information processing module 700, which includes a machine learning model 610A, to obtain first information regarding the future state of individual entities (Step 1).

[0114] Next, we identify the discrepancy (gap) between the prediction and the target value based on user input (Step 2). The degree of this discrepancy (gap) can also be considered a type of metric information.

[0115] Next, the second information processing instruction is appropriately set to obtain a second acquisition result from the second information repository 500A that includes the elements necessary to bridge this gap (step 3).

[0116] Furthermore, the search extension generation module 120 is used to generate an inference request that includes the second acquisition result, and the large-scale language model 200 outputs a brand strategy (inference result) to bridge the discrepancy (gap) (step 4).

[0117] As the implementation of a brand strategy changes entity identification information, users will repeat steps 1 through 4 multiple times. This is equivalent to going through the so-called PDCA (Plan-Do-Check-Action) cycle.

[0118] As a specific application area for the fourth support tool, hands-on investment planning for startup companies is given as an example.

[0119] (5) A support tool (fifth support tool) that outputs action plans and programs to people and systems to deal with predicted dangers and threats. An information processing system 1000 according to one embodiment of the present invention functions as a support tool (fifth support tool) that outputs a plan for systems to implement appropriate security measures from predicted future threat assessment information.

[0120] First, entity identification information and entity-related information are input into the first information processing module 700, which includes a machine learning model 610A, to obtain first information regarding the prediction of the threat level of individual entities in the future (Step 1).

[0121] Next, the second information processing command is appropriately configured to obtain the second acquisition result from the second information repository 500A, which includes the measures necessary to minimize the damage caused by this anticipated threat (Step 2).

[0122] Next, the search extension generation module 120 is used to generate an inference request including the second acquisition result, and a specific execution plan is output from the large-scale language model 200 (step 3).

[0123] Executing the generated plan changes the entity identification information, so the user repeats steps 1 through 3 multiple times. This is equivalent to running a so-called OODA (Observe-Orient-Decide-Act) loop.

[0124] Specific application areas for the fifth support tool include countermeasures against physical and cyber attacks, and countermeasures against natural disasters.

[0125] (6) Support tool for the movement and walking of a bio-type robot (sixth support tool) The information processing system 1000 according to one embodiment of the present invention functions as a support tool (sixth support tool) that outputs a program related to the posture and motion control of the movement and walking of a bio-type robot. In this specification, a bio-type robot includes bipedal robots represented by humanoid robots, multi-legged robots such as two-legged and six-legged robots, robots with fins for moving underwater, and robots with wings for flight (including bird wings, wing membranes of bats, etc., and gliding membranes of flying squirrels, etc.). Furthermore, the movement of a bio-type robot includes changing the relative position of the robot's components (for example, moving its arms), and locomotion such as walking, running, flying, jumping, gliding, swimming, and crawling.

[0126] First, unique information regarding the specifications of the target robot or machine is input into the first information processing module 700, which includes a machine learning model 610A, as entity identification information and entity-related information, to obtain first information including the environment for the robot's movement and walking, movement functions, walking functions, and evaluation of the state of movement and walking (Step 1).

[0127] Next, the second information processing command is appropriately set to obtain the second acquisition result, which includes operation information and operation examples of the target robot's dynamic attitude control, from the second information repository 500A (step 2).

[0128] Next, the search extension generation module 120 is used to generate an inference request including the second acquisition result, and a specific operation program is output from the large-scale language model 200 (step 3).

[0129] Executing the generated plan changes the entity identification information, so the user repeats steps 1 through 3 multiple times. This is equivalent to running a so-called OODA (Observe-Orient-Decide-Act) loop.

[0130] Examples of specific application areas for the sixth support tool include multi-legged, multi-jointed, multi-limbed, multi-toed, and humanoid robots and mechanical systems.

[0131] (7) A support tool for personnel allocation and human resource planning, which presents too many challenges for humans to perform (7th support tool) An information processing system 1000 according to one embodiment of the present invention functions as a support tool (7th support tool) that outputs an appropriate human resource plan and dynamic personnel allocation plan from performance information of candidate personnel to be allocated.

[0132] First, as entity identification information and entity-related information, unique information of the personnel who are candidates for placement is input into the first information processing module 700, which includes a machine learning model 610A, to obtain first information including a prediction and evaluation of suitable tasks for future candidate personnel and their contribution to profitability and efficiency (Step 1).

[0133] Next, the first information, which includes the above-mentioned predictive evaluation, is appropriately used to automatically generate a second information processing command, and the second acquisition result, which includes information on the capabilities, proficiency, experience, personality, and number of personnel required for the operations of the target organization, is obtained from the second information repository 500A (Step 2).

[0134] Next, the search extension generation module 120 is used to generate an inference request including the second acquisition result, and a specific personnel execution plan is output from the large-scale language model 200 (step 3).

[0135] Executing the generated plan changes the entity identification information, so the user repeats steps 1 through 3 multiple times. This is equivalent to running a so-called OODA (Observe-Orient-Decide-Act) loop.

[0136] Examples of specific application areas for the seventh support tool include companies, organizations, government agencies, and general organizations.

[0137] (8) Support tool for constructing and modifying product manufacturing processes (8th support tool) An information processing system 1000 according to one embodiment of the present invention functions as a support tool (8th support tool) that outputs product manufacturing processes and implementation plans for more advanced automation and efficiency in product manufacturing.

[0138] First, entity identification information and entity-related information, such as product / part / manufacturing equipment specifications, personnel work capacity, and costs, are input into the first information processing module 700, which includes a machine learning model 610A, to obtain first information including a predictive evaluation of product accuracy, required time and cost, and efficiency when the work is automated (Step 1).

[0139] Next, the first information, which includes the above-mentioned predictive evaluation, is appropriately used to automatically generate a second information processing command, and the second acquisition result, which includes information on examples of work implemented and devised to fully automate the target manufacturing process, is obtained from the second information repository 500A (Step 2).

[0140] Next, the search extension generation module 120 is used to generate an inference request including the second acquisition result, and a specific process efficiency improvement and automation plan is output from the large-scale language model 200 (step 3).

[0141] Executing the generated plan changes the entity identification information, so the user repeats steps 1 through 3 multiple times. This is equivalent to running a so-called OODA (Observe-Orient-Decide-Act) loop.

[0142] As a specific application area for the eighth support tool, all types of manufacturing industries are given as examples.

[0143] (9) Support tool for outputting business process efficiency plans and programs (9th support tool) An information processing system 1000 according to one embodiment of the present invention functions as a support tool (9th support tool) that outputs implementation plans for automating business processes and configuring and rearranging business flow routes for the purpose of improving the efficiency of an organization's operations.

[0144] First, entity identification information and entity-related information, such as the specifications of the business process and the characteristics of the equipment and personnel required for its implementation, are input into the first information processing module 700, which includes a machine learning model 610A, to obtain first information including the specifications and characteristics of the business process, and predictive evaluations of efficiency through automation of the business flow and changes to the business route (Step 1).

[0145] Next, the first information, which includes the above-mentioned predictive evaluation, is appropriately used to automatically generate a second information processing command, and the second acquisition result, which includes information on the introduction of automation of business processes, the reorganization of routes for efficiency improvements, and invention examples, is obtained from the second information repository 500A (Step 2).

[0146] Next, the search extension generation module 120 is used to generate an inference request including the second acquisition result, and the large-scale language model 200 outputs a plan for automating specific processes and rearranging paths (step 3).

[0147] Executing the generated plan changes the entity identification information, so the user repeats steps 1 through 3 multiple times. This is equivalent to running a so-called OODA (Observe-Orient-Decide-Act) loop.

[0148] Examples of specific areas of application for the ninth support tool include government agencies, public services, service industry organizations including medical institutions, and non-profit organizations.

[0149] The relative positions of the above support tools are shown in Figure 7.

[0150] The following are specific application examples of the information processing system 1000 according to one embodiment of the present invention. (Example 1) Example 1 is also shown in Figure 5. In Example 1, the individual entities are natural persons.

[0151] The entity-related information stored in the entity-related information repository 300A includes at least one of the following: individual health checkup information 310 and social insurance medical fee statement information (receipt information 320) relating to natural persons.

[0152] The first information included in the first acquisition result output by the first information processing module 700 includes one or more selected from the group consisting of health risk data 410 (see Figure 1), stress visualization data 420 (see Figure 1), and health risk fluctuation data 430 (see Figure 1) of a natural person.

[0153] The second information stored in the second information repository 500A includes one or more selected from the group consisting of medical papers 510, health management-related information 520, nutrition and diet-related information 530, exercise function-related information 540, and physician's findings-related information 550, all of which are of natural persons.

[0154] The inference results received by the transmitting / receiving unit 130 include information about a natural person, selected from the group consisting of a health prediction report, a disease prediction report, a health advice report, a proposal regarding medical expenses, and a rehabilitation plan.

[0155] Even if the individual entity identification information output from the input / output unit 110 is only a name, if health examination information for that natural person can be obtained from the entity-related information repository 300A, the first information processing command can include medical information necessary for the health risk prediction engine 610 to output an appropriate response. In this way, if only the entity identification information input by the user is available, the variety of information may be limited, making it difficult to properly perform generation processing of the first information processing module, such as prediction processing. However, by adding entity-related information consisting of multiple types of information, it becomes possible to prepare an appropriate dataset, thereby improving the reliability of generation processing in the first information processing module 700.

[0156] If the first information generated by the health risk prediction engine 610 includes data on health risks, such as a particularly high probability of developing diabetes, then lifestyle habits to reduce the probability of developing diabetes can be included as search conditions for the second information repository 500A. A second information processing command including such search conditions will result in the second acquisition result output from the second information repository 500A including, for example, literature on specific lifestyle habits to reduce the probability of developing diabetes.

[0157] As a result, even though the input / output unit 110 contains only the names of natural persons, information that is particularly useful in creating a health advice report for that natural person can be added to the inference requests (prompts) that the search extension generation module 120 outputs to the large-scale language model 200, including descriptions of literature on specific lifestyle habits to reduce the likelihood of developing diabetes.

[0158] Moreover, in a preferred example, the inference requests (prompts) generated by the search extension generation module 120 do not contain any specific personal information. This prevents the leakage of personal information, regardless of the type of information attribution policy used for the large-scale language model 200 (for example, an information attribution policy that attributes the ownership of all data included in all inference requests and inference results to the model provider or developer).

[0159] (Example 2) In Example 2, the individual entities are Mobility Equipment. Mobility Equipment includes vehicles such as automobiles and motorcycles, ships (including land-based hovercraft), and airplanes, as well as underwater mobility equipment such as submarines.

[0160] The entity-related information stored in the entity-related information repository 300A includes one or more selected from the group consisting of the mobile device's serial number, year of manufacture, repair information, inspection information, accident history, and structural drawings, as well as the mobile device's owner history information.

[0161] The first information included in the first acquisition result output by the first information processing module 700 includes one or more selected from the group consisting of a future risk assessment and a physical service life assessment of the mobile equipment.

[0162] The second information stored in the second information repository 500A includes general information about mobile devices. Specific examples of general information include information on the capabilities (specifications) of mobile devices, the environment in which they are used, and their usage patterns.

[0163] The inference results received by the transmitting / receiving unit 130 include information for creating a contract for damage insurance for mobile equipment. The inference results may also include maintenance plans to extend the service life.

[0164] (Example 3) In Example 3, the individual entities are built structures. Built structures are not limited in type as long as they are identifiable from others, and specific examples include buildings such as detached houses, apartment buildings and buildings where businesses reside, facilities such as sports facilities, amusement facilities and entertainment venues, and civil engineering structures such as bridges, tunnels and dams.

[0165] The entity-related information stored in the entity-related information repository 300A includes one or more selected from the group consisting of a number that identifies the installation location of the artificial structure, and the purchase history and repair history of the artificial structure.

[0166] The first information included in the first acquisition result output by the first information processing module 700 includes at least one selected from the group consisting of the necessity of repairing the artificial structure and the amount required for the repair.

[0167] The second information stored in the second information repository 500A includes general information regarding the estimation of the real estate value of artificial structures.

[0168] The inference results received by the transmitting / receiving unit 130 include information for creating at least one selected from the group consisting of insurance contracts for artificial structures related to earthquakes, disasters, fires, etc., i.e., property insurance, engineering reports, real estate investment information reports, and maintenance plans to extend the service life of artificial structures.

[0169] (Example 4) In Example 4, the individual entities are Learning Agents. The Learning Agents may be natural people, non-human animals such as dogs, horses, or parakeets, or even computers performing machine learning.

[0170] The entity-related information stored in the entity-related information repository 300A includes one or more selected from the group consisting of the learner's identification number, periodic and non-periodic exam results, submitted reports, and learning history.

[0171] The first information included in the first acquisition result output by the first information processing module 700 includes at least one selected from the group consisting of the learning subject's learning achievement evaluation results, personality analysis results, and mental fatigue level check results.

[0172] The second piece of information includes general information about at least one selected from a group consisting of learning subject information, learning subject information, and instructional literature.

[0173] The inference results received by the transmitting / receiving unit 130 include information for creating at least one selected from the group consisting of an individual learning plan, individualized advice, and learning progress management information.

[0174] (Example 5) In Example 5, the individual entities are Useful Biological Entities. Useful organisms include not only organisms classified by the animal kingdom, plant kingdom, fungal kingdom, protist kingdom, prokaryote kingdom, etc., but also organisms that do not fall under the category of living organisms, such as viruses, as long as they have usefulness.

[0175] The first information included in the first acquisition result output by the first information processing module 700 includes at least one selected from the group consisting of brand value information, product characteristic information, rarity information, and growth inhibitory factor-related information of a useful organism.

[0176] The second information stored in the second information repository 500A includes general information about at least one selected from the group consisting of the useful organism's method of use and its safety.

[0177] The inference results received by the transmitting / receiving unit 130 include information for creating at least one selected from the group consisting of explanatory materials for loans for businesses targeting useful organisms, materials that contribute to the sales promotion of useful organisms, and materials for creating brand strategies related to useful organisms.

[0178] If an individual entity is an agricultural organism, the entity-related information stored in the entity-related information repository 300A may include one or more selected from the group consisting of farmland number, seedling registration number, organic certification number, pesticide registration number, and origin certification number, relating to the agricultural organism.

[0179] If an individual entity is an aquatic resource, including farmed and wild aquatic resources, that inhabits aquatic bodies and is used in the fisheries industry, the entity-related information shall include one or more of the following items related to the aquatic resource: catch number, notification number for specific Class I aquatic animals and plants, lot number, fisherman registration number, fishing permit number, fishing vessel registration number, catch tracking number, fishery resource management number, fishing right number, fishery product quality certification number, and origin certificate number for catches, export fisheries business registration number, general importer certification number, and manufacturer importer certification number.

[0180] If an individual entity is livestock, the entity-related information stored in the entity-related information repository 300A may include one or more of the following selected from the group consisting of the livestock's individual identification number, ear tag number, microchip number, livestock registration number, DNA identification code, farm identification number, movement and trade identification number, and international identification number.

[0181] (Example 6) In Example 6, the individual entity is Mineral Resources. In this specification, Mineral Resources means non-living resources obtained by mining (including open-pit mining), and includes substances related to fossil fuels such as petroleum, coal, natural gas, and asphalt.

[0182] The entity-related information stored in Entity-Related Information Repository 300A includes one or more selected from the group consisting of mining registration numbers, mineral mining permit numbers, mineral import / export permit numbers, mineral product quality certification numbers, mineral processing / processing facility registration numbers, mining company numbers, mine worker registration numbers, mineral trading registration numbers, mining environmental protection numbers, mining permit numbers, mining machinery registration numbers, mineral origin certification numbers, mineral product safety standard compliance numbers, mining pollution prevention numbers, and mineral mining rights numbers, relating to mining resources.

[0183] The first information included in the first acquisition result output by the first information processing module 700 includes at least one piece of information selected from the group consisting of the utility value, purity characteristics, rarity, means of transport, and transport routes of the mining resources, as well as the geopolitical location of the mining site of the mining resources.

[0184] The second information stored in the second information repository 500A includes general information regarding the safety, commodity characteristics, usage methods, and processing and transportation costs of mining resources, as well as the political situation of the mining sites.

[0185] The inference results received by the transmitting / receiving unit 130 include information for creating at least one selected from the group consisting of a mining resource, a security strategy, and a commercialization brand strategy. The commercialization brand strategy includes a brand strategy relating to the commercialization of mining resources that have been processed from mining to refining.

[0186] (Example 7) In Example 7, the individual entity is an Industrial Product. In this specification, Industrial Products are not limited to tangible assets such as televisions and refrigerators, but also include intangible assets such as game software and business software.

[0187] The entity-related information stored in the entity-related information repository 300A includes one or more selected from the group consisting of manufacturer registration number, product identification number, quality certification number, manufacturing number, machine equipment number, manufacturing license number, manufacturer ID number, quality certification number, import / export control number, trademark registration number, manufacturing facility license number, patent number, product registration number, and environmental standard compliance number, relating to industrial products.

[0188] The first information included in the first acquisition result output by the first information processing module 700 includes at least one piece of information selected from the group consisting of the manufacturer, price, year of manufacture, accident history, and manual of the industrial product.

[0189] The second information stored in the second information repository 500A includes at least one selected from the group consisting of general information about competing products in the market for industrial products, claim information for industrial products and competing products, and general information about the destinations for industrial products and competing products.

[0190] The inference results received by the transmitting / receiving unit 130 include information for creating at least one selected from the group consisting of information on the existence of repair parts (in the case of software, update patches) for industrial products, safety reports, and warranty documents for the export of used industrial products.

[0191] (Example 8) In Example 8, the individual entity is a legal entity.

[0192] The entity-related information stored in Entity-Related Information Repository 300A includes one or more of the following for a legal entity: corporate number, commercial registration number, securities code / ticker symbol, EDINET code (Electronic Disclosure for Investors' NETwork Code), LEI code (Legal Entity Identifier), qualified invoice issuer number (T number), tax office reference number, and customs duty number (Harmonized System Code).

[0193] The first information included in the first acquisition result output by the first information processing module 700 includes at least one piece of information selected from the group consisting of a quantitative evaluation of an individual stock price and a financial evaluation of a corporation.

[0194] The second information stored in the second information repository 500A includes fundamental information of corporations.

[0195] The inference results received by the transmitting / receiving unit 130 include information for creating a comprehensive stock price valuation of a corporation.

[0196] If the legal entity is a service provider, the entity-related information stored in the entity-related information repository 300A may include one or more of the following items related to the service provider: business license number, customer management number, license number, service contract number, operating license number, franchise number, employee number, consumer complaint number, invoice number, customer service number, product / service identification number (SKU number), and delivery management number.

[0197] (Example 9) In Example 9, the individual entity is an IT company (Service Company), which is a type of legal entity.

[0198] The entity-related information stored in the entity-related information repository 300A includes one or more selected from the group consisting of corporate number, software registration number, domain name registration number, software product ID number, IP address number, engineer qualification number, patent number, database management number, hardware serial number, Internet service provider (ISP) number, IT contract number, server management number, and cloud service account number, relating to an IT company.

[0199] The first information included in the first acquisition result output by the first information processing module 700 includes information about the services provided by individual IT companies.

[0200] The second information stored in the second information repository 500A includes one or more selected from the group consisting of security information, lead time, and unit price information of IT companies.

[0201] The inference results received by the transmitting / receiving unit 130 include information for creating business order support information for an IT company, which includes one or more selected from the group consisting of supplier, unit price, lead time, and accuracy.

[0202] (Example 10) In Example 10, the individual entities are copyrighted works. Specific examples of copyrighted works include books (literary works), musical works, video works, and software.

[0203] The entity-related information stored in Entity-Related Information Repository 300A may include one or more of the following selected from the group consisting of International Standard Book Number (ISBN), International Standard Serial Number (ISSN), JAN code, (National Diet Library Bibliographic ID (NDL Bibliographic ID)), and library management number, relating to copyrighted works, particularly books.

[0204] The first information included in the first acquisition result output by the first information processing module 700 includes at least one selected from the group consisting of the content of the copyrighted work, book information, and content information.

[0205] The second information stored in the second information repository 500A includes one or more selected from the group consisting of general information about the author of a copyrighted work, general information about the literature, general information about the content's holdings, and information about past adaptations of copyrighted works into film.

[0206] The inference results received by the transmitting / receiving unit 130 include information for creating one or more of the following selected from the group consisting of availability information, evaluation / introduction texts, scripts for video works (specific examples include movies, TV dramas, anime, and videos), and manga storyboards for copyrighted works.

[0207] (Example 11) In Example 11, the individual entity is a Performer. In this specification, a Performer is an individual entity (not limited to natural persons) capable of demonstrating a particular ability, and specific examples include entertainers, musicians, and athletes.

[0208] The entity-related information stored in Entity-Related Information Repository 300A includes one or more of the following selected from the group of performers: ID issued by the entertainer's management agency, performer number, copyright registration number, International Standard Score Number (ISRC), International Standard Recording Code (ISMN), movie / music album identification number, fan card number, athlete number, Olympic athlete number, sports team number, competition registration number, leaderboard number, athlete ID number, and sports license number.

[0209] The first information included in the first acquisition result output by the first information processing module 700 includes at least one selected from the group consisting of the performer's contribution to sabermetrics, contribution to distribution, contribution to audience attendance, contribution to music distribution, and contribution to product sales.

[0210] The second information stored in the second information repository 500A includes one or more selected items from a group consisting of general information about the performer's merchandising related to sports and entertainment, and general information about scandals.

[0211] The inference results received by the transmitting / receiving unit 130 include information that is useful in creating a contract for the performer.

[0212] (Example 12) In Example 12, the individual entity is a Gambling Competitive Entity. In this specification, a Gambling Competitive Entity is an individual entity that engages in gambling competitions, and specific examples include natural persons (such as professional cyclists), a collection of individual entities (such as a soccer team), and animals other than humans (such as racehorses).

[0213] The entity-related information stored in the entity-related information repository 300A may include one or more selected from the group consisting of Racehorse Name Registration, Studbook Number, Horse ID Number, JRA Registration Number, Breeder & Owner Information, and Race Record, relating to the gambling entity.

[0214] The first information included in the first acquisition result output by the first information processing module 700 includes at least one selected from the group consisting of bloodline, bloodline performance, performance, and track compatibility, which are primarily related to gambling competitions.

[0215] The second information stored in the second information repository 500A may include one or more selected from a group consisting of racehorse characteristic information, rival horse information, racecourse information, past races, and weather / track information, primarily related to gambling competitions.

[0216] The inference results received by the transmitting / receiving unit 130 include information that contributes to at least one of the group consisting of race-by-race win / loss predictions and stallion value data for the gambling entity.

[0217] (Example 13) In Example 13, the individual entity is an artificial satellite.

[0218] The entity-related information stored in Entity-Related Information Repository 300A includes one or more of the following selected from the group of satellite-related information: International Designator (NSSDC ID), NORAD ID (Satellite Catalog Number / SATCAT), COSPAR ID (Committee on Space Research ID), GSAT number (Geostationary Satellite Number), and ITU registration number (International Telecommunication Union Registration).

[0219] The first information included in the first acquisition result output by the first information processing module 700 includes one or more selected from the group consisting of the satellite's history, movement, orbit, current position, future position, and collision probability with an object in Earth orbit.

[0220] The second information stored in the second information repository 500A includes one or more selected from a group consisting of satellite characteristics, constellation information, and registration information.

[0221] The inference results received by the transmitting / receiving unit 130 include information about the satellite, selected from the group consisting of a strategy for debris removal, information on space debris in orbit, and a plan for ensuring a stable orbit.

[0222] (Example 14) In Example 14, the individual entities are crises or threats caused by external attacks.

[0223] The entity-related information stored in Entity-Related Information Repository 300A includes one or more selected items from a group consisting of CVE numbers (Common Vulnerabilities and Exposures), MITRE ATT&CK identifiers, malware-specific identifiers (MD5 / SHA hashes), AV vendor-specific identification numbers, international malware databases (such as MalwareBazaar), and weapon lists. Specific examples of weapon lists include GlobalSecurity.org, Jane's Information Group, Federation of American Scientists (FAS) Weapons Database, SIPRI Arms Transfers Database, and Wikipedia.

[0224] The first information included in the first acquisition result output by the first information processing module 700 includes one or more selected from the group consisting of the type of crisis or threat caused by an external attack, an assessment of the threat level, and the encounter history.

[0225] The second information stored in the second information repository 500A includes one or more selected from the group consisting of general type numbers, characteristic information, and past occurrence history information of crises or threats caused by external attacks.

[0226] The inference results received by the transmitting / receiving unit 130 include information on at least one selected from the group consisting of strategies as security measures against crises and threats caused by external attacks, tactical plans for defense, and threat warning information for the target of defense.

[0227] (Example 15) In Example 15, the individual entities are the allocation of personnel within an organization.

[0228] The entity-related information stored in the entity-related information repository 300A includes one or more of the following characteristics of a candidate for deployment: age, gender, physical strength, mental strength, experience, and past accident history.

[0229] The first information included in the first acquisition result output by the first information processing module 700 includes one or more selected from a group consisting of the characteristics, history, competency assessment, personality assessment, and relationships with other personnel of the personnel who are candidates for placement in the organization, the suitable tasks for the personnel who are candidates for placement, and a predicted assessment of the contribution of the personnel who are candidates to the profitability and efficiency of the target tasks.

[0230] The second information stored in the second information repository 500A includes one or more selected items from a group consisting of the abilities, proficiency, experience, personality, and number of personnel required for the operations of the target organization in the allocation of personnel within the organization.

[0231] The inference results received by the transmitting / receiving unit 130 include information about candidate personnel, selected from a group consisting of deployment plans and dynamically updated replacement plans.

[0232] (Example 16) In Example 16, the individual entities are the actions of a bio-robot.

[0233] The entity-related information stored in the entity-related information repository 300A includes one or more selected from the group consisting of the size, weight, material and properties of the constituent parts, driving force, range of motion of joints, power, center of gravity, and motion characteristics of the bio-type robot.

[0234] The first information included in the first acquisition result output by the first information processing module 700 includes one or more selected from the group consisting of evaluations of the environment, function, and state of the biological robot's movements.

[0235] The second information stored in the second information repository 500A includes one or more selected from the group consisting of dynamic attitude control operation information and operation examples in the operation of a bio-type robot.

[0236] The inference results received by the transmitting / receiving unit 130 include information regarding the posture of the bio-robot's movements and the program related to the operation of the joints and body parts of the bio-robot in conjunction with the movement control.

[0237] (Example 17) In Example 17, the individual entities are the manufacturing processes of a product.

[0238] The entity-related information stored in the entity-related information repository 300A includes one or more selected from the group consisting of product usage and characteristics, specifications and characteristics of parts necessary for product manufacturing, specifications and characteristics of equipment used in multiple manufacturing operations that constitute the product manufacturing process, specifications and characteristics of jigs and fixtures used in manufacturing operations, the work capabilities of people involved in product manufacturing, the skills and characteristics required of people involved in product manufacturing, and the time and cost required for manufacturing operations.

[0239] The first information included in the first acquisition result output by the first information processing module 700 includes one or more selected from a group consisting of multiple manufacturing operations that constitute the product manufacturing process, the accuracy of the product when the manufacturing operations are automated, the time and cost required to automate the manufacturing operations, and the efficiency of the automated manufacturing operations compared to conventional operations.

[0240] The second information stored in the second information repository 500A includes one or more selected from the group consisting of examples of work performed to fully automate the product manufacturing process and examples of work devised to fully automate the product manufacturing process.

[0241] The inference results received by the transmitting / receiving unit 130 include information selected from the group consisting of manufacturing processes and implementation plans for further automating and streamlining the manufacturing process.

[0242] (Example 18) In Example 18, the individual entity is one selected from a group consisting of a business and the work processes that are components of that business.

[0243] The entity-related information stored in the entity-related information repository 300A includes one or more selected from the group consisting of specifications and characteristics of processing tasks required for business and work processes, specifications of equipment required for the execution of business and work processes, the work capabilities and required skills and characteristics of the people involved in processing business and work processes, and the time and costs required for business and work processes.

[0244] The first information included in the first acquisition result output by the first information processing module 700 includes one or more selected from the group consisting of specifications and characteristics of business and work processes, and evaluations of the processing path configuration of the business flow, evaluations of the feasibility of introducing automation, and evaluations of efficiency after conversion to automation.

[0245] The second information stored in the second information repository 500A includes one or more selected from a group consisting of examples of implemented and devised work related to the introduction of automation into processing paths and the reorganization of processing paths in business and work processes.

[0246] The inference results received by the transmitting / receiving unit 130 include information selected from the group consisting of an implementation plan for automation introduction and an implementation plan for reconfiguring the processing paths of business flows, with the aim of improving the operational efficiency of the organization.

[0247] In the above embodiment, the information processing device 100 included an input / output unit 110, a control unit 100C, and a transmitting / receiving unit 130, but is not limited thereto. Another embodiment of the information processing system 1000 of the present invention includes a first device having input / output functions, i.e., at least a transmitting / receiving unit 130, a second device having a data storage function, and a third device having a control unit 100C, wherein the first device can communicate with the second and third devices.

[0248] The second device includes a model storage device in which a large-scale language model 200 that has undergone machine learning is stored, a first database 400 in which first information linked to individuals identified by personally identifiable information input into the first device is stored, and a second database 500 in which second information consisting of non-personal information is stored.

[0249] The control unit 100C of the third device has a search extension generation module 120 that generates inference requests to be input to a machine-learned large-scale language model 200 based on information input from the first device. The first device includes a transmitting / receiving unit 130 that transmits the inference requests input from the third device to the second device which has the large-scale language model 200, and receives the inference results from the second device.

[0250] The search extension generation module 120 generates an inference request influenced by the first information included in the first search result, based on a first search result containing first information obtained by searching the first database 400 provided by the second device, and a second search result containing second information obtained by searching the second database 500 provided by the second device. Since the search extension generation module 120 is provided by the third device, the search performed by the search extension generation module 120 is executed via the first device.

[0251] In the information processing system 1000 according to another embodiment of the present invention described above, at least two of the first device, the second device, and the third device may be integrated.

[0252] An information processing system 1000 according to another embodiment of the present invention has a plurality of computer devices that constitute a peer-to-peer network. At least one of the plurality of computer devices in the information processing system 1000 has a search extension generation module 120 that generates inference requests to be input to a machine-learned large-scale language model 200 based on input information. Furthermore, at least one of the plurality of computer devices in the information processing system 1000 has a transmitting and receiving unit 130 that transmits inference requests to the large-scale language model 200 and receives inference results from the large-scale language model 200.

[0253] At least one of the multiple computer devices in the information processing system 1000 has an input information generation unit that generates information to be input to the search extension generation module 120. At least one of the multiple computer devices in the information processing system 1000 has a display signal generation unit 140 that generates a display signal that includes at least a portion of the inference results. The search extension generation module 120 generates an inference request influenced by the first information included in the first search result, based on a first search result that includes first information obtained by searching a first database 400 in which first information linked to an individual identified by personally identifiable information is stored, and a second search result that includes second information obtained by searching a second database 500 in which second information consisting of non-personal information is stored.

[0254] From the perspective of enhancing information security, the control unit 100C may have an inference request verification unit to confirm that the inference requests generated by the search extension generation module 120 do not contain personally identifiable information. In this case, only inference requests that the inference request verification unit has confirmed do not contain personally identifiable information will be transmitted from the transmission / reception unit 130 to the large-scale language model 200. The transmission / reception unit 130 may also have the function of an inference request verification unit.

[0255] In relation to Figure 6, the present invention includes the following embodiments.

[0256] An information processing program according to one aspect of the present invention is an information processing program that generates an inference request to be input to a machine-learned large-scale language model 200 based on input information including entity identification information that can identify individual entities, sends the inference request to the large-scale language model 200, and receives an inference result from the large-scale language model 200.

[0257] This information processing program includes a search extension generation process that generates an inference request influenced by entity identification information, based on a first acquisition result obtained by executing a first information processing and including first information linked to individual entities identified by entity identification information, and a second acquisition result obtained by searching a second information repository 500A that stores second information including information other than the first information.

[0258] In the search extension generation process, the first information processing is performed before searching the second information repository 500A, and based on the obtained first acquisition results, the search conditions for the second information repository 500A are set.

[0259] In the above-described information processing program, the first information processing may include at least one of the following: a search process that searches the stored information in the first information repository 400A using first search conditions generated with information including entity identification information as input, and outputs search results that include the first information; and a generation process that generates entity metrics information, which is an indicator of the performance of individual entities, as the first information, using information including entity identification information as input.

[0260] In the above information processing program, the search process of the entity-related information repository 300A, which stores entity-related information that contributes to obtaining the first acquisition result, may be performed using search conditions that include entity identification information to obtain entity-related information related to individual entities. In this case, the first information processing takes information including entity identification information and entity-related information as input and outputs the first acquisition result.

[0261] An information processing system 1000 according to one aspect of the present invention comprises a first device having input / output functions, a second device having data storage functions, and a third device equipped with a control unit 100C, wherein the first device is capable of communicating with the second and third devices.

[0262] The second device has a model storage device in which a machine-learned large-scale language model 200 is stored. The control unit 100C of the third device has a search extension generation module 120 that generates inference requests to be input to the large-scale language model 200 based on input information that is input from the first device and includes entity identification information that can identify individual entities. The first device includes a transmitting and receiving unit 130 that transmits the inference requests input from the third device to the second device which has the large-scale language model 200 and receives the inference results from the second device.

[0263] The search extension generation module 120 generates an inference request influenced by entity identification information based on a first acquisition result obtained from the first information processing module 700, which includes first information linked to individual entities identified by entity identification information, and a second acquisition result obtained by searching the second information repository 500A, which is owned by the second device and stores second information including information other than the first information.

[0264] The search extension generation module 120 performs processing in the first information processing module 700 before searching the second information repository 500A, and sets the search conditions for the second information repository 500A based on the first acquisition results obtained.

[0265] In the information processing system 1000 described above, at least two of the first, second, and third devices may be integrated into one unit. Furthermore, each device may consist of multiple devices, in which case the multiple devices may be located in different locations.

[0266] In the above-described information processing system 1000, the first information processing module 700 may include at least one of the following: a search processing unit 450A that searches the stored information in the first information repository 400A of the second device using first search conditions generated with information including entity identification information as input, and outputs search results including the first information; and a generation processing unit 600A that generates entity metrics information, which is an indicator of the performance of individual entities, as the first information, using information including entity identification information as input.

[0267] In the above-described information processing system 1000, the second device may further have an entity-related information repository 300A in which entity-related information useful for obtaining the first acquisition result is stored. In this case, the search extension generation module 120 executes a search process of the entity-related information repository 300A with search conditions including entity identification information to obtain entity-related information related to individual entities. The first information processing module 700 takes information including entity identification information and entity-related information as input and outputs the first acquisition result.

[0268] In the information processing system 1000 described above, at least one of the first device, the second device, and the third device may be composed of multiple devices.

[0269] Another aspect of the present invention relates to an information processing system 1000 which has a plurality of computer devices that constitute a peer-to-peer network.

[0270] At least one of the multiple computer devices has a search extension generation module 120 that generates inference requests to be input to a machine-learned large-scale language model 200 based on the input information.

[0271] At least one of the multiple computer devices has a transmitting / receiving unit 130 that sends inference requests to the large-scale language model 200 and receives inference results from the large-scale language model 200.

[0272] At least one of the multiple computer devices has an input information generation unit that generates information to be input to the search extension generation module 120. The input information generation unit may be included in a computer device that has an input / output unit 110.

[0273] At least one of the multiple computer devices has a display signal generation unit 140 that generates a display signal including at least a portion of the inference results. The display signal generation unit 140 may be included in a computer device having an input / output unit 110.

[0274] The search extension generation module 120 generates an inference request influenced by entity identification information based on a first acquisition result obtained from the first information processing module 700, which includes first information linked to individual entities identified by entity identification information, and a second acquisition result obtained by searching the second information repository 500A, which stores second information including information other than the first information. Before searching the second information repository 500A, the search extension generation module 120 executes processing in the first information processing module 700 and sets the search conditions for the second information repository 500A based on the obtained first acquisition result.

[0275] In the above-described information processing system 1000, the first information processing module 700 includes at least one of the following: a search processing unit 450A that searches the stored information in the first information repository 400A using first search conditions generated with information including entity identification information as input, and outputs search results including the first information; and a generation processing unit 600A that generates entity metrics information, which is an indicator showing the performance of individual entities, as the first information, with information including entity identification information as input.

[0276] In the above-described information processing system 1000, at least one of the multiple computer devices has an entity-related information repository 300A in which entity-related information useful for obtaining the first acquisition result is stored.

[0277] The search extension generation module 120 executes a search process in the entity-related information repository 300A using search conditions that include entity identification information, and obtains entity-related information related to individual entities.

[0278] The first information processing module 700 takes information including entity identification information and entity-related information as input and outputs a first acquisition result.

[0279] An information processing method according to one aspect of the present invention generates an inference request to be input to a machine-learned large-scale language model 200 based on input information including entity identification information that can identify individual entities, transmits the inference request to the large-scale language model 200, and receives an inference result from the large-scale language model 200.

[0280] In the above information processing method, a search extension generation process is executed to generate an inference request influenced by entity identification information, based on a first acquisition result obtained by executing the first information processing and including first information linked to individual entities identified by entity identification information, and a second acquisition result obtained by searching a second information repository 500A that stores second information including information other than the first information.

[0281] In the search extension generation process, the first information processing is performed before searching the second information repository 500A, and based on the obtained first acquisition results, a process is executed to set the search conditions for the second information repository 500A.

[0282] In the above-described information processing method, the first information processing involves performing at least one of the following: a search process that searches the stored information in the first information repository 400A using first search conditions generated with information including entity identification information as input, and outputs search results that include the first information; and a generation process that generates entity metrics information, which is an indicator of the performance of individual entities, as the first information, using information including entity identification information as input.

[0283] In the above information processing method, the search process of the entity-related information repository 300A, which stores entity-related information that contributes to obtaining the first acquisition result, is performed using search conditions that include entity identification information to obtain entity-related information related to individual entities.

[0284] In the first information processing step, information including entity identification information and entity-related information is taken as input, and the first acquisition result is output.

[0285] An information processing device according to another aspect of the present invention includes a control unit 100C having a search extension generation module 120 that generates an inference request to be input to a machine-learned large-scale language model 200 based on input information including specific information that can identify confidential information which is information that should be kept confidential, and a transmitting / receiving unit 130 that transmits the inference request to the large-scale language model 200 and receives the inference result from the large-scale language model 200. The search extension generation module 120 generates an inference request influenced by confidential information based on a first acquisition result output from a first information processing module 700 which includes first information linked to confidential information identified by the specific information, and a second acquisition result obtained by searching a second information repository 500A which stores second information including information other than the first information. The search extension generation module 120 executes processing in the first information processing module 700 before searching the second information repository 500A, and sets search conditions for the second information repository 500A based on the obtained first acquisition result.

[0286] A concrete example of confidential information is specific information that can identify individual entities, while first information is more abstract than confidential information and therefore requires relatively less confidentiality management. When confidential information is information that can identify individual entities, first information is information that describes the characteristics of the individual entities but does not identify them. By creating search conditions for the second information repository 500A using this first information, information highly related to confidential information can be obtained from the second information repository 500A without revealing the individual entities that constitute confidential information to the second information repository 500A. This improves the accuracy and / or comprehensiveness of inferences regarding confidential information in the large-scale language model 200.

[0287] In other words, the first information processing module 700 is an information transformation module that abstracts specific information in confidential information, thereby enhancing its confidentiality while maintaining its essential characteristics. To put it another way, the first information processing module 700 is an information anonymization transformation device that enhances both the anonymity and clarity of information, particularly the clarity of the search conditions in the second information repository 500A. This makes it possible to obtain more appropriate inference results while reducing the possibility that the inference requests generated by the search extension generation module 120 may contain confidential information about individual entities.

[0288] When performing this information conversion, an entity-related information repository 300A may be further provided, which stores relevant information that contributes to obtaining the first acquisition result, in order to increase the amount of confidential information input. In this case, the search extension generation module 120 executes a search process of the entity-related information repository 300A with search conditions that include specific information to obtain relevant information that is related to the confidential information and contributes to clarifying the search conditions of the second information repository 500A, and the first information processing module 700 takes the confidential information and the information including this relevant information as input and outputs the first acquisition result.

[0289] 1000: Information processing system 100: Information processing device 100C: Control unit 110: Input / output unit 120: Search extension generation module 130: Transmit / receive unit 140: Display signal generation unit 150: Training data creation unit 200: Large-scale language model 300: Personal related information database 300A: Entity related information repository 310: Individual health checkup information 320: Medical claim information 400: First database 400A: First information repository 410: Health risk data 420: Stress visualization data 430: Health risk fluctuation data 450A: Search processing unit 500: Second database 500A: Second information repository 510: Medical papers 520: Health management related information 530: Nutrition and diet related information 540: Exercise function related information 550: Physician's findings related information 600: Information processing unit 600A: Generation processing unit 610: Health risk prediction engine 610A: Machine learning model 620: Stress visualization engine 630: Health risk fluctuation prediction engine 700: First information processing module

Claims

1. An information processing device comprising: a control unit having a search extension generation module that generates an inference request to be input to a machine-learned large-scale language model based on input information including entity identification information that can identify individual entities; and a transmitting / receiving unit that transmits the inference request to the large-scale language model and receives an inference result from the large-scale language model, wherein the search extension generation module generates an inference request using information that does not directly include the entity identification information but is related to the entity identification information, based on: a first acquisition result output from a first information processing module and including first information associated with the individual entities identified by the entity identification information; and a second acquisition result obtained by searching a second information repository that stores second information including information other than the first information; and the search extension generation module executes processing in the first information processing module before searching the second information repository, and sets search conditions for the second information repository based on the obtained first acquisition result.

2. The information processing apparatus according to claim 1, wherein the first information processing module includes at least one of: a search processing unit that searches stored information in a first information repository using first search conditions generated with input information including the entity identification information, and outputs search results including the first information; and a generation processing unit that generates entity metrics information, which is an indicator of the performance of the individual entity, as first information, with input information including the entity identification information.

3. The information processing apparatus according to claim 2, wherein the first information generated by the generation processing unit is stored in the first information repository.

4. The information processing apparatus according to claim 2, wherein the generation processing unit performs information processing using a machine learning model.

5. The information processing apparatus according to claim 2, wherein the entity metrics information includes at least one of the following: information relating to the evaluation results of the current performance of the individual entity and information relating to the prediction of the future performance of the individual entity.

6. The information processing apparatus according to claim 2, further comprising an entity-related information repository in which entity-related information useful for obtaining the first acquisition result is stored, wherein the search extension generation module executes a search process of the entity-related information repository with search conditions including the entity identification information to obtain the entity-related information related to the individual entity, and the first information processing module takes the information including the entity identification information and the entity-related information as input and outputs the first acquisition result.

7. The information processing apparatus according to claim 6, wherein the entity-related information consists of multiple types of information, and the entity identification information consists of a part of the multiple types of information.

8. The information processing apparatus according to claim 1, wherein the first acquisition result does not include information that can identify the individual entity.

9. The information processing apparatus according to claim 1, wherein the search conditions for the second information repository do not include information that can identify the individual entities.

10. The information processing apparatus according to claim 1, wherein the inference request does not include information that can identify the individual entities.

11. The information processing apparatus according to claim 1, wherein the large-scale language model is managed in a closed environment with restricted access.

12. The information processing apparatus according to claim 1, wherein the first information processing module is managed in a closed environment with restricted access.

13. The information processing apparatus according to claim 1, wherein the control unit further comprises a display signal generation unit that receives the inference results from the large-scale language model and generates a display signal that includes at least a portion of the inference results.

14. The information processing apparatus according to claim 1, wherein the control unit further comprises a training data creation unit that creates training data based on the inference request, the inference result, and any modifications to the inference result.

15. An information processing program that generates an inference request to be input to a machine-learned large-scale language model based on input information including entity identification information that can identify individual entities, transmits the inference request to the large-scale language model, and receives an inference result from the large-scale language model, comprising a search extension generation process that generates the inference request using information that does not directly include the entity identification information but is related to the entity identification information, based on a first acquisition result obtained by executing a first information processing and including first information linked to the individual entities identified by the entity identification information, and a second acquisition result obtained by searching a second information repository that stores second information including information other than the first information, wherein the search extension generation process performs the first information processing before searching the second information repository, and performs a process to set search conditions for the second information repository based on the obtained first acquisition result.

16. The information processing program according to claim 15, wherein the first information processing includes: a search process that searches stored information in a first information repository using first search conditions generated with information including the entity identification information as input, and outputs search results including the first information; and a generation process that generates entity metrics information, which is an indicator of the performance of the individual entity, as the first information, with information including the entity identification information as input.

17. An information processing program according to claim 16, wherein the program performs a search process on an entity-related information repository where entity-related information useful for obtaining the first acquisition result is stored, using a search condition that includes the entity identification information to obtain the entity-related information related to the individual entity, and in the first information processing, takes information including the entity identification information and the entity-related information as input and outputs the first acquisition result.

18. An information processing system comprising: a first device having input / output functions; a second device having data storage functions; and a third device having a control unit, wherein the first device is capable of communicating with the second and third devices, the second device having a model storage device storing a machine learning-trained large-scale language model, the control unit of the third device having a search extension generation module that generates inference requests to be input to the large-scale language model based on input information input from the first device, including entity identification information capable of identifying individual entities, the first device having a transmitting / receiving unit that transmits the inference requests input from the third device to the second device having the large-scale language model and receives inference results from the second device, the search extension generation module having: a first acquisition result obtained from the first information processing module, including first information linked to the individual entities identified by the entity identification information; and a second acquisition result obtained by searching a second information repository owned by the second device that stores second information including information other than the first information, An information processing system characterized in that, based on the above, it generates an inference request using information that does not directly include the entity identification information but is related to the entity identification information, and the search extension generation module executes processing in the first information processing module before searching the second information repository, and sets the search conditions for the second information repository based on the obtained first acquisition result.

19. The information processing system according to claim 18, wherein at least two of the first device, the second device, and the third device are integrated.

20. The information processing system according to claim 19, wherein the first information processing module includes at least one of: a search processing unit that searches stored information in a first information repository owned by the second device using first search conditions generated with input information including the entity identification information, and outputs search results including the first information; and a generation processing unit that generates entity metrics information, which is an indicator of the performance of the individual entity, as first information, with input information including the entity identification information.

21. The information processing system according to claim 20, wherein the second device further has an entity-related information repository in which entity-related information useful for obtaining the first acquisition result is stored, the search extension generation module executes a search process of the entity-related information repository with search conditions including the entity identification information to obtain the entity-related information related to the individual entity, and the first information processing module takes the information including the entity identification information and the entity-related information as input and outputs the first acquisition result.

22. The information processing system according to any one of claims 18 to 21, wherein at least one of the first device, the second device, and the third device is composed of a plurality of devices.

23. An information processing system having a plurality of computer devices constituting a peer-to-peer network, wherein at least one of the plurality of computer devices has a search extension generation module that generates an inference request to be input to a machine-learned large-scale language model based on input information, at least one of the plurality of computer devices has a transmitting and receiving unit that transmits the inference request to the large-scale language model and receives the inference result from the large-scale language model, at least one of the plurality of computer devices has an input information generation unit that generates information to be input to the search extension generation module, at least one of the plurality of computer devices has a display signal generation unit that generates a display signal including at least a part of the inference result, and the search extension generation module generates the inference request using information that does not directly include the entity identification information but is related to the entity identification information, based on: a first acquisition result obtained from a first information processing module and including first information linked to individual entities identified by entity identification information, and a second acquisition result obtained by searching a second information repository that stores second information including information other than the first information, The information processing system is characterized in that the search extension generation module performs processing in the first information processing module before searching the second information repository, and sets search conditions for the second information repository based on the obtained first acquisition results.

24. The information processing system according to claim 23, wherein the first information processing module includes at least one of: a search processing unit that searches stored information in a first information repository using first search conditions generated with information including the entity identification information as input, and outputs search results including the first information; and a generation processing unit that generates entity metrics information, which is an indicator of the performance of the individual entity, as first information, with information including the entity identification information as input.

25. The information processing system according to claim 24, wherein at least one of the plurality of computer devices has an entity-related information repository in which entity-related information useful for obtaining the first acquisition result is stored, the search extension generation module executes a search process of the entity-related information repository with search conditions including the entity identification information to obtain the entity-related information related to the individual entity, and the first information processing module takes the information including the entity identification information and the entity-related information as input and outputs the first acquisition result.

26. An information processing method comprising: generating an inference request to be input to a machine-learned large-scale language model based on input information including entity identification information capable of identifying individual entities; transmitting the inference request to the large-scale language model; and receiving an inference result from the large-scale language model, wherein the method executes a search extension generation process to generate an inference request using information that does not directly include the entity identification information but is related to the entity identification information, based on: a first acquisition result obtained by executing a first information processing process and including first information associated with the individual entities identified by the entity identification information; and a second acquisition result obtained by searching a second information repository that stores second information including information other than the first information; and in the search extension generation process, the first information processing is performed before searching the second information repository, and a process is executed to set search conditions for the second information repository based on the obtained first acquisition result.

27. The information processing method according to claim 26, wherein the first information processing includes: a search process that searches stored information in a first information repository using first search conditions generated with information including the entity identification information as input, and outputs search results including the first information; and a generation process that generates entity metrics information, which is an indicator of the performance of the individual entity, as the first information, with information including the entity identification information as input.

28. The information processing method according to claim 27, wherein the search process of an entity-related information repository where entity-related information useful for obtaining the first acquisition result is stored is performed using a search condition that includes the entity identification information to obtain the entity-related information related to the individual entity, and in the first information processing, the information including the entity identification information and the entity-related information is taken as input and the first acquisition result is output.

29. The information processing device according to claim 1, wherein the individual entity is a natural person, the first information includes one or more selected from the group consisting of health risk data, stress visualization data, and health risk fluctuation data of the natural person, the second information includes one or more selected from the group consisting of medical papers, health management-related information, nutrition and diet-related information, exercise function-related information, and physician's findings-related information of the natural person, and the inference result includes information concerning the natural person, including at least one selected from the group consisting of a health prediction report, a disease prediction report, a health advice report, a proposal regarding medical expenses, and a rehabilitation plan.

30. The information processing device according to claim 6, wherein the individual entity is a natural person, and the entity-related information includes at least one of individual health checkup information and social insurance medical fee statement information relating to the natural person.

31. The information processing apparatus according to claim 1, wherein the individual entity is a mobile device, the first information includes one or more selected from the group consisting of a future risk assessment and a physical useful life assessment of the mobile device, the second information includes general information relating to the mobile device, and the inference result includes information for creating a contract relating to damage insurance for the mobile device.

32. The information processing apparatus according to claim 6, wherein the individual entity is a mobile device, and the entity-related information includes one or more selected from the group consisting of the serial number, year of manufacture, repair information, inspection information, accident history, and structural drawings of the mobile device, and the history information of the owner of the mobile device.

33. The information processing device according to claim 1, wherein the individual entity is an artificial structure, the first information includes at least one selected from the group consisting of the need for repair of the artificial structure and the amount required for such repair, the second information includes general information relating to an estimate of the real estate value of the artificial structure, and the inference result includes information for creating at least one selected from the group consisting of a property insurance contract for the artificial structure, an engineering report, a real estate investment information report, and a maintenance plan to extend the service life of the artificial structure.

34. The information processing device according to claim 6, wherein the individual entity is an artificial structure, and the entity-related information includes one or more selected from the group consisting of a number that identifies the installation location of the artificial structure, and the purchase history and repair history of the artificial structure.

35. The information processing apparatus according to claim 33 or claim 34, wherein the artificial structure is a building.

36. The information processing device according to claim 1, wherein the individual entity is a learning subject, the first information includes at least one selected from the group consisting of learning achievement evaluation results, personality analysis results, and mental fatigue level check results of the learning subject, the second information includes general information about the learning subject, at least one selected from the group consisting of learning subject material information and learning guidance literature, and the inference result includes information for creating at least one selected from the group consisting of individual learning plan, individualized advice, and learning progress management information.

37. The information processing device according to claim 6, wherein the individual entity is a learning subject, and the entity-related information includes one or more selected from the group consisting of the learning subject's identification number, periodic and non-periodic exam results, submitted reports, and learning history.

38. The information processing apparatus according to claim 1, wherein the individual entity is a useful organism, the first information includes at least one selected from the group consisting of brand value information, product characteristics information, rarity information, and growth inhibitor-related information of the useful organism, the second information includes general information about at least one selected from the group consisting of methods of use and safety of the useful organism, and the inference result includes information for creating at least one selected from the group consisting of explanatory materials for business financing targeting the useful organism, materials contributing to sales promotion of the useful organism, and materials for creating a brand strategy relating to the useful organism.

39. The information processing device according to claim 6, wherein the individual entity is an agricultural organism, and the entity-related information includes one or more selected from the group consisting of a farmland number, a seedling registration number, an organic certification number, a pesticide registration number, and a place of origin certification number relating to the agricultural organism.

40. The information processing device according to claim 6, wherein the individual entity is an aquatic resource, and the entity-related information includes one or more selected from the group consisting of catch number, notification number for specific Class I aquatic animals and plants, lot number, fisherman registration number, fishing permit number, fishing vessel registration number, catch tracking number, fishery resource management number, fishing right number, aquatic product quality certification number, and catch origin certification number, export fisheries business registration number, general importer certification number, and manufacturer importer certification number, relating to the aquatic resource.

41. The information processing device according to claim 6, wherein the individual entity is livestock, and the entity-related information includes one or more selected from the group consisting of an individual identification number, ear tag number, microchip number, livestock registration number, DNA identification code, farm identification number, movement / trade identification number, and international identification number of the livestock.

42. The information processing apparatus according to claim 1, wherein the individual entity is a mining resource, the first information includes at least one piece of information selected from the group consisting of the utility value, purity characteristics, rarity, means of transport, and transport routes of the mining resource, and the geopolitical location of the mining site of the mining resource, the second information includes general information concerning the safety, commodity characteristics, methods of use, and processing and transport costs of the mining resource, and the political situation of the mining site, and the inference result includes information for creating at least one selected from the group consisting of a security strategy and a commercialization brand strategy for the mining resource.

43. The information processing device according to claim 6, wherein the individual entity is a mining resource, and the entity-related information includes one or more selected from the group consisting of a mine registration number, a mineral mining permit number, a mineral import / export permit number, a mineral product quality certification number, a mineral processing / processing facility registration number, a mining company number, a mine worker registration number, a mineral trading registration number, a mining environmental protection number, a mining permit number, a mining machinery registration number, a mineral origin certification number, a mineral product safety standard compliance number, a mining pollution prevention number, and a mineral mining right number, relating to the mining resource.

44. The information processing apparatus according to claim 1, wherein the individual entity is an industrial product, the first information includes at least one piece of information selected from the group consisting of the manufacturer, price, year of manufacture, accident history, and manual of the industrial product, the second information includes at least one piece of information selected from the group consisting of general information about competing products of the industrial product in the market, claim information for the industrial product and the competing products, and general information about the destinations of the industrial product and the competing products, and the inference result includes information for creating at least one piece of information selected from the group consisting of the existence of repair parts for the industrial product, safety reports, and warranty materials for export of used industrial products.

45. The information processing apparatus according to claim 6, wherein the individual entity is an industrial product, and the entity-related information includes one or more selected from the group consisting of a manufacturer registration number, product identification number, quality certification number, manufacturing number, machine equipment number, manufacturing license number, manufacturer ID number, quality certification number, import / export control number, trademark registration number, manufacturing facility license number, patent number, product registration number, and environmental standards compliance number, relating to the industrial product.

46. ​​The information processing apparatus according to claim 1, wherein the individual entity is a legal entity, the first information includes at least one piece of information selected from the group consisting of a quantitative valuation of the individual stock price and a financial valuation of the legal entity, the second information includes fundamental information of the legal entity, and the inference result includes information for creating an overall stock price valuation of the legal entity.

47. The information processing device according to claim 6, wherein the individual entity is a legal entity, and the entity-related information includes one or more selected from the group consisting of a corporate number, commercial registration number, securities code / ticker symbol, EDINET code (Electronic Disclosure for Investors' NETwork Code), LEI code (Legal Entity Identifier), qualified invoice issuer number (T number), tax office reference number, and customs duty number (Harmonized System Code), relating to the legal entity.

48. The information processing device according to claim 6, wherein the individual entity is a service provider, and the entity-related information includes one or more selected from the group consisting of a business license number, customer management number, license number, service contract number, operating license number, franchise number, employee number, consumer claim number, invoice number, customer service number, product / service identification number (SKU number), and delivery management number relating to the service provider.

49. The information processing device according to claim 1, wherein the individual entity is an IT company, the first information includes information relating to the individual company's services, the second information includes one or more selected from the group consisting of security information, lead time, and unit price information, and the inference result includes information for creating business order support information for the IT company, which includes one or more selected from the group consisting of suppliers, unit price, lead time, and accuracy.

50. The information processing device according to claim 6, wherein the individual entity is an IT company, and the entity-related information includes one or more selected from the group consisting of a corporate number, software registration number, domain name registration number, software product ID number, IP address number, engineer qualification number, patent number, database management number, hardware serial number, Internet service provider (ISP) number, IT contract number, server management number, and cloud service account number relating to the IT company.

51. The information processing device according to claim 1, wherein the individual entity is a copyrighted work, the first information includes at least one selected from the group consisting of the content of the copyrighted work, book information, and content information, the second information includes one or more selected from the group consisting of the author of the copyrighted work, general information about literature, general information about the possession of the content, and information about past film adaptations of the copyrighted work, and the inference result includes information for creating one or more selected from the group consisting of availability information, evaluation / introduction text, screenplay for a film, and manga storyboard for the copyrighted work.

52. The information processing device according to claim 6, wherein the individual entity is a copyrighted work, and the entity-related information includes one or more selected from the group consisting of an International Standard Book Number (ISBN), an International Standard Serial Number (ISSN), a JAN code, (National Diet Library Bibliographic ID (NDL Bibliographic ID)), and a library management number relating to the copyrighted work.

53. The information processing device according to claim 1, wherein the individual entity is a performer, the first information includes at least one selected from the group consisting of the performer's contribution to sabermetrics, contribution to distribution, contribution to audience attendance, contribution to music distribution, and contribution to merchandise sales, the second information includes one or more selected from the group consisting of the performer's general information regarding the commercialization of sports and entertainment, and general information regarding scandals, and the inference result includes information useful for creating a contract for the performer.

54. The information processing device according to claim 6, wherein the individual entity is a performer, and the entity-related information includes one or more selected from the group consisting of an ID issued by the entertainer's agency, a performer number, a copyright registration number, an International Standard Score Number (ISRC), an International Standard Recording Code (ISMN), a movie / music album identification number, a fan card number, athlete number, an Olympic athlete number, a sports team number, a competition registration number, a leaderboard number, athlete ID number, and a sports license number, relating to the performer.

55. The information processing device according to claim 1, wherein the individual entity is a gambling competition subject, the first information includes at least one selected from the group consisting of pedigree, pedigree performance, performance, and track compatibility of the gambling competition subject, the second information includes one or more selected from the group consisting of racehorse characteristic information, rival horse information, racetrack information, past races, and weather and track information of the gambling competition subject, and the inference result includes information that contributes to at least one selected from the group consisting of race-specific win / loss predictions and stallion value data for the gambling competition subject.

56. The information processing device according to claim 6, wherein the individual entity is a gambling competition entity, and the entity-related information includes one or more selected from the group consisting of Racehorse Name Registration, Studbook Number, Horse ID Number, JRA Registration Number, Breeder & Owner Information, and Race Record, relating to the gambling competition entity.

57. The information processing apparatus according to claim 1, wherein the individual entity is a satellite, the first information includes one or more selected from the group consisting of the satellite's history, motion, orbit, current position, future position, and probability of collision with an object in Earth orbit, the second information includes one or more selected from the group consisting of the satellite's characteristics, constellation information, and registration information, and the inference result includes at least one selected from the group consisting of a strategy for debris removal, on-orbit space debris information, and a plan for ensuring a stable orbit for the satellite.

58. The information processing device according to claim 6, wherein the individual entity is a satellite, and the entity-related information includes one or more selected from the group consisting of the International Designator (NSSDC) ID, NORAD ID (Satellite Catalog Number / SATCAT), COSPAR ID (Committee on Space Research) ID, GSAT number (Geostationary Satellite Number), and ITU registration number (International Telecommunication Union Registration) relating to the satellite.

59. The information processing device according to claim 1, wherein the individual entity is a crisis or threat caused by an external attack, the first information includes one or more selected from the group consisting of the type of crisis or threat caused by an external attack, an assessment of the threat level, and an encounter history, the second information includes one or more selected from the group consisting of the general model number, characteristic information, and past occurrence history information of the crisis or threat caused by an external attack, and the inference result includes information on at least one selected from the group consisting of a strategy as a security measure against the crisis or threat caused by an external attack, a tactical plan for defense, and threat warning information for the target to be defended.

60. The information processing device according to claim 6, wherein the individual entity is a crisis or threat resulting from an external attack, and the entity-related information includes one or more selected from the group consisting of CVE numbers (Common Vulnerabilities and Exposures), MITRE ATT&CK identifiers, malware-specific identifiers (MD5 / SHA hashes), AV vendor-specific identification numbers, international malware databases (such as MalwareBazaar), and weapon lists.

61. The information processing device according to claim 1, wherein the individual entity is the allocation of personnel in an organization, the first information includes one or more selected from the group consisting of the characteristics, history, competence assessment, personality assessment, and relationships with other personnel of a candidate for allocation of personnel in the organization, suitable tasks for the candidate for allocation of personnel, and a predicted assessment of the candidate for allocation of personnel's contribution to the profitability and efficiency of the target tasks, the second information includes one or more selected from the group consisting of the abilities, proficiency, experience, personality, and number of personnel required for the tasks of the target organization in the allocation of personnel in the organization, and the inference result includes at least one piece of information selected from the group consisting of an allocation plan and a dynamically updated replacement plan for the candidate personnel.

62. The information processing device according to claim 6, wherein the individual entity is the allocation of personnel in an organization, and the entity-related information includes one or more selected from the group consisting of age, gender, physical strength, mental strength, experience, and past accident history of personnel who are candidates for the allocation.

63. The information processing device according to claim 1, wherein the individual entity is the movement of a bio-robot, the first information includes one or more selected from the group consisting of evaluations of the environment, function, and state of the movement of the bio-robot, the second information includes one or more selected from the group consisting of operation information and operation examples of dynamic posture control in the movement of the bio-robot, and the inference result includes information relating to the posture of the movement of the bio-robot and information relating to a program relating to the operation of joints and parts of the bio-robot associated with motion control.

64. The information processing device according to claim 6, wherein the individual entity is the operation of a bio-robot, and the entity-related information includes one or more selected from the group consisting of the size, weight, material and properties of the constituent parts, driving force, range of motion of joints, power, center of gravity, and operation characteristics of the bio-robot.

65. The information processing apparatus according to claim 1, wherein the individual entity is a product manufacturing process, the first information includes one or more selected from the group consisting of a plurality of manufacturing operations constituting the product manufacturing process, the accuracy of the product when the manufacturing operations are automated, the time and cost required to automate the manufacturing operations, and the efficiency of the automated manufacturing operations compared to conventional operations, the second information includes one or more selected from the group consisting of examples of operations performed to fully automate the product manufacturing process and examples of operations devised to fully automate the product manufacturing process, and the inference result includes information on at least one selected from the group consisting of manufacturing processes and implementation plans for further automating and improving the efficiency of the manufacturing process.

66. The information processing apparatus according to claim 6, wherein the individual entity is a product manufacturing process, and the entity-related information includes one or more selected from the group consisting of the use and characteristics of the product, the specifications and characteristics of the parts necessary for manufacturing the product, the specifications and characteristics of the equipment used for a plurality of manufacturing operations constituting the product manufacturing process, the specifications and characteristics of jigs and the like used for the manufacturing operations, the work capabilities of the people involved in manufacturing the product, the skills and characteristics required of the people involved in manufacturing the product, and the time and cost required for the manufacturing operations.

67. The information processing device according to claim 1, wherein the individual entity is one selected from a group consisting of business operations and work processes that are components of said business operations, the first information includes one or more selected from a group consisting of specifications and characteristics of said business operations and work processes, and evaluation of the processing path configuration of the business flow, evaluation of the feasibility of introducing automation, and evaluation of efficiency after conversion to automation for said business operations and work processes, the second information includes one or more selected from a group consisting of examples of implemented and devised work relating to the introduction of automation to the processing paths of said business operations and work processes and the reconfiguration of the processing paths, and the inference result includes information relating to at least one selected from a group consisting of implementation plans for introducing automation and implementation plans for reconfiguring the processing paths of said business flows for business processes aimed at improving the operational efficiency of the organization.

68. The information processing apparatus according to claim 6, wherein the individual entity is one selected from a group consisting of a business and work processes that are components of the business, and the entity-related information includes one or more selected from a group consisting of specifications and characteristics of processing work required for the business and the work processes, specifications of equipment required for the execution of the business and the work processes, the work capabilities and required skills and characteristics of the people involved in processing the business and the work processes, and the time and costs required for the business and the work processes.

69. The information processing apparatus according to claim 1, wherein the individual entities are confidential information which is information that should be kept confidential, the first information is more abstract than the confidential information and has a relatively lower need for confidentiality management, and is more clearly defined as a search condition for the second information repository than the confidential information, and the second information includes information related to the confidential information.

70. The information processing apparatus according to claim 6, wherein the individual entities are confidential information which is information that should be kept confidential, and the entity-related information includes information which enhances the anonymity of the confidential information and which is related to the confidential information and helps to clarify the search conditions of the second information repository.

Images