system

Abstract

We provide the system. [Solution] This invention relates to a drug discovery support system that integrates diverse data such as scientific papers, clinical trials, patient records, and genetic information, and autonomously and rapidly performs tasks such as extracting therapeutic targets for each disease, designing novel compounds, virtual biological simulations, toxicity and efficacy evaluations, optimization proposals, and automatic report generation using generative AI. This significantly shortens the entire drug discovery process, which previously took a long time, and enables efficient and highly accurate responses to diverse medical needs, such as rare diseases and intractable diseases, thereby improving the success rate of drug discovery and maximizing social value.

Description

TECHNICAL FIELD

[0001] The technology of the present disclosure relates to a system. BACKGROUND ART

[0002] Patent Document 1 discloses a persona chatbot control method executed by at least one processor, the method comprising: a step of receiving a user utterance; a step of adding the user utterance to a prompt including an instruction text associated with a description related to a character of the chatbot; a step of encoding the prompt; and a step of inputting the encoded prompt into a language model to generate a chatbot utterance that responds to the user utterance. PRIOR ART DOCUMENTS PATENT DOCUMENTS

[0003] PATENT DOCUMENT 1 Japanese Unexamined Patent Publication No. 2022-180282 SUMMARY OF THE INVENTION PROBLEMS TO BE SOLVED BY THE INVENTION

[0004] In the drug discovery process, there are serious challenges: the inability to fully utilize the vast and diverse data from various sources such as scientific papers, clinical trial data, patient records, and genetic information makes it difficult to discover therapeutic targets and promising compound candidates for each disease; the entire process from designing new compounds to in vivo experiments, results analysis, and proposals for next-generation compounds remains manual or partially automated, taking an average of more than 10 years and resulting in an extremely low success rate; efficient new drug development has hardly been achieved, especially in areas where research and development is difficult under conventional economic constraints, such as rare and intractable diseases; and the inability to make reliable predictions and selections in the stages prior to compound generation and in vivo testing using actual compounds often leads to failure despite enormous costs and time. To solve these problems, it is necessary to comprehensively analyze vast amounts of diverse data and execute compound design, experimental planning, simulation, results analysis, and proposals for next-generation compounds in a seamless, high-speed, and highly accurate manner using generative AI, thereby simultaneously shortening the overall drug discovery process and improving the success rate. [Means for solving the problem]

[0005] This invention provides a system that receives diverse and vast amounts of information such as scientific paper data, clinical trial data, patient record data, and genetic information data, semantically integrates this information using natural language processing and database integration technologies, extracts therapeutic targets and disease-related genes from the integrated knowledge base, designs novel compound candidates based on these using molecular design algorithms and structure generation AI, performs biological experiment simulations on the designed compound candidates using generative AI, employs molecular dynamics simulations, virtual screening technologies, and machine learning models, evaluates the efficacy and toxicity of each compound, selects promising compounds that meet predetermined thresholds, and autonomously proposes next-generation compound designs based on the selected compounds. By autonomously and continuously repeating this series of processes using generative AI, the system provides a system that simultaneously shortens the overall drug discovery process time and improves the success rate, thereby solving the above-mentioned problems. [Brief explanation of the drawing]

[0006] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Modes for carrying out the invention]

[0007] Hereinafter, an example of an embodiment of the system relating to the technology of this disclosure will be described with reference to the attached drawings.

[0008] First, let's explain the terminology used in the following explanation.

[0009] In the following embodiments, the signed processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Furthermore, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), and APU (Accelerated Processing Unit).

[0010] In the following embodiments, signed RAM (Random Access Memory) is a memory that temporarily stores information and is used as work memory by the processor.

[0011] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0012] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B". That is, "A and / or B" means that it may be only A, only B, or a combination of A and B. Further, in the present specification, when three or more matters are expressed in a linked manner using "and / or", the same concept as that for "A and / or B" is applied.

[0014] [First Embodiment]

[0015] Figure 1 illustrates an example of the configuration of a data processing system 10 according to the first embodiment.

[0016] As shown in Figure 1, the data processing system 10 includes a data processing apparatus 12 and a smart device 14. An example of the data processing apparatus 12 is a server.

[0017] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of the "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network) and the like.

[0018] The smart device 14 includes a computer 36, a reception apparatus 38, an output apparatus 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception apparatus 38, the output apparatus 40, the camera 42, and the communication I / F 44 are also connected to the bus 52.

[0019] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A receives user input via contact of a pointing body by detecting the contact of the pointing body (for example, a pen, a finger, or the like). The microphone 38B receives voice-based user input by detecting the user's voice. The control unit 46A in the processor 46 transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0020] The output device 40 includes a display 40A, a speaker 40B, and the like, and presents data to the user 20 by outputting the data in a representation form perceptible to the user 20 (for example, voice and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs voice in accordance with instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0021] The communication I / F 44 is connected to the network 54. The communication I / Fs 44 and 26 manage the exchange of various types of information between the processor 46 and the processor 28 via the network 54.

[0022] Figure 2 illustrates an example of the main functions of the data processing device 12 and the smart device 14.

[0023] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0024] Storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0025] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output processing program 60 executed on the RAM 48. The smart device 14 also has a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0026] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device having the data generation model 58. The data processing device 12 may also be a server device or a terminal device owned by a user (e.g., a mobile phone, robot, home appliance, etc.). Next, an example of processing by the data processing system 10 according to the first embodiment will be described.

[0027] (Example 1)

[0028] The flow of a specific process in Example 1 will now be described. Each part of the system described below is implemented by a data processing device 12 and a smart device 14. The data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0029] (Embodiments for carrying out the present invention)

[0030] Embodiments of the present invention will be described in more specific and detail. The system of the present invention consists of a server and a terminal. The server side is equipped with major data processing functions, an execution environment for generative AI, a database, a molecular design engine, a simulation engine, and an analysis module, while the terminal side is equipped with a user interface, data input / output functions, and lightweight data preprocessing functions. The server can be configured in various forms such as a cloud server or an on-premise server, and the terminal can be a personal computer, tablet, smartphone, or dedicated terminal.

[0031] First, during the data collection phase, researchers and medical professionals upload scientific paper data (e.g., full text, abstracts, figures and tables), clinical trial data (e.g., clinical trial protocols, patient characteristics, treatment results, outcome data), patient record data (e.g., electronic medical records, medical records, laboratory values, image data), and genetic information data (e.g., genome sequences, epigenomic information, gene expression data) to the server from their terminals. Examples include sets of papers automatically obtained via API from a paper database, patient medical history and test results extracted from an electronic medical record system of a medical institution, and FASTQ and VCF files output from a genome analysis device. Furthermore, it is possible to link with a variety of external data sources, such as patent documents, conference presentation materials, drug side effect databases, drug interaction databases, disease ontologs, protein structure data banks, and compound libraries.

[0032] The data integration unit on the server semantically integrates heterogeneous data by combining a natural language processing engine (e.g., a large-scale language model using generative AI), a structured data integration module, image analysis AI, and knowledge graph generation AI. For example, it automatically extracts disease names, drug names, molecular targets, mechanisms of action, clinical outcomes, patient characteristics, dosage, treatment duration, and side effect incidence rates from research paper texts, matches them with patient attributes and treatment outcomes from clinical trial data and patient record data, and extracts disease-related genes, mutation sites, expression patterns, and epigenetic modification information from genetic information data. For image data, it uses image recognition technology with generative AI to extract, for example, morphological features and immunohistochemical staining patterns of tumors from tissue section images and associates them with diagnostic information and treatment response. This data is structured as a knowledge graph and built on the server as a massive network with multidimensional relationships such as disease-gene-molecular target-drug-patient population-treatment outcomes-side effects-molecular structure-biological response.

[0033] The Target Extraction and Compound Design Department operates on a server utilizing generative AI. For example, the generative AI analyzes knowledge graphs and automatically extracts molecules and pathways that could serve as therapeutic targets for each disease. Specifically, it comprehensively evaluates factors such as mutation frequency and expression levels of disease-related genes, efficacy and side effect patterns of existing drugs, and treatment outcomes for each patient population, enabling the discovery of novel proteins and signaling pathways that have not been previously targeted. Furthermore, the generative AI works in conjunction with molecular design algorithms to autonomously perform diverse molecular designs, such as automatically generating novel molecular skeletons that fit into binding pockets based on the three-dimensional structural information of target proteins, designing derivatives that partially modify the structure of existing drugs to reduce side effects, designing multifunctional molecules that act on multiple targets simultaneously, and designing biopharmaceutical candidates such as peptides and nucleic acid drugs. For example, the generative AI can learn from past drug efficacy and side effect data and propose optimal compound structures while predicting the impact of specific structural motifs on efficacy and safety.

[0034] For the designed compound candidates, the simulation and analysis unit on the server performs virtual biological experiments using generative AI. For example, molecular dynamics simulation AI predicts the binding affinity, stability, and dynamics of the compound and target protein at the atomic level. Virtual screening AI rapidly screens hundreds of thousands to millions of compound candidates and evaluates their binding ability and selectivity to the target. Using machine learning models, it predicts various toxicity risks, such as hepatotoxicity, nephrotoxicity, cardiotoxicity, teratogenicity, and immunotoxicity, and predicts pharmacokinetic (ADME: absorption, distribution, metabolism, and excretion) characteristics. Furthermore, the generative AI learns from past experimental failures and adverse event occurrences, and if, for example, a specific functional group or molecular size causes toxicity or metabolic dysfunction, it reflects that knowledge in the design of next-generation compounds. For example, it can also predict selectivity for mutant forms of the target protein and efficacy considering the genetic diversity of each patient population.

[0035] The simulation results are aggregated and analyzed on the server, and the analysis unit automatically selects only promising compounds that meet predetermined thresholds based on multidimensional evaluation indicators such as efficacy score, toxicity score, pharmacokinetic score, and ease of manufacture score. For example, selection criteria can be set by combining multiple conditions, such as advancing only compounds with an efficacy score of 0.85 or higher, a toxicity score of less than 0.15, a pharmacokinetic score of 0.8 or higher, and an ease of manufacture score of 0.7 or higher to the next step. For the selected compounds, the generative AI repeatedly performs structural optimization and proposes novel modifications, automatically executing various optimizations such as adjusting the balance of hydrophilicity and lipophilicity, improving target selectivity, improving metabolic stability, reducing synthesis costs, and designing patent-avoiding structures. Furthermore, the generative AI can be applied to, for example, multi-target compounds that act on multiple targets simultaneously, prodrug design suitable for drug delivery systems, biopharmaceutical design such as peptides and oligonucleotides, and the design of antibody drugs and cell therapies.

[0036] On the terminal side, a user interface is provided that allows researchers and medical professionals to view and manipulate the processing status on the server, simulation results, proposed compound designs, and the rationale and design intent of the AI ​​analysis in real time. For example, it has a variety of functions, such as rotating and zooming in and out of a 3D model of the compound structure on the terminal screen, viewing graphs, heatmaps, distribution maps, and time-series changes of simulation results, having the AI ​​explain the design intent and rationale of the next-generation compound proposed by the AI ​​in natural language, comparing it with past experimental data and failure cases, and downloading reports and draft papers automatically generated by the AI. It is also possible to upload new data from the terminal, change simulation conditions (for example, specifying variant types of target proteins, specifying attributes of patient populations, changing dosages and administration routes), and request additional analysis or redesign by the AI.

[0037] Furthermore, the terminal side can implement a variety of functions that comprehensively support the entire drug discovery process, such as a collaboration function that allows multiple researchers to use the system simultaneously to share, compare, and discuss their analysis results and design proposals, an AI-powered automatic report generation and paper writing support function, and an automatic generation function for application documents to regulatory authorities.

[0038] Thus, the system of the present invention realizes a data integration unit, a target extraction / compound design unit, and a simulation / analysis unit on the server side, and by making generative AI essential and central to each unit, it can perform large-scale data analysis, molecular design, virtual experiments, results analysis, next-generation compound proposals, user feedback learning, collaboration, and automatic report generation end-to-end, autonomously, and at high speed, which were not possible in conventional drug discovery processes. For example, concrete effects can be obtained such as completing compound screening and structural optimization, which previously took several years, in a few days to a few weeks, efficiently creating new drug candidates for rare and intractable diseases beyond conventional economic constraints, realizing drug discovery innovation through collaboration between AI and humans, and significantly streamlining the application process to regulatory authorities.

[0039] (System Configuration)

[0040] The system according to this embodiment comprises a data collection unit, a data integration unit, a knowledge graph generation unit, a target extraction unit, a compound design unit, a simulation unit, an analysis unit, an optimization proposal unit, a user interface unit, a feedback learning unit, and a report generation unit. The data collection unit automatically or manually collects data from a variety of external data sources, such as scientific paper data, clinical trial data, patient record data, genetic information data, patent documents, conference presentation materials, drug side effect databases, drug interaction databases, disease ontologs, protein structure data banks, and compound libraries. For example, it can obtain full texts and abstracts of papers from a paper database via API, extract patient medical history and test values ​​from an electronic medical record system of a medical institution, import genome sequence data and mutation information output from a genome analysis device, and obtain side effect incidence rates and drug interaction information from an external drug side effect database. The data integration unit semantically integrates collected heterogeneous data using natural language processing engines, structured data integration modules, and image analysis AI. For example, it automatically extracts disease names, drug names, molecular targets, mechanisms of action, clinical outcomes, patient characteristics, dosage, treatment duration, and side effect incidence rates from research paper texts, matches them with patient attributes and treatment outcomes from clinical trial data and patient record data, extracts disease-related genes, mutation sites, expression patterns, and epigenetic modification information from genetic information data, and extracts morphological features and immunohistochemical staining patterns from image data to correlate with diagnostic information and treatment response. In short, it centrally organizes and integrates diverse data. The knowledge graph generation unit generates a massive knowledge graph with multidimensional relationships between diseases, genes, molecular targets, drugs, patient populations, treatment outcomes, side effects, molecular structures, and biological responses based on the integrated data. For example, it visualizes and structures networks of diseases and related genes, relationships between drugs and side effects, differences in treatment outcomes across patient populations, and correlations between molecular structures and biological responses. The target extraction unit analyzes knowledge graphs to automatically extract molecules and pathways that could serve as therapeutic targets for each disease. It comprehensively evaluates factors such as the mutation frequency and expression levels of disease-related genes, the efficacy and side effect patterns of existing drugs, and treatment outcomes for different patient populations, enabling the discovery of novel proteins and signaling pathways that have not been previously targeted.The Compound Design Department, based on therapeutic targets extracted by the Target Extraction Department, automatically generates novel compound candidates using molecular design algorithms and structure generation AI that utilize generative AI. For example, it generates novel molecular skeletons that fit into binding pockets based on the three-dimensional structural information of target proteins, designs derivatives that reduce side effects by partially modifying the structure of existing drugs, and designs biopharmaceutical candidates such as multifunctional molecules, peptides, and nucleic acid drugs that act on multiple targets simultaneously. The Simulation Department uses generative AI to perform molecular dynamics simulations, virtual screening, and toxicity and pharmacokinetic predictions using machine learning models on the designed compound candidates. For example, it predicts the binding affinity, stability, and dynamics of compounds and target proteins at the atomic level, rapidly screens hundreds of thousands to millions of compound candidates to evaluate their binding ability and selectivity to targets, predicts various toxicity risks and pharmacokinetic characteristics such as hepatotoxicity, nephrotoxicity, cardiotoxicity, teratogenicity, and immunotoxicity, and learns from past experimental failures and adverse event occurrences to predict the impact of specific structural motifs on efficacy and safety. The analysis unit aggregates the output results from the simulation unit and automatically selects only promising compounds that meet predetermined thresholds based on multidimensional evaluation indicators such as efficacy score, toxicity score, pharmacokinetic score, and ease of manufacture score. It can set selection criteria that combine multiple conditions, such as advancing only compounds with an efficacy score of 0.85 or higher, a toxicity score of less than 0.15, a pharmacokinetic score of 0.8 or higher, and an ease of manufacture score of 0.7 or higher to the next step. The optimization proposal unit repeatedly performs structural optimization and proposes novel modifications for the selected compounds using a generative AI. It automatically executes a variety of optimizations, such as adjusting the balance of hydrophilicity and lipophilicity, improving target selectivity, improving metabolic stability, reducing synthesis costs, designing patent-avoiding structures, designing multi-target compounds that act on multiple targets simultaneously, designing prodrugs suitable for drug delivery systems, and designing biopharmaceuticals.The user interface section provides a screen that allows researchers and medical professionals to view and manipulate the processing status, simulation results, compound design proposals, and AI analysis rationale and design intent on the server in real time. It offers a variety of functions, such as rotating and scaling 3D models of compound structures, viewing graphs, heatmaps, distribution maps, and time-series changes of simulation results, having the AI ​​explain the design intent and rationale of next-generation compounds proposed by the AI ​​in natural language, comparing with past experimental data and failure cases, and downloading reports and draft papers automatically generated by the AI. The feedback learning section collects feedback from users, allowing researchers to input usefulness and areas for improvement regarding the AI's proposals, and the generative AI to learn from this feedback and reflect it in future designs and analyses. The report generation section uses the generative AI to automatically generate reports, draft papers, and application documents for regulatory authorities based on analysis results, design proposals, and simulation results. It can automatically create detailed documents, such as summaries of the design rationale and simulation results of new compounds, comparisons with past data, and assessments of compliance with regulatory requirements.

[0041] Specific examples of prompts to be fed into the generative AI necessary for implementing the present invention include, for example, "Integrate all research paper data, clinical trial data, patient record data, and genetic information data related to disease X, and extract molecular pathways that could serve as novel therapeutic targets for disease X," "Automatically design a novel molecular skeleton with high binding affinity and low toxicity risk based on the three-dimensional structural information of target protein Y," "Perform molecular dynamics simulations on compound Z and predict binding stability to target protein Y, pharmacokinetic properties, and major toxicity risks," "Learn from past failures and adverse event occurrences and propose structural optimization plans to avoid similar risks," "Calculate efficacy scores, toxicity scores, and pharmacokinetic scores based on the simulation results and select only compounds that meet predetermined thresholds," "For the selected compounds, propose multiple optimization plans considering hydrophilic / lipidophilic balance, target selectivity, ease of synthesis, patent avoidance structures, etc.," and "Automatically generate a detailed report including analysis results and design rationale, and summarize it in natural language that is easy for the user to understand." These prompt statements are generated and input sequentially and dynamically when activating the generative AI in each part of the system, playing a central role in realizing an autonomous and highly accurate drug discovery process for the entire system.

[0042] (Implementation Steps)

[0043] Step 1: Data collection and input In this step, data is automatically or manually collected from a variety of external data sources, including scientific paper data, clinical trial data, patient record data, genetic information data, patent documents, conference presentation materials, drug adverse event databases, drug interaction databases, disease ontologs, protein structure data banks, and compound libraries, and aggregated on a server. For example, full texts and abstracts of papers are obtained from paper databases via APIs, patient medical history and test results are extracted from medical institutions' electronic medical record systems, and genome sequence data and mutation information are imported from genome analysis equipment. Furthermore, all relevant data is comprehensively collected, such as adverse event rates and drug interaction information obtained from external drug adverse event databases.

[0044] Step 2: Data Integration and Knowledge Graph Generation In this step, the data integration unit on the server semantically integrates the collected heterogeneous data using a natural language processing engine, structured data integration module, image analysis AI, etc., and the knowledge graph generation unit generates a massive knowledge graph with multidimensional relationships between diseases, genes, molecular targets, drugs, patient populations, treatment outcomes, side effects, molecular structures, and biological responses. For example, it automatically extracts disease names, drug names, molecular targets, mechanisms of action, clinical outcomes, patient characteristics, dosage, treatment duration, and side effect incidence rates from the text of research papers, matches them with patient attributes and treatment outcomes from clinical trial data and patient record data, extracts disease-related genes, mutation sites, expression patterns, and epigenetic modification information from genetic information data, and extracts morphological features and immunohistochemical staining patterns of tumors from image data and associates them with diagnostic information and treatment response. A specific example of a prompt when using the generation AI is, "Integrate all research paper data, clinical trial data, patient record data, and genetic information data related to disease X, and extract molecular pathways that could be novel therapeutic targets for disease X."

[0045] Step 3: Therapeutic target extraction and compound design In this step, the knowledge graph is analyzed to automatically extract molecules and pathways that could serve as therapeutic targets for each disease. The target extraction unit then comprehensively evaluates factors such as the mutation frequency and expression levels of disease-related genes, the efficacy and side effect patterns of existing drugs, and treatment outcomes for each patient population, discovering novel proteins and signaling pathways that have not been previously targeted. Based on the therapeutic targets extracted by the target extraction unit, the compound design unit automatically generates novel compound candidates using molecular design algorithms and structure generation AI that utilize generative AI. For example, it can generate novel molecular skeletons that fit into the binding pocket based on the three-dimensional structural information of target proteins, design derivatives that reduce side effects by partially modifying the structure of existing drugs, or design biopharmaceutical candidates such as multifunctional molecules, peptides, and nucleic acid drugs that act on multiple targets simultaneously. Specific examples of prompts when using generative AI include: "Automatically design a novel molecular skeleton with high binding affinity and low toxicity risk based on the three-dimensional structural information of target protein Y," and "Extract novel therapeutic targets for disease Z and propose multiple optimal compound structures."

[0046] Step 4: Virtual bio-simulation and evaluation In this step, the simulation unit uses generative AI to perform molecular dynamics simulations, virtual screening, and toxicity and pharmacokinetic predictions using machine learning models for the designed compound candidates. For example, it can predict the binding affinity, stability, and dynamics of compounds to target proteins at the atomic level, rapidly screen hundreds of thousands to millions of compound candidates to evaluate their binding ability and selectivity to targets, predict various toxicity risks and pharmacokinetic properties such as hepatotoxicity, nephrotoxicity, cardiotoxicity, teratogenicity, and immunotoxicity, and predict the impact of specific structural motifs on efficacy and safety by learning from past experimental failures and adverse event occurrences. Specific examples of prompts when using generative AI include: "Perform molecular dynamics simulations for compound Z and predict its binding stability to target protein Y, pharmacokinetic properties, and major toxicity risks," and "Learn from past failures and adverse event occurrences and propose structural optimizations to avoid similar risks."

[0047] Step 5: Analysis, Selection, and Optimization Proposal In this step, the analysis unit aggregates the output results from the simulation unit and automatically selects only promising compounds that meet predetermined thresholds based on multidimensional evaluation indicators such as efficacy score, toxicity score, pharmacokinetic score, and ease of manufacture score. The optimization proposal unit repeatedly performs structural optimization and proposes novel modifications for the selected compounds using a generative AI, automatically executing a variety of optimizations such as adjusting the balance of hydrophilicity and lipophilicity, improving target selectivity, improving metabolic stability, reducing synthesis costs, designing patent-avoiding structures, designing multi-target compounds that act on multiple targets simultaneously, designing prodrugs suitable for drug delivery systems, and designing biopharmaceuticals. Specific examples of prompts when using the generative AI include: "Calculate efficacy score, toxicity score, and pharmacokinetic score based on the simulation results and select only compounds that meet predetermined thresholds," and "For the selected compounds, propose multiple optimization options considering the balance of hydrophilicity and lipophilicity, target selectivity, ease of synthesis, patent-avoiding structures, etc."

[0048] Step 6: Visualization, operation, and feedback through the user interface In this step, the user interface provides researchers and medical professionals with a screen that allows them to view and manipulate the server's processing status, simulation results, compound design proposals, and the AI's analysis rationale and design intent in real time. For example, it offers a variety of functions, such as rotating and scaling a 3D model of the compound structure, viewing simulation result graphs, heatmaps, distribution diagrams, and time-series changes, having the AI ​​explain the design intent and rationale of the next-generation compound it proposed in natural language, comparing it with past experimental data and failure cases, and downloading reports and draft papers automatically generated by the AI. Furthermore, users can upload new data, change simulation conditions, request additional analysis or redesign by the AI, and input feedback on the AI's proposals, allowing the generative AI to learn from this feedback and reflect it in future designs and analyses.

[0049] Step 7: Report generation and document output In this step, the report generation unit uses a generative AI to automatically generate reports, paper drafts, and application documents for regulatory authorities based on analysis results, design proposals, and simulation results. For example, it automatically creates detailed documents including the design rationale for new compounds, summaries of simulation results, comparisons with past data, and assessments of compliance with regulatory requirements. Specific examples of prompts when using the generative AI include: "Automatically generate a detailed report including analysis results and design rationale, and summarize it in natural language that is easy for the user to understand," and "Automatically generate application documents for regulatory authorities in a format that conforms to the latest regulatory requirements."

[0050] (Specific use cases)

[0051] For example, when the goal is to develop new therapeutic drugs for rare diseases, the data collection department first comprehensively collects relevant information from worldwide scientific paper data, clinical trial data, patient record data, genetic information data, patent documents, drug adverse event databases, disease ontologs, and protein structure databanks related to rare diseases. From the paper database, full texts and abstracts are automatically retrieved using the rare disease name and related gene name as keywords. From the electronic medical record systems of medical institutions, the medical history, test values, treatment history, and adverse event occurrence information of patients with the relevant disease are extracted. From genome analysis equipment, genome sequences and mutation information for each patient are imported. Furthermore, adverse event incidence rates and drug interaction information of existing drugs are obtained from external drug adverse event databases, and disease classification and three-dimensional structure information of target proteins are obtained from disease ontologs and protein structure databanks.

[0052] The data integration unit semantically integrates these heterogeneous data using natural language processing engines, structured data integration modules, and image analysis AI, and the knowledge graph generation unit generates a massive knowledge graph with multidimensional relationships between diseases, genes, molecular targets, drugs, patient populations, treatment outcomes, side effects, molecular structures, and biological responses. For example, it automatically extracts disease names, drug names, molecular targets, mechanisms of action, clinical outcomes, patient characteristics, dosage, treatment duration, and side effect incidence rates from research paper texts, matches them with patient attributes and treatment outcomes from clinical trial data and patient record data, extracts disease-related genes, mutation sites, expression patterns, and epigenetic modification information from genetic information data, and extracts morphological features and immunohistochemical staining patterns of tumors from image data and associates them with diagnostic information and treatment response.

[0053] The Target Extraction Unit analyzes knowledge graphs to automatically extract molecules and pathways that could serve as therapeutic targets for each rare disease. It comprehensively evaluates factors such as mutation frequency and expression levels of disease-related genes, the efficacy and side effect patterns of existing drugs, and treatment outcomes for each patient population, discovering novel proteins and signaling pathways that have not been previously targeted. The Compound Design Unit, based on the therapeutic targets extracted by the Target Extraction Unit, automatically generates novel compound candidates using molecular design algorithms and structural generation AI that utilize generative AI. For example, it can generate novel molecular skeletons that fit into binding pockets based on the three-dimensional structural information of target proteins, design derivatives that partially modify the structure of existing drugs to reduce side effects, or design biopharmaceutical candidates such as multifunctional molecules, peptides, and nucleic acid drugs that act simultaneously on multiple targets.

[0054] The simulation department uses generative AI to perform molecular dynamics simulations and virtual screenings of designed compound candidates, as well as toxicity and pharmacokinetic predictions using machine learning models. For example, it predicts the binding affinity, stability, and dynamics of compounds to target proteins at the atomic level, rapidly screens hundreds of thousands to millions of compound candidates to evaluate their binding ability and selectivity to targets, predicts various toxicity risks and pharmacokinetic properties such as hepatotoxicity, nephrotoxicity, cardiotoxicity, teratogenicity, and immunotoxicity, and learns from past experimental failures and adverse event occurrences to predict the impact of specific structural motifs on efficacy and safety.

[0055] The analysis unit aggregates the output results from the simulation unit and automatically selects only promising compounds that meet predetermined thresholds based on multidimensional evaluation indicators such as efficacy score, toxicity score, pharmacokinetic score, and ease of manufacture score. The optimization proposal unit repeatedly performs structural optimization and proposes novel modifications for the selected compounds using a generative AI, automatically executing a variety of optimizations such as adjusting the balance of hydrophilicity and lipophilicity, improving target selectivity, improving metabolic stability, reducing synthesis costs, designing patent-avoiding structures, designing multi-target compounds that act on multiple targets simultaneously, designing prodrugs suitable for drug delivery systems, and designing biopharmaceuticals.

[0056] The user interface provides researchers and medical professionals with a screen that allows them to view and manipulate the server's processing status, simulation results, compound design proposals, and the AI's analysis rationale and design intent in real time. It offers a variety of functions, including rotating and scaling 3D models of compound structures, viewing graphs, heatmaps, distribution maps, and time-series changes of simulation results, having the AI ​​explain the design intent and rationale of next-generation compounds proposed by the AI ​​in natural language, comparing with past experimental data and failure cases, and downloading reports and draft papers automatically generated by the AI. Furthermore, users can upload new data, change simulation conditions, request additional analysis and redesign by the AI, and input feedback on the AI's proposals, allowing the generative AI to learn from this feedback and reflect it in future designs and analyses.

[0057] The report generation unit uses a generational AI to automatically generate reports, paper drafts, and application documents for regulatory authorities based on analysis results, design proposals, and simulation results. It automatically creates detailed documents that include the rationale for the design of new compounds, summaries of simulation results, comparisons with past data, and assessments of compliance with regulatory requirements.

[0058] Specific examples of prompts to be fed into the generating AI include: "Integrate all research paper data, clinical trial data, patient record data, and genetic information data related to rare disease A, and extract molecular pathways that could serve as novel therapeutic targets for rare disease A," "Automatically design a novel molecular skeleton with high binding affinity and low toxicity risk based on the three-dimensional structural information of target protein B," "Perform molecular dynamics simulations on compound C and predict its binding stability to target protein B, pharmacokinetic properties, and major toxicity risks," "Calculate efficacy scores, toxicity scores, and pharmacokinetic scores based on the simulation results, and select only compounds that meet predetermined thresholds," "For the selected compounds, propose multiple optimization options considering hydrophilic / lipidophilic balance, target selectivity, ease of synthesis, patent avoidance structures, etc.," and "Automatically generate a detailed report including analysis results and design rationale, and summarize it in natural language that is easy for the user to understand."

[0059] (Application Example 1) The flow of a specific process in Application Example 1 will be explained below. Each part of the system described below is implemented by a data processing device 12 and a smart device 14. The data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0060] (Embodiments for carrying out the present invention)

[0061] Embodiments of the present invention will be described in more specific and detail. The system of the present invention consists of a server and a terminal. The server side includes a data integration unit, a care plan generation unit, a care plan optimization unit, a simulation unit, an evaluation and analysis unit, a feedback learning unit, a report generation unit, and a database management unit. The terminal side includes a user interface unit, a data input / output unit, a notification / alert unit, and a lightweight pre-processing unit. The server can be configured in various forms, such as a cloud server or an on-premise server, and the terminal can be a personal computer, tablet, smartphone, wearable device, or dedicated terminal for caregiving.

[0062] First, during the data collection phase, care staff, medical professionals, and family members input care record data from terminals. This care record data includes, for example, daily food intake (calories consumed at breakfast, lunch, and dinner, nutritional balance, and whether or not assistance was provided with meals), excretion status (number of times urinating and defecating, presence or absence of incontinence, occurrence of constipation or diarrhea), sleep duration (number of times waking up during the night, presence or absence of naps), presence or absence of falls or accidents (time, location, circumstances, and severity of injury), and observations by care staff (facial expressions, speech, gait, presence and severity of dementia symptoms, emotional stability, signs of agitation or anxiety). Vital data includes regularly measured values ​​such as blood pressure, pulse, body temperature, weight, blood oxygen saturation, blood glucose levels, and electrocardiogram data. These can be manually entered or automatically acquired from wearable devices, bed sensors, blood pressure monitors, thermometers, smartwatches, etc. Medication history data includes information such as medication details (drug name, dosage, administration time), whether medication was missed, occurrence of side effects (rash, nausea, drowsiness, dizziness, etc.), medication change history, and drug allergy information. Rehabilitation progress data includes details of walking training and muscle training (number of sessions, duration, load), changes in range of motion, evaluation comments from the rehabilitation therapist, whether rehabilitation was refused or discontinued, and the degree of achievement of rehabilitation goals. Living environment information data includes room temperature and humidity, noise level, lighting conditions, presence of barrier-free facilities, presence of cohabiting family members and their daily routines, presence of pets, size of living space and traffic flow, furniture arrangement, and presence of areas with a high risk of falls. Family request data includes specific wishes and constraints such as "I want to spend as much time at home as possible," "I want to increase outings and hobby activities," "I want thorough medication management," "I want enhanced nighttime monitoring," and "I want consideration given to the times when family members can participate in caregiving." Medical institution collaboration data includes referral letters from the attending physician, the latest diagnosis and treatment plan, test results (blood tests, imaging tests, electrocardiograms, etc.), emergency response instructions, the need for medical care (suctioning, tube feeding, insulin injections, etc.), infection risk and isolation instructions, and discharge summaries. Past care case data includes care plans and their outcomes for other users with similar symptoms and living backgrounds, failure cases and improvement measures, whether or not there were readmissions or relapses, trends in family satisfaction, staff burden, and utilization of community resources.Furthermore, it can integrate with external care case databases, electronic medical record systems of medical institutions, rehabilitation support systems, medication management systems, environmental sensor networks, and data from community-based integrated care centers, enabling the comprehensive collection of all relevant data.

[0063] The data integration unit on the server semantically integrates this diverse data using natural language processing engines, structured data integration modules, image analysis AI, and time-series data analysis AI, enabling a unified understanding of the overall picture for each user. For example, it extracts daily living patterns and abnormal behaviors from care records, analyzes trends in health status and the risk of sudden changes from vital data, evaluates medication adherence, side effect risks, and drug interaction risks from medication history, quantifies the degree of maintenance and improvement of physical function and the tendency to refuse rehabilitation from rehabilitation progress, extracts fall risks, points for optimizing the sleep environment, and stress factors such as noise and lighting from living environment information, and extracts priorities and emergency response items that should be reflected in the care plan from family requests and medical institution collaboration information. Furthermore, from past care case data, it extracts care plans that were effective in similar cases, factors of failure, improvement measures, recurrence prevention measures, and patterns of utilization of local resources, and evaluates their applicability to current users. For example, past data from users with similar stages of dementia progression can be used to extract information on what rehabilitation and environmental adjustments were effective, and what side effects or accidents were likely to occur, and this information can be reflected in the care plan design for the current user.

[0064] The care plan generation unit uses a generation AI based on integrated information to automatically generate multiple care plan candidates for each individual user. For example, meal plans might include "providing high-protein, low-salt meals three times a day," "introducing pureed food to accommodate impaired swallowing function," "adding snacks if food intake is decreasing," "selecting low-carbohydrate menus if there is a high risk of diabetes," and "setting up special meals to accommodate food allergies." Rehabilitation plans might include "conducting walking training for 30 minutes each in the morning and evening," "adding muscle training three times a week," "conducting range of motion exercises daily," "conducting cognitive function training twice a week," and "introducing a motivational program if there is a tendency to refuse rehabilitation." Medication management plans might include "staff confirming medication after meals" and "alerts to prevent missed medications." The system can simultaneously design diverse care plans that combine various elements, such as setting up "automatic notification to the doctor when side effects occur," "automatic contact to family when medication is changed," and "automatic consultation to a pharmacist if there is a high risk of drug interaction," as well as adjusting the living environment with features like "activating fall prevention sensors at night," "automatic adjustment of room lighting," "automatic management of room temperature and humidity," "proposing barrier-free renovations," and "recommending soundproofing measures if noise levels are high," and recommending social activities with features like "recommending family visits twice a week," "supporting participation in community salons," "securing time for hobby activities," "introducing outing support programs," and "setting up online social gatherings." Furthermore, the generative AI learns from past care cases, the latest medical knowledge, guidelines, local resource information, family lifestyles and caregiving burdens, and the user's values ​​and preferences. It then autonomously designs and proposes the optimal care plan based on scientific evidence, without relying on conventional rules of thumb or subjective judgment, by comprehensively considering various factors such as the user's condition and background, family requests, and instructions from medical institutions.

[0065] The Care Plan Optimization Department uses generative AI to perform virtual simulations of the generated care plan candidates, predicting outcomes from multiple perspectives such as physical improvement (maintenance and improvement of walking ability, muscle strengthening, reduction of fall risk, prevention of pressure ulcers, maintenance of swallowing function), mental stability (inhibition of dementia progression, improvement of depressive symptoms, prevention of social isolation, improvement of motivation), implementation compliance rate (whether users and families can continue the plan, burden on staff, implementation costs), and family satisfaction (family sense of security and reduction of burden, level of care participation, quality of communication among family members), and calculating quantified evaluation indicators. For example, detailed evaluation indicators can be obtained for each plan, such as "how much maintenance and improvement of physical function can be expected if this plan is implemented," "how much it contributes to mental stability and improved motivation," "how well users and families can continue the plan," "how much family satisfaction and sense of security will increase," "how much the burden on staff will increase or decrease," and "how much medical and nursing care costs will change." A comprehensive score is calculated by weighting these evaluation indicators, and the care plan with the highest score is automatically selected. For example, if physical improvement is a priority, its weight can be set higher, and if family demands are strong, the weight of family satisfaction can be increased, allowing for flexible optimization depending on the situation.

[0066] Furthermore, whenever new data such as daily vital signs, new accidents, additional requests from family members, new instructions from doctors, new data from environmental sensors, changes in rehabilitation progress, decreased medication adherence, or changes in social activity participation are input, the AI ​​reintegrates all the data and autonomously repeats the process of regenerating care plan candidates, performing virtual simulations, and re-selecting the optimal plan. For example, if a user's weight drops sharply, the system immediately redesigns their diet and rehabilitation content; if a fall occurs, it adjusts the living environment and strengthens the rehabilitation content; if the family requests to increase the frequency of going out, it strengthens recommendations for social activities; and if a doctor indicates a new treatment plan, it immediately reflects this in medication management and medical care content. This enables continuous optimization of the care plan in response to changes in the user's condition and environment.

[0067] The Evaluation and Analysis Department collects and analyzes outcome data after the implementation of care plans (changes in physical function, changes in mental state, accident rate, family satisfaction, trends in medical and nursing care costs, readmission rate, social activity participation rate, etc.), continuously evaluates the AI's prediction accuracy and the effectiveness of care plans, and contributes to improving the overall accuracy of the system. The Feedback Learning Department collects feedback from care staff and families ("This plan was easy to implement," "It reduced the burden on the family," "The user disliked it," "The rehabilitation was ineffective," etc.), and the generative AI learns from this feedback and reflects it in future care plan design and simulations. The Report Generation Department automatically generates the basis for care plan design, simulation results, post-implementation evaluations, explanatory materials for families and staff, reports to government and medical institutions, application documents, etc., to provide information in an easy-to-understand manner to various stakeholders such as users, families, staff, medical institutions, and government agencies. The Database Management Department manages all data safely and efficiently, and is responsible for privacy protection, security measures, access rights management, backup, and data linkage.

[0068] On the terminal side, a user interface is provided that allows care staff, medical professionals, and family members to view and manipulate the processing status, simulation results, care plan proposals, and the AI's analysis basis and design intent in real time. For example, they can edit and confirm each element of the care plan, view graphs, heatmaps, and time-series changes of the simulation results, have the AI ​​explain the design intent and basis of the proposed care plan in natural language, compare it with past implementations and failures, and download reports and application documents automatically generated by the AI. The notification and alert unit immediately notifies staff, family members, and medical institutions when events such as missed medication, increased risk of falls, abnormal vital signs, failure to perform rehabilitation, urgent requests from family members, or new instructions from doctors occur, prompting a quick response. The data input and output unit handles daily record keeping, input of new data, output of reports and documents, and data linkage with external systems. A lightweight pre-processing unit reduces the load on the server side by performing format conversion, simple checks, and privacy protection processing on the terminal side.

[0069] As described above, the system of the present invention implements a data integration unit, a care plan generation unit, a care plan optimization unit, a simulation unit, an evaluation and analysis unit, a feedback learning unit, a report generation unit, and a database management unit on the server side. By making essential and core use of generative AI in each unit, it is possible to perform large-scale data analysis, automatic generation of individually optimized care plans, virtual simulation, result analysis, continuous optimization, user feedback learning, automatic report generation, privacy protection, and security management in a seamless, autonomous, and high-speed manner, which were not possible in conventional nursing care settings. For example, care plan design, which conventionally relied on rules of thumb and subjective judgment, can be replaced with highly accurate automation based on scientific evidence and vast amounts of case data, enabling the continuous provision of optimal care tailored to the condition and needs of each individual user without increasing the burden on on-site staff and families. Furthermore, since a vast amount of past care cases and the latest medical knowledge are reflected in real time, advanced care that surpasses conventional limits can be realized, dramatically improving the quality and efficiency of the entire nursing care setting.

[0070] (System Configuration)

[0071] The system according to this embodiment comprises a data integration unit, a care plan generation unit, a care plan optimization unit, a simulation unit, an evaluation and analysis unit, a feedback learning unit, a report generation unit, a database management unit, a user interface unit, a data input / output unit, a notification / alert unit, and a lightweight preprocessing unit. The data integration unit collects diverse information such as care record data, vital data, medication history data, rehabilitation progress data, living environment information data, family request data, medical institution collaboration data, past care case data, external care case databases, medical institution electronic medical record systems, rehabilitation support systems, medication management systems, environmental sensor networks, and community comprehensive support center data, and semantically integrates it using a natural language processing engine, structured data integration module, image analysis AI, time series data analysis AI, etc. For example, it extracts daily living patterns and abnormal behaviors from care records, and vital data We analyze trends in changes in health status and the risk of sudden changes from data, evaluate medication adherence, side effect risk, and drug interaction risk from medication history, quantify the degree of maintenance and improvement of physical function and the tendency to refuse rehabilitation from rehabilitation progress, extract fall risk, points for optimizing the sleep environment, and stress factors such as noise and lighting from living environment information, extract priorities and emergency response items that should be reflected in the care plan from family requests and medical institution collaboration information, and extract effective care plans, failure factors, improvement measures, recurrence prevention measures, and patterns of utilization of local resources from past care case data to evaluate their applicability to current users.The care plan generation unit uses a generation AI based on integrated information to automatically generate multiple care plan candidates for each individual user. For example, the meal plan might include providing high-protein, low-salt meals three times a day, introducing pureed food to accommodate swallowing difficulties, adding snacks if food intake is decreasing, selecting low-carbohydrate menus if there is a high risk of diabetes, and setting up special meals to accommodate food allergies. The rehabilitation plan might include 30 minutes of walking training in the morning and evening, adding muscle training three times a week, performing range of motion exercises daily, conducting cognitive function training twice a week, and introducing motivational programs if there is a tendency to refuse rehabilitation. Medication management and other aspects are also included. To ensure medication adherence is maintained, staff will check after meals; alerts will be set to prevent missed doses; doctors will be automatically notified if side effects occur; families will be automatically contacted if medication changes are made; and pharmacists will be automatically consulted if there is a high risk of drug interactions. To adjust the living environment, fall prevention sensors will be activated at night; room lighting will be automatically adjusted; room temperature and humidity will be automatically managed; barrier-free renovations will be proposed; and soundproofing measures will be recommended if noise levels are high. To encourage social activities, family visits twice a week will be recommended; participation in community salons will be supported; time for hobbies will be secured; outing support programs will be introduced; and online social gatherings will be set up. A diverse care plan combining all these elements will be designed simultaneously. The Care Plan Optimization Department uses a generative AI to perform virtual simulations of the generated care plan candidates, predicting outcomes from multiple perspectives, including physical improvement (maintenance and improvement of walking ability, muscle strengthening, fall risk reduction, pressure ulcer prevention, maintenance of swallowing function), mental stability (inhibition of dementia progression, improvement of depressive symptoms, prevention of social isolation, increased motivation), implementation compliance rate (whether users and families can continue the plan, staff burden, implementation costs), and family satisfaction (family sense of security and reduction of burden, level of care participation, quality of family communication). It calculates quantified evaluation indicators, weights these indicators to calculate an overall score, and automatically selects the care plan with the highest score. The Simulation Department generates a virtual implementation scenario for each care plan, predicting, for example, changes in the fall accident rate, trends in rehabilitation effectiveness, improvement in medication adherence, time-series changes in family satisfaction, trends in medical and nursing care costs, and the effect of reducing readmission rates.The Evaluation and Analysis Department collects and analyzes outcome data after care plan implementation, continuously evaluating the AI's prediction accuracy and the effectiveness of care plans, contributing to the overall improvement of the system's accuracy. The Feedback Learning Department collects feedback from care staff and families, and the generative AI learns from this feedback to reflect it in future care plan design and simulations. The Report Generation Department automatically generates care plan design rationale, simulation results, post-implementation evaluations, explanatory materials for families and staff, reports to government and medical institutions, and application documents, ensuring that information is provided in an easily understandable manner to various stakeholders, including users, families, staff, medical institutions, and government agencies. The Database Management Department manages all data safely and efficiently, handling privacy protection, security measures, access rights management, backups, and data linkage. The user interface section provides a screen that allows care staff, medical professionals, and family members to view and manipulate the processing status, simulation results, care plan proposals, and AI analysis rationale and design intent in real time. They can edit and confirm each element of the care plan, view graphs, heatmaps, and time-series changes of simulation results, have the AI ​​explain the design intent and rationale of the proposed care plan in natural language, compare it with past implementations and failures, and download reports and application documents automatically generated by the AI. The data input / output section handles daily record keeping, input of new data, output of reports and documents, and data linkage with external systems. The notification / alert section immediately notifies staff, family members, and medical institutions when events such as missed medication, increased risk of falls, abnormal vital signs, failure to perform rehabilitation, urgent requests from family members, or new instructions from doctors occur, prompting a quick response. The lightweight pre-processing section reduces the load on the server by performing format conversion, simple checks, and privacy protection processing on the terminal side.Specific examples of prompts to be fed into the generating AI necessary for implementing the present invention include: "Integrate user X's care records, vital data, medication history, rehabilitation progress, living environment information, family requests, medical institution collaboration information, and past care case data to automatically generate five optimal care plan candidates," "For each care plan candidate, predict the degree of physical improvement, mental stability, implementation compliance rate, and family satisfaction through virtual simulation and calculate an overall score," "If there are new vital sign abnormalities, accidents, additions to family requests, or changes in physician orders, reintegrate all data, autonomously regenerate and simulate care plan candidates, and re-select the optimal plan," and "Automatically generate a detailed report including the design basis of the care plan, simulation results, and post-implementation evaluation, and summarize it in an easy-to-understand manner for family, staff, and medical institutions."

[0072] (Implementation Steps)

[0073] Step 1: Collection and input of multifaceted care data In this step, care staff, medical professionals, and family members use terminals to collect a variety of information on a server, including care record data (food intake, excretion status, sleep duration, presence or absence of falls or accidents, observation records), vital data (blood pressure, pulse, body temperature, weight, blood oxygen saturation, blood glucose level, electrocardiogram), medication history data (drug name, dosage, administration time, missed doses, occurrence of side effects), rehabilitation progress data (training content, number of sessions, load, evaluation comments, degree of goal achievement), living environment information data (temperature / humidity, noise, lighting, barrier-free facilities, cohabiting family members, pets, furniture arrangement), family request data (desire to stay at home, desire to go out, strengthened medication management, strengthened nighttime monitoring, time of care participation), medical institution collaboration data (medical information provision letter, diagnosis, treatment plan, test results, emergency instructions, medical care), and past care case data (care plans for similar symptoms, results, failure cases, readmission status, family satisfaction, staff burden, utilization of local resources). Furthermore, the system will integrate with external care case databases, electronic medical records from healthcare institutions, rehabilitation support systems, medication management systems, environmental sensor networks, and community-based integrated care centers to conduct comprehensive data collection.

[0074] Step 2: Data Integration and Building a Comprehensive User Profile In this step, the data integration unit on the server semantically integrates the diverse data collected using a natural language processing engine, structured data integration module, image analysis AI, time series data analysis AI, etc., enabling a unified understanding of the overall picture for each user. For example, it extracts lifestyle patterns and abnormal behaviors from care records, analyzes trends in health status and the risk of sudden changes from vital data, evaluates medication adherence, side effect risk, and drug interaction risk from medication history, quantifies the degree of maintenance and improvement of physical function and the tendency to refuse rehabilitation from rehabilitation progress, extracts fall risk and points for optimizing the sleep environment from living environment information, extracts priorities and emergency response items to be reflected in the care plan from family requests and medical institution collaboration information, and extracts effective care plans, failure factors, improvement measures, and patterns of utilization of local resources from past care case data, and evaluates their applicability to current users.

[0075] Step 3: Automatic generation of care plan candidates using AI. In this step, based on the overall user profile constructed by the data integration unit, the care plan generation unit uses a generation AI to automatically generate multiple care plan candidates for each individual user. For example, it simultaneously designs care plans that combine diverse elements such as: dietary content (high-protein, low-salt diet, diet for impaired swallowing function, additional snacks, low-carbohydrate menus, allergy-friendly diet); rehabilitation content (walking training, muscle training, range of motion training, cognitive function training, motivational programs); medication management (medication confirmation, alert settings, automatic side effect notifications, automatic notification when medication is changed, automatic pharmacist consultation); living environment adjustments (fall prevention sensors, automatic lighting adjustment, automatic room temperature and humidity management, barrier-free renovations, soundproofing measures); and social activity recommendations (family visits, participation in community salons, hobby activities, outing support, online social gatherings). A specific example of a prompt message to be fed into the generation AI is: "Integrate user X's care records, vital data, medication history, rehabilitation progress, living environment information, family requests, medical institution collaboration information, and past care case data, and automatically generate five optimal care plan candidates."

[0076] Step 4: Virtual simulation and calculation of evaluation metrics using generative AI In this step, the care plan optimization unit uses a generative AI to conduct virtual simulations for each care plan candidate, predicting outcomes from multiple perspectives such as physical improvement (maintenance / improvement of walking ability, muscle strengthening, reduction of fall risk, prevention of pressure ulcers, maintenance of swallowing function), mental stability (inhibition of dementia progression, improvement of depressive symptoms, prevention of social isolation, improvement of motivation), implementation compliance rate (feasibility of continuation, burden on staff, cost), and family satisfaction (sense of security, degree of burden reduction, level of care participation, family communication), and calculating quantified evaluation indicators. A specific example of a prompt message to be fed into the generative AI is, "For each care plan candidate, predict the degree of physical improvement, mental stability, implementation compliance rate, and family satisfaction through virtual simulation, and calculate an overall score."

[0077] Step 5: Automatic selection and continuous optimization of the optimal care plan In this step, the evaluation and analysis unit weights the evaluation indicators for each care plan candidate to calculate an overall score and automatically selects the care plan with the highest score. Furthermore, whenever daily vital sign measurements, accidents occur, family requests are added, physician orders are changed, new data is received from environmental sensors, rehabilitation progress changes, decreased medication adherence, or changes in social activity participation are entered, the AI ​​reintegrates all data and autonomously repeats the process of regenerating care plan candidates, performing virtual simulations, and re-selecting the optimal plan. A specific example of a prompt message to be fed into the generating AI is: "If there are new vital sign abnormalities, accidents occur, family requests are added, or physician orders are changed, reintegrate all data, autonomously regenerate and simulate care plan candidates, and re-select the optimal plan."

[0078] Step 6: Outcome evaluation, feedback learning, and report generation In this step, the evaluation and analysis unit collects and analyzes outcome data after the implementation of the care plan (changes in physical function, changes in mental state, accident rate, family satisfaction, trends in medical and nursing care costs, readmission rate, social activity participation rate, etc.) to continuously evaluate the AI's prediction accuracy and the effectiveness of the care plan. The feedback learning unit collects feedback from staff and families, and the generating AI learns from this and reflects it in future care plan design and simulations. The report generation unit automatically generates the basis for the care plan design, simulation results, post-implementation evaluations, explanatory materials for families and staff, reports to government and medical institutions, and application documents, providing information to stakeholders in an easy-to-understand manner. A specific example of a prompt sentence to be fed into the generating AI is, "Automatically generate a detailed report including the basis for the care plan design, simulation results, and post-implementation evaluations, and summarize it in an easy-to-understand manner for families, staff, and medical institutions."

[0079] (Specific use cases)

[0080] For example, consider a use case for home care support for elderly people requiring long-term care. In this case, care staff or family members use a terminal to input care record data such as the user's daily food intake, excretion status, sleep duration, presence or absence of falls or accidents, presence and severity of dementia symptoms, emotional stability, walking ability, facial expressions, and speech content. Furthermore, vital data such as blood pressure, pulse, body temperature, weight, blood oxygen saturation, blood glucose level, and electrocardiogram are automatically acquired from wearable devices or bed sensors, and medication history data such as medication content, administration time, missed doses, occurrence of side effects, medication change history, and drug allergy information are also input. Rehabilitation progress data includes the content, number of repetitions, load, evaluation comments, whether rehabilitation was refused or discontinued, and the degree of achievement of rehabilitation goals. Living environment information data includes room temperature and humidity, noise level, lighting conditions, presence or absence of barrier-free facilities, presence or absence of cohabiting family members, presence or absence of pets, furniture arrangement, and presence or absence of areas with a high risk of falls. Family request data includes specific wishes such as "I want to spend as much time at home as possible," "I want to increase outings and hobby activities," "I want thorough medication management," "I want enhanced nighttime monitoring," and "I want consideration given to times when family members can participate in caregiving." Medical institution collaboration data includes medical information from the attending physician, the latest diagnosis and treatment plan, test results, emergency response instructions, necessity of medical care, infection risk and isolation instructions, and discharge summary. Furthermore, past care case data is collected, including care plans and their outcomes for other users with similar symptoms and living backgrounds, failures and improvement measures, whether there were readmissions or relapses, trends in family satisfaction, staff burden, and utilization of local resources.

[0081] The Data Integration Department semantically integrates all of this data to gain a unified understanding of the user's overall situation. The Care Plan Generation Department uses generative AI to automatically generate multiple care plan candidates that combine meals, rehabilitation, medication management, living environment adjustments, and recommendations for social activities. Examples include "providing high-protein, low-salt meals three times a day," "introducing pureed food to accommodate impaired swallowing function," "conducting 30 minutes of walking training in the morning and evening," "staff checking medication after meals," "activating fall prevention sensors at night," "recommending family visits twice a week," and "securing time for hobbies." The Care Plan Optimization Department uses generative AI to perform virtual simulations for each care plan candidate, predicting outcomes from various perspectives such as maintaining and improving walking ability, strengthening muscles, reducing fall risk, slowing dementia progression, improving depressive symptoms, preventing social isolation, implementation compliance rate, family satisfaction, staff burden, and trends in medical and nursing care costs, and calculating quantified evaluation indicators. The Evaluation and Analysis Department weights these evaluation indicators to calculate an overall score and automatically selects the care plan with the highest score. Furthermore, whenever daily vital signs, accidents, additional family requests, changes in physician orders, new data from environmental sensors, changes in rehabilitation progress, decreased medication adherence, or changes in social activity participation are entered, the AI ​​reintegrates all data and autonomously repeats the process of regenerating care plan candidates, conducting virtual simulations, and re-selecting the optimal plan. The feedback learning unit collects feedback from staff and family members, and the generative AI learns from this and incorporates it into future care plan design and simulations. The report generation unit automatically generates the rationale for care plan design, simulation results, post-implementation evaluations, explanatory materials for family and staff, reports to government and medical institutions, and application documents, providing information to stakeholders in an easily understandable manner.

[0082] Specific examples of prompts to be fed into the generating AI include: "Integrate user A's care records, vital data, medication history, rehabilitation progress, living environment information, family requests, medical institution collaboration information, and past care case data to automatically generate five optimal care plan candidates," "For each care plan candidate, predict maintenance / improvement of walking ability, reduction of fall risk, suppression of dementia progression, implementation compliance rate, and family satisfaction through virtual simulation, and calculate an overall score," "If there are new vital sign abnormalities, accidents, additions of family requests, or changes in physician orders, reintegrate all data, autonomously regenerate and simulate care plan candidates, and re-select the optimal plan," and "Automatically generate a detailed report including the rationale for the care plan design, simulation results, and post-implementation evaluation, and summarize it in an easy-to-understand manner for family, staff, and medical institutions."

[0083] The specific processing unit 290 transmits the result of the specific processing obtained in the embodiment 1 or application example 1 described above to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0084] Data generation model 58 is a form of so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> Examples of generative AI include text generation AI, image generation AI, and multimodal generation AI. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats from audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVMs), k-means clustering, convolutional neural networks (CNNs), recurrent neural networks (RNNs), generative adversarial networks (GANs), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each of the above parts is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example.Furthermore, processing performed by AI, including generative AI, may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by AI, including generative AI.

[0085] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart device 14 or an external device, and the smart device 14 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0086] For example, the data collection unit is implemented by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit acquires step count data using the camera 42 and communication I / F 44 of the smart device 14 and processes it with the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the data from the data collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 and generates a cooking menu using generation AI. For example, the serving unit is implemented by the output device 40 of the smart device 14 or the specific processing unit 290 of the data processing device 12 and provides the generated cooking menu to the user. The correspondence between each unit and the devices and control units is not limited to the examples described above and can be modified in various ways.

[0087] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0088] [Second Embodiment]

[0089] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0090] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0091] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0092] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and communication interface 44 are also connected to the bus 52.

[0093] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0094] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0095] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0096] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0097] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0098] Storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0099] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48. The smart glasses 214 also have a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the identification processing unit 290 using these models.

[0100] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is realized by the data processing device 12 and the smart glasses 214. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0101] (Example 1)

[0102] Since the process flow is the same as in Example 1 described in the First Embodiment above, the explanation will be omitted.

[0103] (Application Example 1) Since the process flow is the same as in Application Example 1 described in the first embodiment above, the explanation will be omitted.

[0104] The specific processing unit 290 transmits the result of the specific processing obtained in the embodiment 1 or application example 1 described above to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0105] Furthermore, the processing by the data processing system 10 described above is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but it may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart glasses 214 or an external device, and the smart glasses 214 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0106] For example, the data collection unit is implemented by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit acquires step count data using the camera 42 and communication I / F 44 of the smart glasses 214 and processes it using the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the data from the data collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 and generates a cooking menu using generation AI. For example, the serving unit is implemented by the speaker 240 of the smart glasses 214 or the specific processing unit 290 of the data processing device 12 and provides the generated cooking menu to the user. The correspondence between each unit and the device or control unit is not limited to the examples described above and can be modified in various ways.

[0107] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0108] [Third Embodiment]

[0109] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0110] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0111] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0112] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, display 343, and communication interface 44 are also connected to the bus 52.

[0113] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0114] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0115] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0116] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0117] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0118] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0119] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0120] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the headset terminal 314. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0121] (Example 1) Since the process flow is the same as in Example 1 described in the First Embodiment above, the explanation will be omitted.

[0122] (Application Example 1) Since the process flow is the same as in Application Example 1 described in the first embodiment above, the explanation will be omitted.

[0123] The specific processing unit 290 transmits the result of the specific processing obtained in the embodiment 1 or application example 1 described above to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0124] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the headset terminal 314, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the headset terminal 314. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the headset terminal 314 or an external device, and the headset terminal 314 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0125] For example, the data collection unit is implemented by the control unit 46A of the headset terminal 314 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit acquires step count data using the camera 42 and communication I / F 44 of the headset terminal 314 and processes it using the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the data from the data collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 and generates a cooking menu using generation AI. For example, the serving unit is implemented by the speaker 240 and display 343 of the headset terminal 314 or the specific processing unit 290 of the data processing device 12 and provides the generated cooking menu to the user. The correspondence between each unit and the device or control unit is not limited to the examples described above and can be modified in various ways.

[0126] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0127] [Fourth Embodiment]

[0128] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0129] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[0130] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0131] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, controlled object 443, and communication interface 44 are also connected to the bus 52.

[0132] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0133] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0134] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0135] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0136] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0137] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0138] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0139] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0140] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the robot 414. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0141] (Example 1) Since the process flow is the same as in Example 1 described in the First Embodiment above, the explanation will be omitted.

[0142] (Application Example 1) Since the process flow is the same as in Application Example 1 described in the first embodiment above, the explanation will be omitted.

[0143] The specific processing unit 290 transmits the result of the specific processing obtained in the embodiment 1 or application example 1 described above to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0144] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the robot 414 or external devices, and the robot 414 acquires or collects information necessary for processing from the data processing device 12 or external devices.

[0145] For example, the data collection unit is implemented by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit acquires step count data using the camera 42 and communication I / F 44 of the robot 414 and processes it using the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the data from the data collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 and generates a cooking menu using a generation AI. For example, the serving unit is implemented by the speaker 240 and controlled object 443 of the robot 414 and the specific processing unit 290 of the data processing device 12 and provides the generated cooking menu to the user. The correspondence between each unit and the device or control unit is not limited to the examples described above and can be modified in various ways.

[0146] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0147] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0148] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0149] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0150] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0151] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0152] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0153] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0154] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0155] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0156] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0157] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[0158] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0159] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using this memory.

[0160] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0161] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0162] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0163] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0164] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0165] Furthermore, the following additional information is disclosed regarding the above explanation.

[0166] (Claim 1)

[0167] The system includes a data integration unit that receives diverse information such as care records, vital data, medication history data, rehabilitation progress data, living environment information data, family requests data, medical institution collaboration data, and past care case data, and integrates and analyzes this information; a care plan generation unit that uses generative AI to automatically generate multiple care plan candidates for each individual user based on the integrated information; and a virtual simulation of the generated care plan candidates using generative AI to predict outcomes from multiple perspectives such as physical improvement, mental stability, implementation compliance rate, and family satisfaction, and automatically selects the optimal care plan, further enabling new data input and utilization. A system comprising a care plan optimization unit that autonomously and repeatedly re-optimizes the care plan in response to changes in the user's condition, wherein the data integration unit centrally integrates a wide variety of data to enable a comprehensive understanding of the user's overall situation, the care plan generation unit generates diverse care plans combining multiple elements such as meal content, rehabilitation content, medication management, living environment adjustments, and recommendations for social activities, and the care plan optimization unit autonomously regenerates and re-selects the care plan by reflecting new information such as daily vital sign measurements, accident occurrences, family requests, and doctor's orders, and this series of processes is automatically and continuously performed by a generation AI. (Claim 2)

[0168] The system according to claim 1, characterized in that the care plan optimization unit learns from past care case data and the latest medical knowledge using generative AI, and comprehensively considers various factors such as the user's condition and living background, family requests, and instructions from medical institutions to autonomously design and propose an optimal care plan based on scientific evidence without relying on conventional rules of thumb or subjective judgment, and further continuously repeats the optimization in response to daily changes in circumstances. (Claim 3)

[0169] The system according to claim 1, characterized in that the care plan generation unit and the care plan optimization unit use a generation AI to quantify multiple evaluation indicators such as the degree of physical improvement, mental stability, implementation compliance rate, and family satisfaction for each care plan candidate through virtual simulation, weight these evaluation indicators to calculate an overall score, automatically select the care plan with the highest score, and repeat the same process in response to new data input or changes in the user's condition. [Explanation of symbols]

[0170] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots

Claims

1. The system comprises a data integration unit that receives diverse information such as scientific paper data, clinical trial data, patient record data, and genetic information data, and integrates and analyzes them; a target extraction and compound design unit that extracts therapeutic targets and disease-related genes from the integrated knowledge base and designs novel compound candidates based on them; and a simulation and analysis unit that uses generative AI to perform simulations of biological experiments on the designed compound candidates, evaluates the efficacy and toxicity of each compound, selects promising compounds that meet predetermined thresholds, and autonomously proposes the design of next-generation compounds based on the selected compounds. The system is characterized by the fact that the data integration unit semantically integrates data of different formats and contents using natural language processing and database integration technology, the target extraction and compound design unit designs chemical structures using molecular design algorithms and structure generation AI, and the simulation and analysis unit predicts biological reactions using generative AI including molecular dynamics simulation, virtual screening technology, and machine learning models, and by autonomously and continuously repeating this series of processes, the system achieves a reduction in the time and improvement in the success rate of the entire drug discovery process.

2. The system according to claim 1, wherein the generative AI autonomously and continuously performs tasks ranging from analysis of integrated data to discovery of novel therapeutic targets, design of compound candidates, virtual screening, efficacy and toxicity evaluation through simulation, and proposal of next-generation compounds, even in disease areas where research has been difficult under conventional economic constraints, such as rare diseases and intractable diseases, thereby enabling efficient new drug development for diverse medical needs that have been difficult to address with conventional drug discovery approaches.

3. The system according to claim 1, wherein the simulation and analysis unit autonomously performs simulations of biological experiments 24 hours a day, 365 days a year using a generative AI, automatically selects compounds that are worth proceeding to actual biological experiments based on the evaluation results of the efficacy and toxicity of each compound candidate, and further repeatedly proposes structural optimization and novel modifications of the selected compounds, thereby accelerating the drug discovery process and creating highly novel drug candidates.

Images