System

Abstract

A system is provided.SOLUTION: A system comprising: means for obtaining a current risk map; means for obtaining historical weather data; means for pre-processing the current risk map and the historical weather data as data sources; means for analyzing using generative artificial intelligence using the pre-processed data to predict disaster risk after multiple years; and means for visualizing the predicted disaster risk.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] In recent years, the frequent occurrence of natural disasters due to climate change has become a social issue. In particular, floods and landslides caused by abnormal weather and heavy rains are occurring frequently, and effective prediction and countermeasures are required. However, the current disaster prevention systems only perform risk assessments based on the most recent data, and the accuracy of predicting disaster risks several years into the future is an issue. In addition, there is a lack of forecasting methods that effectively utilize past weather data, which makes it difficult for residents and local governments to develop adequate disaster prevention plans. [Means for solving the problem]

[0005] The present invention solves the above-mentioned problems by providing a system that uses generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future using a current risk map and past weather data. Specifically, the present invention includes the following means.

[0006] A means of obtaining the current risk map;

[0007] a means for obtaining historical weather data;

[0008] means for preprocessing the current risk map and the historical weather data as data sources;

[0009] A means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now;

[0010] A means for visualizing the predicted disaster risk;

[0011] Includes:

[0012] This makes it possible to effectively predict future disaster risks based on highly accurate data that complements outliers and missing values, and to provide information that will be useful in planning disaster prevention measures. Furthermore, by utilizing rainfall data and meteorological chart data as weather data, more precise and reliable forecast results can be obtained.

[0013] A "risk map" is a map showing the risk of natural disasters occurring, visually representing the degree of danger in a particular area.

[0014] "Weather data" is a general term for time-series data that represents weather information such as rainfall, temperature, and wind speed observed in the past.

[0015] "Generative artificial intelligence" is a machine learning technique that generates new data or makes predictions based on given data, and in particular, recurrent neural networks (RNN) and long short-term memory (LSTM) networks fall into this category.

[0016] "Preprocessing" is the process of correcting outliers in data and filling in missing values ​​to prepare the data in a format suitable for analysis and model training.

[0017] "Analysis" refers to the process of using generative artificial intelligence to extract patterns and trends from past data and predict future data.

[0018] "Visualization" is a technique for displaying analysis and prediction results in visual formats such as graphs and maps, providing information in a way that is easy for users to understand. [Brief explanation of the drawings]

[0019] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11]FIG. 3 is a sequence diagram illustrating a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0020] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0021] First, the terms used in the following description will be explained.

[0022] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0023] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0024] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0025] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0026] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0027] [First embodiment]

[0028] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0029] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0030] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "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).

[0031] The smart device 14 includes a computer 36, a reception device 38, an output device 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 device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0032] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A 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.

[0033] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio 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 audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0034] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0035] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0036] 2, in the data processing device 12, a specific process 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" according to the technology of the present 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 process 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.

[0037] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0038] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0039] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0040] The present invention relates to a system that uses a generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future, based on a current risk map and past weather data. Below, we will explain in natural language the mode for carrying out the present invention and the program processing. We will also provide specific examples.

[0041] Data retrieval by the server

[0042] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including, for example, rainfall data and weather chart data from the past 30 years. The server retrieves this data and stores it in a database.

[0043] Data preprocessing by the server

[0044] The server preprocesses the stored risk map data and weather data by imputing outliers and missing values, integrating the data, and converting it into a time series format. It also normalizes the data to make it suitable for machine learning models.

[0045] Server-based analysis using generative AI

[0046] The server uses the pre-processed data and uses generative artificial intelligence (e.g., LSTM models) to learn from past data, allowing it to extract temporal patterns and trends and predict disaster risk for several years into the future.

[0047] Server-based future prediction

[0048] The server uses the trained model to predict future disaster risks, such as flood risk over the next five years, and the results are stored in a database.

[0049] Visualization of prediction results on the device

[0050] The device receives the prediction results from the server and provides them to the user in a visual format, such as graphs and maps, in a way that is easy for the user to understand.

[0051] Specific examples

[0052] For example, suppose a user wants to know the risk of flooding. The server retrieves current hazard map data from the government's hazard map API, and also retrieves past rainfall data and weather chart data from the weather data API. These data are preprocessed on the server, and outliers and missing values ​​are filled in.

[0053] Next, using the preprocessed data, the server trains the data using a generative artificial intelligence (AI) model with an LSTM model. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database.

[0054] When a user accesses the system via a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps. For example, a graph showing flood risk forecasts for the next five years based on data from the past 30 years can be displayed in an easy-to-understand manner.

[0055] In this way, the system of the present invention effectively utilizes current risk maps and past weather data to perform highly accurate disaster predictions using generative artificial intelligence, and visualizes the results and provides them to users, allowing them to take appropriate measures against future disaster risks.

[0056] The processing flow will be explained below.

[0057] Step 1:

[0058] The server connects to external data sources to retrieve the latest risk maps and historical weather data, specifically accessing APIs from government and meteorological agencies to retrieve rainfall data and weather chart data for the past 30 years.

[0059] Step 2:

[0060] The server stores the acquired risk map data and weather data in a database, along with timestamps and related information corresponding to the acquired data.

[0061] Step 3:

[0062] The server preprocesses the data stored in the database. Specifically, it detects and complements outliers, estimates missing values ​​based on previous and subsequent data, and fills in missing values. It also organizes rainfall data and meteorological map data and converts them into a time series format.

[0063] Step 4:

[0064] The server normalizes the preprocessed data, calculating the mean and standard deviation of the data and standardizing each data point to prepare it for efficient learning by the generative AI.

[0065] Step 5:

[0066] The server uses the normalized data to train generative AI models such as LSTM models, and builds models to predict future disaster risks based on past time-series data.

[0067] Step 6:

[0068] The server uses the trained model to predict disaster risk over the next five years, generating data that quantifies monthly flood risk, for example, and storing the prediction results in a database.

[0069] Step 7:

[0070] A user accesses the system from a terminal and sends a request for disaster prediction results to the server through an application or web browser on the terminal.

[0071] Step 8:

[0072] The device receives the prediction results from the server, and the server sends the latest disaster prediction results stored in the database to the device.

[0073] Step 9:

[0074] The device visualizes the obtained prediction results, specifically using a visualization library such as Matplotlib to display the predicted disaster risk as graphs and maps.

[0075] Step 10:

[0076] Users can view the visualized forecast results on their devices, which allows them to understand future disaster risks and take appropriate measures.

[0077] Example 1

[0078] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0079] In recent years, the frequency and scale of natural disasters have increased, making it important to accurately predict future disaster risks. However, conventional systems are unable to effectively utilize past data, making it difficult to accurately predict disaster risks several years into the future. The present invention aims to solve this problem by providing a system that utilizes current risk maps and past weather data obtained from external data sources to achieve highly accurate disaster risk predictions.

[0080] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0081] In this invention, the server includes means for periodically acquiring a current risk map from an external data source, means for periodically acquiring past weather data from an external data source, means for preprocessing the current risk map and the past weather data by complementing, integrating, and normalizing outliers and missing values, means for training a generative artificial intelligence model using the preprocessed data to predict disaster risk several years from now, and means for visualizing the predicted disaster risk in the form of a graph or map. This makes it possible to effectively utilize past data to accurately predict future disaster risk and provide it to users visually.

[0082] A "risk map" is a map that geographically represents risks such as natural disasters and accidents.

[0083] "Weather data" refers to data that records past weather conditions such as rainfall, temperature, and wind speed on an hourly basis.

[0084] An "outlier" is a value that is significantly different from the rest of the data, and is a value that deviates significantly from the normal distribution.

[0085] "Missing values" refer to portions of a dataset where no values ​​are recorded.

[0086] "Preprocessing" refers to the process of preparing acquired data in a format suitable for analysis and model learning, and includes correcting outliers, filling in missing values, and integrating and normalizing data.

[0087] A "generative artificial intelligence model" is an algorithm that learns from input data and generates and predicts future data and results; specifically, it refers to neural network models such as LSTM (long short-term memory).

[0088] "Visualization" refers to displaying data and prediction results in a visual format such as a graph or map so that people can understand them intuitively.

[0089] A "data source" is an external resource or system from which data is obtained, such as an API or database.

[0090] "Integration" is the process of combining different datasets into a single dataset.

[0091] "Regularly" means to repeat at regular intervals.

[0092] "Normalization" is a process of converting each data point to fall within a certain range in order to unify the scale of the data.

[0093] MODE FOR CARRYING OUT THE INVENTION

[0094] The present invention provides a system that uses a generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future, based on a current risk map and past weather data. The following describes specific embodiments of the present invention.

[0095] Data retrieval by the server

[0096] The server periodically accesses external data sources to obtain the latest risk maps and historical weather data. For example, the server sends requests to government hazard map APIs and weather data APIs, analyzes the data received in response, and stores it in a database. This data includes rainfall data and weather map data from the past 30 years.

[0097] Data preprocessing by the server

[0098] The server preprocesses the acquired data. This includes imputing outliers and missing values, integrating risk map data with weather data, and normalizing the data. Specifically, anomalies are detected using an anomaly detection algorithm and imputed with appropriate values. Missing data is imputed with the mean or median value. The integrated dataset is then preprocessed using techniques such as z-score normalization.

[0099] Server-based analysis using generative AI

[0100] The server uses the preprocessed data to train a generative artificial intelligence (e.g., an LSTM model), which can learn patterns and trends from past data and predict future disaster risks. The server builds the LSTM model and splits the dataset into training data and validation data for training. After this process is complete, the server saves the trained model.

[0101] For example, the LSTM model predicts flood risk for the next five years based on rainfall data from the past 30 years. The server stores the prediction results in a database for future analysis and visualization.

[0102] Visualization of prediction results on the device

[0103] The terminal retrieves the forecast results from the server and provides them visually to the user. This includes displaying the forecast data in the form of graphs and maps. When a user accesses the system, the terminal sends a request to the server to retrieve the forecast results. Heat maps and graphs are then generated based on the retrieved data and displayed on the user interface. This allows the user to intuitively understand the flood risk forecast for the next five years based on data from the past 30 years.

[0104] Specific examples

[0105] For example, suppose a user wants to know the flood risk for the next five years. The server first obtains the latest risk map data from the government's hazard map API, and simultaneously collects rainfall data and weather chart data from the past 30 years from the weather data API. Next, the server preprocesses this data and uses appropriate algorithms to fill in outliers and missing values. The server then trains this data using an LSTM model to predict the flood risk for the next five years. After saving this forecast data in a database, when the user accesses the system using a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps.

[0106] Prompt Sentence Examples

[0107] "Predict flood risk over the next five years. Use rainfall data from the past 30 years and current risk map data. The generative AI used is an LSTM model."

[0108] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0109] Step 1: Data Acquisition

[0110] The server periodically retrieves current risk maps and historical weather data from external data sources (government hazard map APIs and weather data APIs). Specifically, the server sends requests to the API, analyzes the JSON or CSV format data received as a response, and stores it in a database. The input is the response data from the API, and the output is data saved in the database.

[0111] Step 2: Data Preprocessing

[0112] The server preprocesses the risk map data and weather data obtained from the database. First, it detects outliers and missing values ​​and fills them with appropriate values. Next, it integrates the risk map data and weather data and converts them into time-series data. At this time, it normalizes each data point using methods such as z-score normalization. The input is raw data obtained from the database, and the output is preprocessed, well-formatted data.

[0113] Specific working example:

[0114] An outlier detection algorithm is used to impute outliers.

[0115] Impute missing data with the mean or median.

[0116] Combine multiple datasets and transform them into time series data.

[0117] The data is scaled using a normalization algorithm.

[0118] Step 3: Model training

[0119] The server uses the preprocessed data to build a generative artificial intelligence model (such as an LSTM model) and trains it. It splits the data into training and validation data, and then trains and evaluates the model. The input is the preprocessed data, and the output is the trained LSTM model.

[0120] Specific working example:

[0121] Initialize the LSTM model and set the hyperparameters.

[0122] Split the dataset into training and validation sets.

[0123] A learning algorithm is used to train the model.

[0124] Validation data is used to evaluate the accuracy of the model and adjust the model as needed.

[0125] Step 4: Predict the future

[0126] The server uses a trained LSTM model to predict future disaster risks. The prediction period is multiple years, and risk assessments are performed for each region. The input is the latest risk map data, and the output is predicted future disaster risk data.

[0127] Specific working example:

[0128] Load a trained model.

[0129] Provide the latest risk map data as input to the model.

[0130] Predict future disaster risks using models.

[0131] The prediction results are saved in a database.

[0132] Step 5: Visualize the results

[0133] The terminal obtains the prediction results from the server and provides them visually to the user. Specifically, it generates heat maps and graphs and displays them on the user interface. The input is the prediction data obtained from the server, and the output is a display in the form of a graph or map.

[0134] Specific working example:

[0135] The terminal sends a request to the server to obtain the prediction data.

[0136] Generate graphs and heat maps based on the acquired data.

[0137] The generated graphs and heat maps are displayed in a user interface.

[0138] (Application example 1)

[0139] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0140] In recent years, the frequent occurrence of natural disasters has created a need for damage prediction and prevention. However, conventional disaster risk prediction systems do not fully utilize past data and current risk information, resulting in limitations in the accuracy of predictions and the implementation of disaster prevention measures. Furthermore, many systems only visualize disaster risks and do not adequately communicate them to users or present effective countermeasures. This presents a problem in that it is difficult for users to be aware of risks on a daily basis and take prompt and appropriate measures.

[0141] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0142] In this invention, the server includes means for acquiring a current risk map, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risk several years from now, means for visualizing the predicted disaster risk, means for sending a real-time alert to the user based on the predicted disaster risk, and means for the user to check the disaster risk around their place of residence or workplace on a map. This not only enables highly accurate disaster risk prediction, but also enables the user to be aware of disaster risk on a daily basis and take prompt and appropriate disaster prevention measures.

[0143] A "current risk map" is map information that shows the current risk of natural disasters in a specific area.

[0144] "Past weather data" refers to historical data on weather information such as rainfall and weather charts for a specific period of time.

[0145] "Preprocessing" is the process of correcting abnormal and missing values ​​in the acquired data, and integrating and normalizing it to prepare it into an analyzable format.

[0146] "Generative AI" is an artificial intelligence technology that learns from time-series data and predicts future disaster risks.

[0147] "Disaster risk several years from now" is forecast information that indicates the possibility of a natural disaster occurring in a specific area several years from now.

[0148] "Visualization" refers to the display of numerical data and forecast results in visual formats such as graphs and maps.

[0149] "Real-time alerts" is a function that instantly notifies users of warnings based on predicted disaster risks.

[0150] "Means of checking on a map" refers to a function that allows users to visually check disaster risks around their place of residence or workplace using a geographic information system.

[0151] This invention is a system that uses generative artificial intelligence to predict and visualize disaster risks several years into the future, using a current risk map and past weather data. Specific embodiments for implementing this invention are described below.

[0152] System Program

[0153] 1. Data Acquisition

[0154] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to retrieve the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years.

[0155] 2. Data Preprocessing

[0156] The server complements outliers and missing values ​​in the acquired risk map data and weather data, integrates the data, converts it into a time series format, and normalizes the data to make it suitable for machine learning models.

[0157] 3. Analysis using generative artificial intelligence

[0158] The server uses the pre-processed data and uses generative artificial intelligence (e.g., LSTM models) to learn from past data, allowing it to extract temporal patterns and trends and predict disaster risk for several years into the future.

[0159] 4. Predicting the future

[0160] The server uses the trained model to predict future disaster risks, for example, predicting flood risk over the next five years, and stores the results in a database.

[0161] 5. Visualization of prediction results

[0162] The device receives the prediction results from the server and provides them to the user in a visual format, such as graphs and maps, in a way that is easy for the user to understand.

[0163] 6. Real-time alerts

[0164] Based on the predicted disaster risk, the server sends users real-time alerts to urge them to take measures, and also provides a function that allows users to check the disaster risk around their residence or workplace on a map.

[0165] Hardware and software used

[0166] Hardware:

[0167] Server: A high-performance server handles the analysis, performing data acquisition, preprocessing, and predictive analysis.

[0168] Device: The device used by the user, such as a smartphone or tablet.

[0169] software:

[0170] API client: Uses government hazard map APIs and weather data APIs to obtain data.

[0171] Data preprocessing: Data integration, normalization, and outlier processing using the pandas library.

[0172] Generative AI: Analysis is performed using TENSORFLOW (registered trademark) and LSTM models.

[0173] Visualization: Use matplotlib to visualize the prediction results.

[0174] Specific examples

[0175] For example, if a user requests the system to "predict flood risk for the next five years and display it by region," the server retrieves current hazard map data from the government's hazard map API and collects historical rainfall data and weather chart data from the weather data API. These data are preprocessed and trained using generative artificial intelligence (AI) with an LSTM model. The server then predicts flood risk for the next five years and stores the results in a database.

[0176] When a user accesses the system from their smartphone, the device retrieves the forecast results from the server and displays them in graphs and maps. For example, it provides a graph and map showing flood risk forecasts for the next five years, making it easy for users to understand.

[0177] In this way, the system of the present invention effectively utilizes current risk maps and past weather data to perform highly accurate disaster predictions using generative artificial intelligence, and visualizes the results and provides them to users, allowing them to take appropriate measures against future disaster risks.

[0178] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0179] Step 1: Data Acquisition

[0180] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to retrieve current risk map data and historical weather data. Inputs include API endpoints and parameters, and outputs JSON-formatted data. The server stores this data in a local database.

[0181] Step 2: Data Preprocessing

[0182] The server preprocesses the acquired risk map data and weather data. Specifically, it complements outliers and missing values, integrates the data, and converts it into a time series format. The acquired JSON data is the input, and the formatted data is the output. This processing uses the pandas library. The server detects outliers and complements them with appropriate values. It also merges data obtained from multiple data sources and formats it into time series data.

[0183] Step 3: Generative AI learning

[0184] The server uses the preprocessed data to learn from past data using a generative artificial intelligence (e.g., an LSTM model). The input is the preprocessed time series data, and the output is the trained model. The server uses the TensorFlow library to build the LSTM model and train the model on the time series data.

[0185] Step 4: Predict the future

[0186] The server uses the trained LSTM model to predict future disaster risks. For example, it predicts flood risks over the next five years. The inputs are the trained model and time series data, and the output is future disaster risk prediction data. The server stores this prediction data in a database.

[0187] Step 5: Visualize the prediction results

[0188] The terminal retrieves the prediction results from the server and provides them to the user in a visual format. Specifically, it displays the prediction data in the form of graphs and maps. The input is the prediction result data, and the output is visualized information. The terminal uses the matplotlib library and a geographic information system (GIS) to display the information in a visually easy-to-understand format.

[0189] Step 6: Real-time alerts

[0190] The server sends real-time alerts to users based on predicted disaster risks. The input is predicted disaster risk data, and the output is a notification message. The server sends push notifications to users' devices to provide immediate warnings. Based on the notifications received, users can immediately take appropriate disaster prevention measures.

[0191] This configuration allows users to receive highly accurate disaster prediction information based on current risk maps and past weather data, enabling them to take appropriate action in real time.

[0192] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0193] The present invention relates to a system that uses a current risk map and past weather data to predict disaster risks several years into the future using generative artificial intelligence, and also recognizes the user's emotions to adjust the way information is presented. Below, we will explain in natural language the modes for implementing the present invention and the program's processing. Specific examples will also be included.

[0194] Data retrieval by the server

[0195] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years. The server stores the obtained data in a database.

[0196] Data preprocessing by the server

[0197] The server preprocesses the stored risk map data and weather data. Specifically, it detects and complements outliers, and fills in missing values ​​by estimating them from previous and subsequent data. It also organizes rainfall data and weather map data and converts them into a time series format. It also normalizes the data and formats it in a way that is suitable for machine learning models.

[0198] Server-based analysis using generative AI

[0199] The server uses the preprocessed data to train generative AI models such as LSTM models, which build models to predict future disaster risks based on past time-series data.

[0200] Server-based future prediction

[0201] The server uses the trained model to predict disaster risk over the next five years, generating data that quantifies monthly flood risk, for example, and storing the prediction results in a database.

[0202] Emotion engine that recognizes user emotions

[0203] The server is equipped with an emotion engine that recognizes the user's emotions from their inputs, actions, images, and voice. For example, it analyzes the text and voice entered when the user accesses the system and determines the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0204] Server-adjusted presentation of information

[0205] After the emotion engine recognizes the user's emotion, the server adjusts the way disaster risk information is presented based on the recognized emotion. For example, if the user expresses strong anxiety, the server will present disaster risk information in a more reassuring tone. It will also alleviate the user's anxiety by displaying specific evacuation locations and countermeasures.

[0206] Visualization of prediction results on the device

[0207] The device retrieves the forecast results from the server and provides them to the user in a visual format. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, providing a design that is easy for users to understand.

[0208] Device and user interaction

[0209] Users can understand future disaster risks by viewing the visualized forecast results on their devices. Furthermore, based on the user's input and feedback, the emotion engine continuously monitors the user's emotional state and dynamically adjusts the way information is presented accordingly.

[0210] Specific examples

[0211] For example, suppose a user wants to know the risk of flooding. The server retrieves current hazard map data from the government's hazard map API, and also retrieves past rainfall data and weather chart data from the weather data API. These data are preprocessed on the server, and outliers and missing values ​​are filled in.

[0212] Next, the server uses the preprocessed data to train the data using a generative artificial intelligence (AI) model with an LSTM model. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database.

[0213] When a user accesses the system via a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps. For example, a graph showing flood risk forecasts for the next five years based on data from the past 30 years can be displayed in an easy-to-understand manner.

[0214] Furthermore, the emotion engine recognizes the user's emotional state and presents information that provides a sense of security if the user feels anxious. In this way, the system of the present invention effectively utilizes current risk maps and past weather data, and uses generative artificial intelligence and the emotion engine to provide highly accurate disaster predictions and appropriate information. This allows users to take more effective measures against future disaster risks.

[0215] The processing flow will be explained below.

[0216] Step 1:

[0217] The server connects to external data sources to retrieve the latest risk maps and historical weather data. Specifically, it accesses the government's hazard map API to retrieve risk map data, and weather data API to retrieve rainfall data and weather chart data for the past 30 years.

[0218] Step 2:

[0219] The server stores the acquired risk map data and weather data in a database, along with the corresponding timestamp and related meta information.

[0220] Step 3:

[0221] The server preprocesses the risk map data and weather data stored in the database. Specifically, it detects outliers and interpolates them based on the preceding and following data. Missing values ​​are also interpolated based on the preceding and following data. The server also converts the data into a time series format and normalizes it.

[0222] Step 4:

[0223] The server then uses the preprocessed data to train a generative artificial intelligence model (such as an LSTM model). Using past time-series data, the server builds a model to predict future disaster risk. This model quantifies flood risk and other natural disaster risks.

[0224] Step 5:

[0225] The server uses the trained model to predict disaster risk for the next five years. Specifically, it predicts the next five years on a monthly basis, quantifying the flood risk for each month and storing the results in a database.

[0226] Step 6:

[0227] Users access the system using a terminal and request prediction results through a web browser or application.

[0228] Step 7:

[0229] The server receives the user's request, retrieves the latest disaster risk prediction results from the database, and sends the results to the terminal.

[0230] Step 8:

[0231] The device visualizes the disaster risk prediction results it receives. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, making it easy for users to understand visually.

[0232] Step 9:

[0233] The emotion engine recognizes the user's emotions based on their input and actions. For example, it analyzes the text, voice, and facial expressions entered by the user to determine whether the user is feeling anxious or relieved.

[0234] Step 10:

[0235] The server then adjusts the presentation of disaster risk information based on the user's perceived emotions: for example, if the user feels anxious, it will present the information in a more reassuring tone and provide additional information on precautions and evacuation locations.

[0236] Step 11:

[0237] The device then revisits the adjusted information and provides it to the user, allowing them to check disaster risk information with greater peace of mind and take appropriate measures.

[0238] Example 2

[0239] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0240] Current disaster risk prediction systems predict disaster risk based on past data and current conditions, but no systems provide information that takes user emotions into consideration. This makes users more likely to feel anxious and makes it difficult for them to take appropriate disaster prevention actions. Furthermore, there are problems with insufficient data processing and imputation of outliers and missing values ​​to accurately predict disaster risk.

[0241] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0242] In this invention, the server includes means for acquiring a current risk map from an external data source, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now, means for recognizing a user's emotions and adjusting the method of presenting information, and means for visualizing the predicted disaster risks. This enables highly accurate disaster risk predictions and appropriate information provision while taking user emotions into consideration.

[0243] "External Data Source" refers to a source of data from an external source. Examples include government hazard map APIs and weather data APIs.

[0244] "Current risk maps" refer to map data showing disaster risks at the present time. These usually include risk information for earthquakes, floods, landslides, etc.

[0245] "Historical Weather Data" refers to datasets based on past weather information, including precipitation, weather maps, and temperatures.

[0246] "Preprocessing" refers to the process of cleaning, completing, and transforming data into a format suitable for machine learning models, including imputing outliers, filling in missing values, converting to time series data, and normalizing.

[0247] "Generative AI" refers to AI technology that uses past data to make future predictions, including deep learning models such as LSTM (Long Short-Term Memory).

[0248] "Disaster risk" is an indicator that shows the possibility of natural disasters such as earthquakes, floods, and landslides occurring.

[0249] "Emotion recognition" refers to the technology of analyzing and determining a user's emotional state (e.g., anxiety, relief, surprise, etc.) based on input data such as text, voice, and images.

[0250] "Adjusting the way information is presented" refers to the process of changing the means and content of information presentation based on the user's recognized emotions. For example, for a user who is feeling anxious, adding a reassuring message or specific countermeasures.

[0251] "Visualization" refers to the technology of displaying predicted disaster risks in visual formats such as graphs and maps, making it easier for users to understand the prediction results.

[0252] An "outlier" is a value that is extremely different from the other data points in a data set.

[0253] "Missing values" refers to the absence of values ​​in a dataset.

[0254] The present invention is a system that uses a current risk map and past weather data to predict disaster risks several years into the future using generative artificial intelligence, and further recognizes the user's emotions and adjusts the way information is presented. Specific embodiments for implementing the present invention are described below.

[0255] Data retrieval by the server

[0256] The server connects to external data sources (e.g., government hazard map APIs and weather data APIs) and periodically retrieves current risk map data and historical weather data (e.g., rainfall data and weather chart data from the past 30 years). The server stores this data in a database (e.g., PostgreSQL) and manages it securely.

[0257] Data preprocessing by the server

[0258] The server preprocesses the stored data. Specifically, it uses Python's Pandas library to detect and impute outliers. Missing values ​​are estimated and filled using linear interpolation and moving averages. Furthermore, rainfall data and weather map data are converted into time series data and normalized, allowing machine learning models to learn from the data efficiently.

[0259] Server-based analysis using generative AI

[0260] The server uses the preprocessed data to build and train a generative artificial intelligence model (e.g., an LSTM model) using a deep learning framework such as TensorFlow or Keras. The model is optimized over multiple epochs using a dataset split into training and test data.

[0261] Server-based future prediction

[0262] The server uses the trained model to predict disaster risk for the next five years. Specifically, it generates data quantifying monthly flood risk and stores the prediction results in a database. For example, it provides input to the generative AI model based on prompts such as, "Please predict the flood risk for the next five years."

[0263] Emotion engine that recognizes user emotions

[0264] The server is equipped with an emotion engine that recognizes the user's emotions from their input, actions, or image and audio data. Specifically, it uses Python's NLTK library and OpenCV to analyze text and audio and determine the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0265] Server-adjusted presentation of information

[0266] After the emotion engine recognizes the user's emotions, the server adjusts the way disaster risk information is presented based on the user's emotions. For example, if the user expresses strong anxiety, the server will provide information in a reassuring tone and additionally display specific evacuation locations and countermeasures to alleviate the user's anxiety.

[0267] Visualization of prediction results on the device

[0268] The terminal uses visualization libraries such as Matplotlib to display the forecast results obtained from the server in graphs and maps. For example, it can visually display flood risk forecast results for the next five years, providing a screen that is easy for users to understand.

[0269] Device and user interaction

[0270] Users can view the visualized forecast results on their devices to understand future disaster risks. Furthermore, the emotion engine continuously monitors the user's emotional state based on their input and feedback, and dynamically adjusts the way information is presented accordingly.

[0271] Specific examples

[0272] For example, if a user wants to know the "flood risk for the next five years," the server retrieves the latest hazard map data from the government's hazard map API and downloads historical rainfall data from the weather data API. These data are then interpolated for outliers and missing values ​​and converted into time-series data.

[0273] The data is then trained using an LSTM model to predict flood risk over the next five years. The results are stored in a database, and when users access the system from their devices, the forecast results are displayed in graphs and maps. The emotion engine recognizes the user's emotions and adds reassuring messages such as "There is a high risk of flooding, but there are many evacuation sites in this area and measures are in place."

[0274] This allows the system to provide highly accurate disaster risk predictions and appropriate information while taking into consideration the user's feelings.

[0275] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0276] Step 1:

[0277] The server connects to external data sources (for example, government hazard map APIs or weather data APIs) and periodically retrieves current risk map data and weather data from the past 30 years (rainfall data and weather chart data). The input is data provided through API requests, and the output is raw data that is stored in the server's database. Specifically, it uses the Python Requests library to access the API and retrieve and store the data.

[0278] Step 2:

[0279] The server preprocesses the stored data. The input is the data already retrieved in the database, and the output is the preprocessed data. Specifically, it uses the Python Pandas library to detect and impute outliers. It also estimates missing values ​​using linear interpolation and moving averages, converts them into time series data, and normalizes the data to make it suitable for machine learning models.

[0280] Step 3:

[0281] The server uses the preprocessed data to train a generative AI model, such as an LSTM model. The input is the preprocessed time series data, and the output is a trained LSTM model. Specifically, the model is defined using TensorFlow or Keras and trained through multiple epochs.

[0282] Step 4:

[0283] The server uses the trained model to predict disaster risk for the next five years. The input is the trained LSTM model, the latest risk map, and weather data, and the output is quantified disaster risk prediction data. For example, the server performs predictions based on prompt statements such as, "Please predict the flood risk for the next five years."

[0284] Step 5:

[0285] The server uses an emotion engine to recognize the user's emotional state. The input is the user's text, voice, and image data, and the output is the user's emotional state (anxiety, relief, surprise, etc.). Specifically, the server analyzes the text and voice data using Python's NLTK library and OpenCV.

[0286] Step 6:

[0287] The server adjusts the way disaster risk information is presented based on the recognized emotion. The input is the user's emotional state and disaster risk prediction data, and the output is customized information presentation. For example, if the user expresses anxiety, it displays a reassuring message along with specific evacuation locations and countermeasures.

[0288] Step 7:

[0289] The terminal obtains the forecast results from the server and provides them to the user using a visualization library such as Matplotlib. The input is disaster risk forecast data, and the output is visual information in the form of graphs and maps. Specifically, the terminal generates graphs and maps and displays them in a format that is easy for the user to understand.

[0290] Step 8:

[0291] The user checks and understands the visualized forecast results on their device. The input is the visualized disaster risk forecast data, and the output is the user's feedback and changes in emotional state. Based on the user's feedback, the emotion engine continuously monitors changes in emotions and dynamically adjusts the way information is presented.

[0292] In this way, the system takes into consideration the user's feelings while achieving highly accurate disaster risk prediction and providing appropriate information.

[0293] (Application example 2)

[0294] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0295] Currently, many disaster risk prediction systems simply predict future risks based on past data and do not take into account the user's emotional state or how that information is presented. This can lead to users feeling unnecessarily anxious or, conversely, becoming indifferent. Furthermore, particularly in electronic payment services, it is important to increase the security of transactions by providing appropriate disaster risk information. The objective of this invention is to provide a system that recognizes the user's emotions and presents appropriate disaster risk information based on those emotions, allowing users to use electronic payment services with peace of mind.

[0296] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0297] In this invention, the server includes means for acquiring a current risk map, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now, means for visualizing the predicted disaster risks, means for determining the emotional state of a user using an emotion engine, and means for adjusting the presentation method of disaster risk information based on the emotional state, thereby enabling the user to receive disaster risk information that is appropriately adjusted to match their emotional state.

[0298] A "current risk map" is data provided by governments and public institutions that shows the current risk of natural disasters in map form.

[0299] "Past weather data" refers to observational data on weather over a certain period of time in the past, including rainfall, temperature, and air pressure.

[0300] "Preprocessing" refers to processes such as organizing data, correcting outliers, and estimating missing values ​​in order to prepare raw data in a format suitable for machine learning models.

[0301] "Generative AI" is an AI technology that has the ability to generate new information from input data, and is often used to predict time series data.

[0302] An "emotion engine" is a technology that recognizes and classifies a user's emotional state based on their input, actions, images, voice, etc.

[0303] "Visualization" is a method of presenting data and prediction results in a visually easy-to-understand format, such as a graph or map.

[0304] "Adjusting presentation method" is a technology that provides information more appropriately by changing the method and content of information transmission depending on the user's current emotional state.

[0305] The present invention provides a system for predicting disaster risks in an area where a user resides and providing information that gives the user a sense of security based on the predicted disaster risks. Specific embodiments and their program processing will be described in detail below.

[0306] Data Acquisition Method

[0307] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years. The server stores the obtained data in a database.

[0308] Data preprocessing methods

[0309] The server preprocesses the stored risk map data and weather data. Specifically, it detects and complements outliers, and fills in missing values ​​by estimating them from previous and subsequent data. It also organizes rainfall data and weather map data and converts them into a time series format. It also normalizes the data and formats it in a way that is suitable for machine learning models.

[0310] A means of predictive analysis using generative artificial intelligence

[0311] The server uses the preprocessed data to train a generative artificial intelligence such as an LSTM model. Based on past time series data, it builds a model to predict future disaster risk. The server then uses the trained model to predict disaster risk over the next five years. For example, it generates data that quantifies monthly flood risk and stores the prediction results in a database.

[0312] A method for determining user emotions using an emotion engine

[0313] The server is equipped with an emotion engine that recognizes the user's emotions from their input, actions, images, and voice. For example, it analyzes the text and voice entered when the user accesses the system and determines the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0314] A means of adjusting information presentation

[0315] After the emotion engine recognizes the user's emotion, the server adjusts the way disaster risk information is presented based on the recognized emotion. For example, if the user expresses strong anxiety, the server will present disaster risk information in a more reassuring tone. It will also alleviate the user's anxiety by displaying specific evacuation locations and countermeasures.

[0316] Visualization of prediction results

[0317] The device retrieves the forecast results from the server and provides them to the user in a visual format. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, providing a design that is easy for users to understand.

[0318] Device and user interaction

[0319] Users can understand future disaster risks by viewing the visualized forecast results on their devices. Furthermore, based on the user's input and feedback, the emotion engine continuously monitors the user's emotional state and dynamically adjusts the way information is presented accordingly.

[0320] Examples of concrete examples and prompts

[0321] For example, suppose a user wants to know about flood risk. The server obtains current hazard map data from the government's hazard map API, and also obtains historical rainfall and weather chart data from a weather data API. These data are preprocessed on the server, and outliers and missing values ​​are interpolated. Next, using the preprocessed data, the server uses an LSTM model to train the data using generative artificial intelligence. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database. When the user then accesses the system on their device, the device retrieves the prediction results from the server and displays them as graphs or maps. For example, a graph predicting flood risk for the next five years based on data from the past 30 years can be displayed, providing a clear picture to the user. Furthermore, an emotion engine recognizes the user's emotional state, and if the user is anxious, information that provides reassurance is presented.

[0322] Example prompt for a generative AI model:

[0323] "Train an LSTM model to predict flood risk for the next five years based on rainfall data and meteorological map data from the past 30 years. Then, recognize user-input emotions and generate messages to provide reassurance and risk information if the user is anxious."

[0324] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0325] Step 1:

[0326] The server accesses the government's hazard map API and weather data API to obtain current risk map data and weather data for the past 30 years. The data obtained in this step is stored in a database as raw data.

[0327] Input: Government hazard map API and weather data API

[0328] Output: Current risk map data and weather data from the past 30 years

[0329] Step 2:

[0330] The server preprocesses the raw data stored in the database, detecting outliers and filling in missing values ​​by estimating them from surrounding data. It also converts the data into a time series format and normalizes it, making it suitable for machine learning models.

[0331] Input: Raw data

[0332] Output: Preprocessed data

[0333] Step 3:

[0334] The server uses the preprocessed data to build and train an LSTM model. The trained LSTM model is used to predict disaster risk for several years into the future from past data. The trained model then predicts monthly disaster risk for the next five years and stores the prediction results in a database.

[0335] Input: Preprocessed data

[0336] Output: Predicted disaster risk data

[0337] Step 4:

[0338] The server uses an emotion engine to analyze the user's input text and voice data, thereby determining the user's emotional state (e.g., anxiety, relief, surprise, etc.). The emotion recognition results are also stored in a database.

[0339] Input: User-entered text and voice data

[0340] Output: User's emotional state data

[0341] Step 5:

[0342] The server adjusts the way it presents disaster risk information based on the user's emotional state, as determined by the emotion engine. Specifically, if the user expresses anxiety, the server presents the information in a reassuring tone and additionally displays specific evacuation locations and countermeasures.

[0343] Input: User emotional state data and predicted disaster risk data

[0344] Output: Tailored disaster risk information

[0345] Step 6:

[0346] The device retrieves the prediction results from the server and presents them to the user in a visually understandable format, such as graphs or maps, using visualization libraries like Matplotlib, and adjusts the tone of the message to match the user's emotional state, if necessary.

[0347] Input: Tailored disaster risk information

[0348] Output: Disaster risk information that is visually easy for users to understand

[0349] Step 7:

[0350] Users can view the visualized forecast results on their devices to understand future disaster risks. Furthermore, the emotion engine continuously monitors the user's emotional state based on their input and feedback, dynamically adjusting how information is presented.

[0351] Input: Visualized disaster risk information, user feedback

[0352] Output: Continuously updated emotional state data and adjusted risk information

[0353] The specific processing unit 290 transmits the result of the specific processing 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 a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the 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.

[0354] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search<url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0355] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0356] [Second embodiment]

[0357] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0358] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0359] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "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).

[0360] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, 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 microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0361] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0362] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0363] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0364] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0365] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[0366] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0367] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0368] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0369] The present invention relates to a system that uses a generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future, based on a current risk map and past weather data. Below, we will explain in natural language the mode for carrying out the present invention and the program processing. We will also provide specific examples.

[0370] Data retrieval by the server

[0371] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including, for example, rainfall data and weather chart data from the past 30 years. The server retrieves this data and stores it in a database.

[0372] Data preprocessing by the server

[0373] The server preprocesses the stored risk map data and weather data by imputing outliers and missing values, integrating the data, and converting it into a time series format. It also normalizes the data to make it suitable for machine learning models.

[0374] Server-based analysis using generative AI

[0375] The server uses the pre-processed data and uses generative artificial intelligence (e.g., LSTM models) to learn from past data, allowing it to extract temporal patterns and trends and predict disaster risk for several years into the future.

[0376] Server-based future prediction

[0377] The server uses the trained model to predict future disaster risks, such as flood risk over the next five years, and the results are stored in a database.

[0378] Visualization of prediction results on the device

[0379] The device receives the prediction results from the server and provides them to the user in a visual format, such as graphs and maps, in a way that is easy for the user to understand.

[0380] Specific examples

[0381] For example, suppose a user wants to know the risk of flooding. The server retrieves current hazard map data from the government's hazard map API, and also retrieves past rainfall data and weather chart data from the weather data API. These data are preprocessed on the server, and outliers and missing values ​​are filled in.

[0382] Next, using the preprocessed data, the server trains the data using a generative artificial intelligence (AI) model with an LSTM model. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database.

[0383] When a user accesses the system via a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps. For example, a graph showing flood risk forecasts for the next five years based on data from the past 30 years can be displayed in an easy-to-understand manner.

[0384] In this way, the system of the present invention effectively utilizes current risk maps and past weather data to perform highly accurate disaster predictions using generative artificial intelligence, and visualizes the results and provides them to users, allowing them to take appropriate measures against future disaster risks.

[0385] The processing flow will be explained below.

[0386] Step 1:

[0387] The server connects to external data sources to retrieve the latest risk maps and historical weather data, specifically accessing APIs from government and meteorological agencies to retrieve rainfall data and weather chart data for the past 30 years.

[0388] Step 2:

[0389] The server stores the acquired risk map data and weather data in a database, along with timestamps and related information corresponding to the acquired data.

[0390] Step 3:

[0391] The server preprocesses the data stored in the database. Specifically, it detects and complements outliers, estimates missing values ​​based on previous and subsequent data, and fills in missing values. It also organizes rainfall data and meteorological map data and converts them into a time series format.

[0392] Step 4:

[0393] The server normalizes the preprocessed data, calculating the mean and standard deviation of the data and standardizing each data point to prepare it for efficient learning by the generative AI.

[0394] Step 5:

[0395] The server uses the normalized data to train generative AI models such as LSTM models, and builds models to predict future disaster risks based on past time-series data.

[0396] Step 6:

[0397] The server uses the trained model to predict disaster risk over the next five years, generating data that quantifies monthly flood risk, for example, and storing the prediction results in a database.

[0398] Step 7:

[0399] A user accesses the system from a terminal and sends a request for disaster prediction results to the server through an application or web browser on the terminal.

[0400] Step 8:

[0401] The device receives the prediction results from the server, and the server sends the latest disaster prediction results stored in the database to the device.

[0402] Step 9:

[0403] The device visualizes the obtained prediction results, specifically using a visualization library such as Matplotlib to display the predicted disaster risk as graphs and maps.

[0404] Step 10:

[0405] Users can view the visualized forecast results on their devices, which allows them to understand future disaster risks and take appropriate measures.

[0406] Example 1

[0407] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0408] In recent years, the frequency and scale of natural disasters have increased, making it important to accurately predict future disaster risks. However, conventional systems are unable to effectively utilize past data, making it difficult to accurately predict disaster risks several years into the future. The present invention aims to solve this problem by providing a system that utilizes current risk maps and past weather data obtained from external data sources to achieve highly accurate disaster risk predictions.

[0409] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0410] In this invention, the server includes means for periodically acquiring a current risk map from an external data source, means for periodically acquiring past weather data from an external data source, means for preprocessing the current risk map and the past weather data by complementing, integrating, and normalizing outliers and missing values, means for training a generative artificial intelligence model using the preprocessed data to predict disaster risk several years from now, and means for visualizing the predicted disaster risk in the form of a graph or map. This makes it possible to effectively utilize past data to accurately predict future disaster risk and provide it to users visually.

[0411] A "risk map" is a map that geographically represents risks such as natural disasters and accidents.

[0412] "Weather data" refers to data that records past weather conditions such as rainfall, temperature, and wind speed on an hourly basis.

[0413] An "outlier" is a value that is significantly different from the rest of the data, and is a value that deviates significantly from the normal distribution.

[0414] "Missing values" refer to portions of a dataset where no values ​​are recorded.

[0415] "Preprocessing" refers to the process of preparing acquired data in a format suitable for analysis and model learning, and includes correcting outliers, filling in missing values, and integrating and normalizing data.

[0416] A "generative artificial intelligence model" is an algorithm that learns from input data and generates and predicts future data and results; specifically, it refers to neural network models such as LSTM (long short-term memory).

[0417] "Visualization" refers to displaying data and prediction results in a visual format such as a graph or map so that people can understand them intuitively.

[0418] A "data source" is an external resource or system from which data is obtained, such as an API or database.

[0419] "Integration" is the process of combining different datasets into a single dataset.

[0420] "Regularly" means to repeat at regular intervals.

[0421] "Normalization" is a process of converting each data point to fall within a certain range in order to unify the scale of the data.

[0422] MODE FOR CARRYING OUT THE INVENTION

[0423] The present invention provides a system that uses a generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future, based on a current risk map and past weather data. The following describes specific embodiments of the present invention.

[0424] Data retrieval by the server

[0425] The server periodically accesses external data sources to obtain the latest risk maps and historical weather data. For example, the server sends requests to government hazard map APIs and weather data APIs, analyzes the data received in response, and stores it in a database. This data includes rainfall data and weather map data from the past 30 years.

[0426] Data preprocessing by the server

[0427] The server preprocesses the acquired data. This includes imputing outliers and missing values, integrating risk map data with weather data, and normalizing the data. Specifically, anomalies are detected using an anomaly detection algorithm and imputed with appropriate values. Missing data is imputed with the mean or median value. The integrated dataset is then preprocessed using techniques such as z-score normalization.

[0428] Server-based analysis using generative AI

[0429] The server uses the preprocessed data to train a generative artificial intelligence (e.g., an LSTM model), which can learn patterns and trends from past data and predict future disaster risks. The server builds the LSTM model and splits the dataset into training data and validation data for training. After this process is complete, the server saves the trained model.

[0430] For example, the LSTM model predicts flood risk for the next five years based on rainfall data from the past 30 years. The server stores the prediction results in a database for future analysis and visualization.

[0431] Visualization of prediction results on the device

[0432] The terminal retrieves the forecast results from the server and provides them visually to the user. This includes displaying the forecast data in the form of graphs and maps. When a user accesses the system, the terminal sends a request to the server to retrieve the forecast results. Heat maps and graphs are then generated based on the retrieved data and displayed on the user interface. This allows the user to intuitively understand the flood risk forecast for the next five years based on data from the past 30 years.

[0433] Specific examples

[0434] For example, suppose a user wants to know the flood risk for the next five years. The server first obtains the latest risk map data from the government's hazard map API, and simultaneously collects rainfall data and weather chart data from the past 30 years from the weather data API. Next, the server preprocesses this data and uses appropriate algorithms to fill in outliers and missing values. The server then trains this data using an LSTM model to predict the flood risk for the next five years. After saving this forecast data in a database, when the user accesses the system using a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps.

[0435] Prompt Sentence Examples

[0436] "Predict flood risk over the next five years. Use rainfall data from the past 30 years and current risk map data. The generative AI used is an LSTM model."

[0437] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0438] Step 1: Data Acquisition

[0439] The server periodically retrieves current risk maps and historical weather data from external data sources (government hazard map APIs and weather data APIs). Specifically, the server sends requests to the API, analyzes the JSON or CSV format data received as a response, and stores it in a database. The input is the response data from the API, and the output is data saved in the database.

[0440] Step 2: Data Preprocessing

[0441] The server preprocesses the risk map data and weather data obtained from the database. First, it detects outliers and missing values ​​and fills them with appropriate values. Next, it integrates the risk map data and weather data and converts them into time-series data. At this time, it normalizes each data point using methods such as z-score normalization. The input is raw data obtained from the database, and the output is preprocessed, well-formatted data.

[0442] Specific working example:

[0443] An outlier detection algorithm is used to impute outliers.

[0444] Impute missing data with the mean or median.

[0445] Combine multiple datasets and transform them into time series data.

[0446] The data is scaled using a normalization algorithm.

[0447] Step 3: Model training

[0448] The server uses the preprocessed data to build a generative artificial intelligence model (such as an LSTM model) and trains it. It splits the data into training and validation data, and then trains and evaluates the model. The input is the preprocessed data, and the output is the trained LSTM model.

[0449] Specific working example:

[0450] Initialize the LSTM model and set the hyperparameters.

[0451] Split the dataset into training and validation sets.

[0452] A learning algorithm is used to train the model.

[0453] Validation data is used to evaluate the accuracy of the model and adjust the model as needed.

[0454] Step 4: Predict the future

[0455] The server uses a trained LSTM model to predict future disaster risks. The prediction period is multiple years, and risk assessments are performed for each region. The input is the latest risk map data, and the output is predicted future disaster risk data.

[0456] Specific working example:

[0457] Load a trained model.

[0458] Provide the latest risk map data as input to the model.

[0459] Predict future disaster risks using models.

[0460] The prediction results are saved in a database.

[0461] Step 5: Visualize the results

[0462] The terminal obtains the prediction results from the server and provides them visually to the user. Specifically, it generates heat maps and graphs and displays them on the user interface. The input is the prediction data obtained from the server, and the output is a display in the form of a graph or map.

[0463] Specific working example:

[0464] The terminal sends a request to the server to obtain the prediction data.

[0465] Generate graphs and heat maps based on the acquired data.

[0466] The generated graphs and heat maps are displayed in a user interface.

[0467] (Application example 1)

[0468] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0469] In recent years, the frequent occurrence of natural disasters has created a need for damage prediction and prevention. However, conventional disaster risk prediction systems do not fully utilize past data and current risk information, resulting in limitations in the accuracy of predictions and the implementation of disaster prevention measures. Furthermore, many systems only visualize disaster risks and do not adequately communicate them to users or present effective countermeasures. This presents a problem in that it is difficult for users to be aware of risks on a daily basis and take prompt and appropriate measures.

[0470] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0471] In this invention, the server includes means for acquiring a current risk map, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risk several years from now, means for visualizing the predicted disaster risk, means for sending a real-time alert to the user based on the predicted disaster risk, and means for the user to check the disaster risk around their place of residence or workplace on a map. This not only enables highly accurate disaster risk prediction, but also enables the user to be aware of disaster risk on a daily basis and take prompt and appropriate disaster prevention measures.

[0472] A "current risk map" is map information that shows the current risk of natural disasters in a specific area.

[0473] "Past weather data" refers to historical data on weather information such as rainfall and weather charts for a specific period of time.

[0474] "Preprocessing" is the process of correcting abnormal and missing values ​​in the acquired data, and integrating and normalizing it to prepare it into an analyzable format.

[0475] "Generative AI" is an artificial intelligence technology that learns from time-series data and predicts future disaster risks.

[0476] "Disaster risk several years from now" is forecast information that indicates the possibility of a natural disaster occurring in a specific area several years from now.

[0477] "Visualization" refers to the display of numerical data and forecast results in visual formats such as graphs and maps.

[0478] "Real-time alerts" is a function that instantly notifies users of warnings based on predicted disaster risks.

[0479] "Means of checking on a map" refers to a function that allows users to visually check disaster risks around their place of residence or workplace using a geographic information system.

[0480] This invention is a system that uses generative artificial intelligence to predict and visualize disaster risks several years into the future, using a current risk map and past weather data. Specific embodiments for implementing this invention are described below.

[0481] System Program

[0482] 1. Data Acquisition

[0483] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to retrieve the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years.

[0484] 2. Data Preprocessing

[0485] The server complements outliers and missing values ​​in the acquired risk map data and weather data, integrates the data, converts it into a time series format, and normalizes the data to make it suitable for machine learning models.

[0486] 3. Analysis using generative artificial intelligence

[0487] The server uses the pre-processed data and uses generative artificial intelligence (e.g., LSTM models) to learn from past data, allowing it to extract temporal patterns and trends and predict disaster risk for several years into the future.

[0488] 4. Predicting the future

[0489] The server uses the trained model to predict future disaster risks, for example, predicting flood risk over the next five years, and stores the results in a database.

[0490] 5. Visualization of prediction results

[0491] The device receives the prediction results from the server and provides them to the user in a visual format, such as graphs and maps, in a way that is easy for the user to understand.

[0492] 6. Real-time alerts

[0493] Based on the predicted disaster risk, the server sends users real-time alerts to urge them to take measures, and also provides a function that allows users to check the disaster risk around their residence or workplace on a map.

[0494] Hardware and software used

[0495] Hardware:

[0496] Server: A high-performance server handles the analysis, performing data acquisition, preprocessing, and predictive analysis.

[0497] Device: The device used by the user, such as a smartphone or tablet.

[0498] software:

[0499] API client: Uses government hazard map APIs and weather data APIs to obtain data.

[0500] Data preprocessing: Data integration, normalization, and outlier processing using the pandas library.

[0501] Generative AI: Analysis is performed using TensorFlow and LSTM models.

[0502] Visualization: Use matplotlib to visualize the prediction results.

[0503] Specific examples

[0504] For example, if a user requests the system to "predict flood risk for the next five years and display it by region," the server retrieves current hazard map data from the government's hazard map API and collects historical rainfall data and weather chart data from the weather data API. These data are preprocessed and trained using generative artificial intelligence (AI) with an LSTM model. The server then predicts flood risk for the next five years and stores the results in a database.

[0505] When a user accesses the system from their smartphone, the device retrieves the forecast results from the server and displays them in graphs and maps. For example, it provides a graph and map showing flood risk forecasts for the next five years, making it easy for users to understand.

[0506] In this way, the system of the present invention effectively utilizes current risk maps and past weather data to perform highly accurate disaster predictions using generative artificial intelligence, and visualizes the results and provides them to users, allowing them to take appropriate measures against future disaster risks.

[0507] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0508] Step 1: Data Acquisition

[0509] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to retrieve current risk map data and historical weather data. Inputs include API endpoints and parameters, and outputs JSON-formatted data. The server stores this data in a local database.

[0510] Step 2: Data Preprocessing

[0511] The server preprocesses the acquired risk map data and weather data. Specifically, it complements outliers and missing values, integrates the data, and converts it into a time series format. The acquired JSON data is the input, and the formatted data is the output. This processing uses the pandas library. The server detects outliers and complements them with appropriate values. It also merges data obtained from multiple data sources and formats it into time series data.

[0512] Step 3: Generative AI learning

[0513] The server uses the preprocessed data to learn from past data using a generative artificial intelligence (e.g., an LSTM model). The input is the preprocessed time series data, and the output is the trained model. The server uses the TensorFlow library to build the LSTM model and train the model on the time series data.

[0514] Step 4: Predict the future

[0515] The server uses the trained LSTM model to predict future disaster risks. For example, it predicts flood risks over the next five years. The inputs are the trained model and time series data, and the output is future disaster risk prediction data. The server stores this prediction data in a database.

[0516] Step 5: Visualize the prediction results

[0517] The terminal retrieves the prediction results from the server and provides them to the user in a visual format. Specifically, it displays the prediction data in the form of graphs and maps. The input is the prediction result data, and the output is visualized information. The terminal uses the matplotlib library and a geographic information system (GIS) to display the information in a visually easy-to-understand format.

[0518] Step 6: Real-time alerts

[0519] The server sends real-time alerts to users based on predicted disaster risks. The input is predicted disaster risk data, and the output is a notification message. The server sends push notifications to users' devices to provide immediate warnings. Based on the notifications received, users can immediately take appropriate disaster prevention measures.

[0520] This configuration allows users to receive highly accurate disaster prediction information based on current risk maps and past weather data, enabling them to take appropriate action in real time.

[0521] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0522] The present invention relates to a system that uses a current risk map and past weather data to predict disaster risks several years into the future using generative artificial intelligence, and also recognizes the user's emotions to adjust the way information is presented. Below, we will explain in natural language the modes for implementing the present invention and the program's processing. Specific examples will also be included.

[0523] Data retrieval by the server

[0524] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years. The server stores the obtained data in a database.

[0525] Data preprocessing by the server

[0526] The server preprocesses the stored risk map data and weather data. Specifically, it detects and complements outliers, and fills in missing values ​​by estimating them from previous and subsequent data. It also organizes rainfall data and weather map data and converts them into a time series format. It also normalizes the data and formats it in a way that is suitable for machine learning models.

[0527] Server-based analysis using generative AI

[0528] The server uses the preprocessed data to train generative AI models such as LSTM models, which build models to predict future disaster risks based on past time-series data.

[0529] Server-based future prediction

[0530] The server uses the trained model to predict disaster risk over the next five years, generating data that quantifies monthly flood risk, for example, and storing the prediction results in a database.

[0531] Emotion engine that recognizes user emotions

[0532] The server is equipped with an emotion engine that recognizes the user's emotions from their inputs, actions, images, and voice. For example, it analyzes the text and voice entered when the user accesses the system and determines the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0533] Server-adjusted presentation of information

[0534] After the emotion engine recognizes the user's emotion, the server adjusts the way disaster risk information is presented based on the recognized emotion. For example, if the user expresses strong anxiety, the server will present disaster risk information in a more reassuring tone. It will also alleviate the user's anxiety by displaying specific evacuation locations and countermeasures.

[0535] Visualization of prediction results on the device

[0536] The device retrieves the forecast results from the server and provides them to the user in a visual format. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, providing a design that is easy for users to understand.

[0537] Device and user interaction

[0538] Users can understand future disaster risks by viewing the visualized forecast results on their devices. Furthermore, based on the user's input and feedback, the emotion engine continuously monitors the user's emotional state and dynamically adjusts the way information is presented accordingly.

[0539] Specific examples

[0540] For example, suppose a user wants to know the risk of flooding. The server retrieves current hazard map data from the government's hazard map API, and also retrieves past rainfall data and weather chart data from the weather data API. These data are preprocessed on the server, and outliers and missing values ​​are filled in.

[0541] Next, the server uses the preprocessed data to train the data using a generative artificial intelligence (AI) model with an LSTM model. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database.

[0542] When a user accesses the system via a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps. For example, a graph showing flood risk forecasts for the next five years based on data from the past 30 years can be displayed in an easy-to-understand manner.

[0543] Furthermore, the emotion engine recognizes the user's emotional state and presents information that provides a sense of security if the user feels anxious. In this way, the system of the present invention effectively utilizes current risk maps and past weather data, and uses generative artificial intelligence and the emotion engine to provide highly accurate disaster predictions and appropriate information. This allows users to take more effective measures against future disaster risks.

[0544] The processing flow will be explained below.

[0545] Step 1:

[0546] The server connects to external data sources to retrieve the latest risk maps and historical weather data. Specifically, it accesses the government's hazard map API to retrieve risk map data, and weather data API to retrieve rainfall data and weather chart data for the past 30 years.

[0547] Step 2:

[0548] The server stores the acquired risk map data and weather data in a database, along with the corresponding timestamp and related meta information.

[0549] Step 3:

[0550] The server preprocesses the risk map data and weather data stored in the database. Specifically, it detects outliers and interpolates them based on the preceding and following data. Missing values ​​are also interpolated based on the preceding and following data. The server also converts the data into a time series format and normalizes it.

[0551] Step 4:

[0552] The server then uses the preprocessed data to train a generative artificial intelligence model (such as an LSTM model). Using past time-series data, the server builds a model to predict future disaster risk. This model quantifies flood risk and other natural disaster risks.

[0553] Step 5:

[0554] The server uses the trained model to predict disaster risk for the next five years. Specifically, it predicts the next five years on a monthly basis, quantifying the flood risk for each month and storing the results in a database.

[0555] Step 6:

[0556] Users access the system using a terminal and request prediction results through a web browser or application.

[0557] Step 7:

[0558] The server receives the user's request, retrieves the latest disaster risk prediction results from the database, and sends the results to the terminal.

[0559] Step 8:

[0560] The device visualizes the disaster risk prediction results it receives. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, making it easy for users to understand visually.

[0561] Step 9:

[0562] The emotion engine recognizes the user's emotions based on their input and actions. For example, it analyzes the text, voice, and facial expressions entered by the user to determine whether the user is feeling anxious or relieved.

[0563] Step 10:

[0564] The server then adjusts the presentation of disaster risk information based on the user's perceived emotions: for example, if the user feels anxious, it will present the information in a more reassuring tone and provide additional information on precautions and evacuation locations.

[0565] Step 11:

[0566] The device then revisits the adjusted information and provides it to the user, allowing them to check disaster risk information with greater peace of mind and take appropriate measures.

[0567] Example 2

[0568] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0569] Current disaster risk prediction systems predict disaster risk based on past data and current conditions, but no systems provide information that takes user emotions into consideration. This makes users more likely to feel anxious and makes it difficult for them to take appropriate disaster prevention actions. Furthermore, there are problems with insufficient data processing and imputation of outliers and missing values ​​to accurately predict disaster risk.

[0570] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0571] In this invention, the server includes means for acquiring a current risk map from an external data source, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now, means for recognizing a user's emotions and adjusting the method of presenting information, and means for visualizing the predicted disaster risks. This enables highly accurate disaster risk predictions and appropriate information provision while taking user emotions into consideration.

[0572] "External Data Source" refers to a source of data from an external source. Examples include government hazard map APIs and weather data APIs.

[0573] "Current risk maps" refer to map data showing disaster risks at the present time. These usually include risk information for earthquakes, floods, landslides, etc.

[0574] "Historical Weather Data" refers to datasets based on past weather information, including precipitation, weather maps, and temperatures.

[0575] "Preprocessing" refers to the process of cleaning, completing, and transforming data into a format suitable for machine learning models, including imputing outliers, filling in missing values, converting to time series data, and normalizing.

[0576] "Generative AI" refers to AI technology that uses past data to make future predictions, including deep learning models such as LSTM (Long Short-Term Memory).

[0577] "Disaster risk" is an indicator that shows the possibility of natural disasters such as earthquakes, floods, and landslides occurring.

[0578] "Emotion recognition" refers to the technology of analyzing and determining a user's emotional state (e.g., anxiety, relief, surprise, etc.) based on input data such as text, voice, and images.

[0579] "Adjusting the way information is presented" refers to the process of changing the means and content of information presentation based on the user's recognized emotions. For example, for a user who is feeling anxious, adding a reassuring message or specific countermeasures.

[0580] "Visualization" refers to the technology of displaying predicted disaster risks in visual formats such as graphs and maps, making it easier for users to understand the prediction results.

[0581] An "outlier" is a value that is extremely different from the other data points in a data set.

[0582] "Missing values" refers to the absence of values ​​in a dataset.

[0583] The present invention is a system that uses a current risk map and past weather data to predict disaster risks several years into the future using generative artificial intelligence, and further recognizes the user's emotions and adjusts the way information is presented. Specific embodiments for implementing the present invention are described below.

[0584] Data retrieval by the server

[0585] The server connects to external data sources (e.g., government hazard map APIs and weather data APIs) and periodically retrieves current risk map data and historical weather data (e.g., rainfall data and weather chart data from the past 30 years). The server stores this data in a database (e.g., PostgreSQL) and manages it securely.

[0586] Data preprocessing by the server

[0587] The server preprocesses the stored data. Specifically, it uses Python's Pandas library to detect and impute outliers. Missing values ​​are estimated and filled using linear interpolation and moving averages. Furthermore, rainfall data and weather map data are converted into time series data and normalized, allowing machine learning models to learn from the data efficiently.

[0588] Server-based analysis using generative AI

[0589] The server uses the preprocessed data to build and train a generative artificial intelligence model (e.g., an LSTM model) using a deep learning framework such as TensorFlow or Keras. The model is optimized over multiple epochs using a dataset split into training and test data.

[0590] Server-based future prediction

[0591] The server uses the trained model to predict disaster risk for the next five years. Specifically, it generates data quantifying monthly flood risk and stores the prediction results in a database. For example, it provides input to the generative AI model based on prompts such as, "Please predict the flood risk for the next five years."

[0592] Emotion engine that recognizes user emotions

[0593] The server is equipped with an emotion engine that recognizes the user's emotions from their input, actions, or image and audio data. Specifically, it uses Python's NLTK library and OpenCV to analyze text and audio and determine the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0594] Server-adjusted presentation of information

[0595] After the emotion engine recognizes the user's emotions, the server adjusts the way disaster risk information is presented based on the user's emotions. For example, if the user expresses strong anxiety, the server will provide information in a reassuring tone and additionally display specific evacuation locations and countermeasures to alleviate the user's anxiety.

[0596] Visualization of prediction results on the device

[0597] The terminal uses visualization libraries such as Matplotlib to display the forecast results obtained from the server in graphs and maps. For example, it can visually display flood risk forecast results for the next five years, providing a screen that is easy for users to understand.

[0598] Device and user interaction

[0599] Users can view the visualized forecast results on their devices to understand future disaster risks. Furthermore, the emotion engine continuously monitors the user's emotional state based on their input and feedback, and dynamically adjusts the way information is presented accordingly.

[0600] Specific examples

[0601] For example, if a user wants to know the "flood risk for the next five years," the server retrieves the latest hazard map data from the government's hazard map API and downloads historical rainfall data from the weather data API. These data are then interpolated for outliers and missing values ​​and converted into time-series data.

[0602] The data is then trained using an LSTM model to predict flood risk over the next five years. The results are stored in a database, and when users access the system from their devices, the forecast results are displayed in graphs and maps. The emotion engine recognizes the user's emotions and adds reassuring messages such as "There is a high risk of flooding, but there are many evacuation sites in this area and measures are in place."

[0603] This allows the system to provide highly accurate disaster risk predictions and appropriate information while taking into consideration the user's feelings.

[0604] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0605] Step 1:

[0606] The server connects to external data sources (for example, government hazard map APIs or weather data APIs) and periodically retrieves current risk map data and weather data from the past 30 years (rainfall data and weather chart data). The input is data provided through API requests, and the output is raw data that is stored in the server's database. Specifically, it uses the Python Requests library to access the API and retrieve and store the data.

[0607] Step 2:

[0608] The server preprocesses the stored data. The input is the data already retrieved in the database, and the output is the preprocessed data. Specifically, it uses the Python Pandas library to detect and impute outliers. It also estimates missing values ​​using linear interpolation and moving averages, converts them into time series data, and normalizes the data to make it suitable for machine learning models.

[0609] Step 3:

[0610] The server uses the preprocessed data to train a generative AI model, such as an LSTM model. The input is the preprocessed time series data, and the output is a trained LSTM model. Specifically, the model is defined using TensorFlow or Keras and trained through multiple epochs.

[0611] Step 4:

[0612] The server uses the trained model to predict disaster risk for the next five years. The input is the trained LSTM model, the latest risk map, and weather data, and the output is quantified disaster risk prediction data. For example, the server performs predictions based on prompt statements such as, "Please predict the flood risk for the next five years."

[0613] Step 5:

[0614] The server uses an emotion engine to recognize the user's emotional state. The input is the user's text, voice, and image data, and the output is the user's emotional state (anxiety, relief, surprise, etc.). Specifically, the server analyzes the text and voice data using Python's NLTK library and OpenCV.

[0615] Step 6:

[0616] The server adjusts the way disaster risk information is presented based on the recognized emotion. The input is the user's emotional state and disaster risk prediction data, and the output is customized information presentation. For example, if the user expresses anxiety, it displays a reassuring message along with specific evacuation locations and countermeasures.

[0617] Step 7:

[0618] The terminal obtains the forecast results from the server and provides them to the user using a visualization library such as Matplotlib. The input is disaster risk forecast data, and the output is visual information in the form of graphs and maps. Specifically, the terminal generates graphs and maps and displays them in a format that is easy for the user to understand.

[0619] Step 8:

[0620] The user checks and understands the visualized forecast results on their device. The input is the visualized disaster risk forecast data, and the output is the user's feedback and changes in emotional state. Based on the user's feedback, the emotion engine continuously monitors changes in emotions and dynamically adjusts the way information is presented.

[0621] In this way, the system takes into consideration the user's feelings while achieving highly accurate disaster risk prediction and providing appropriate information.

[0622] (Application example 2)

[0623] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0624] Currently, many disaster risk prediction systems simply predict future risks based on past data and do not take into account the user's emotional state or how that information is presented. This can lead to users feeling unnecessarily anxious or, conversely, becoming indifferent. Furthermore, particularly in electronic payment services, it is important to increase the security of transactions by providing appropriate disaster risk information. The objective of this invention is to provide a system that recognizes the user's emotions and presents appropriate disaster risk information based on those emotions, allowing users to use electronic payment services with peace of mind.

[0625] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0626] In this invention, the server includes means for acquiring a current risk map, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now, means for visualizing the predicted disaster risks, means for determining the emotional state of a user using an emotion engine, and means for adjusting the presentation method of disaster risk information based on the emotional state, thereby enabling the user to receive disaster risk information that is appropriately adjusted to match their emotional state.

[0627] A "current risk map" is data provided by governments and public institutions that shows the current risk of natural disasters in map form.

[0628] "Past weather data" refers to observational data on weather over a certain period of time in the past, including rainfall, temperature, and air pressure.

[0629] "Preprocessing" refers to processes such as organizing data, correcting outliers, and estimating missing values ​​in order to prepare raw data in a format suitable for machine learning models.

[0630] "Generative AI" is an AI technology that has the ability to generate new information from input data, and is often used to predict time series data.

[0631] An "emotion engine" is a technology that recognizes and classifies a user's emotional state based on their input, actions, images, voice, etc.

[0632] "Visualization" is a method of presenting data and prediction results in a visually easy-to-understand format, such as a graph or map.

[0633] "Adjusting presentation method" is a technology that provides information more appropriately by changing the method and content of information transmission depending on the user's current emotional state.

[0634] The present invention provides a system for predicting disaster risks in an area where a user resides and providing information that gives the user a sense of security based on the predicted disaster risks. Specific embodiments and their program processing will be described in detail below.

[0635] Data Acquisition Method

[0636] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years. The server stores the obtained data in a database.

[0637] Data preprocessing methods

[0638] The server preprocesses the stored risk map data and weather data. Specifically, it detects and complements outliers, and fills in missing values ​​by estimating them from previous and subsequent data. It also organizes rainfall data and weather map data and converts them into a time series format. It also normalizes the data and formats it in a way that is suitable for machine learning models.

[0639] A means of predictive analysis using generative artificial intelligence

[0640] The server uses the preprocessed data to train a generative artificial intelligence such as an LSTM model. Based on past time series data, it builds a model to predict future disaster risk. The server then uses the trained model to predict disaster risk over the next five years. For example, it generates data that quantifies monthly flood risk and stores the prediction results in a database.

[0641] A method for determining user emotions using an emotion engine

[0642] The server is equipped with an emotion engine that recognizes the user's emotions from their input, actions, images, and voice. For example, it analyzes the text and voice entered when the user accesses the system and determines the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0643] A means of adjusting information presentation

[0644] After the emotion engine recognizes the user's emotion, the server adjusts the way disaster risk information is presented based on the recognized emotion. For example, if the user expresses strong anxiety, the server will present disaster risk information in a more reassuring tone. It will also alleviate the user's anxiety by displaying specific evacuation locations and countermeasures.

[0645] Visualization of prediction results

[0646] The device retrieves the forecast results from the server and provides them to the user in a visual format. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, providing a design that is easy for users to understand.

[0647] Device and user interaction

[0648] Users can understand future disaster risks by viewing the visualized forecast results on their devices. Furthermore, based on the user's input and feedback, the emotion engine continuously monitors the user's emotional state and dynamically adjusts the way information is presented accordingly.

[0649] Examples of concrete examples and prompts

[0650] For example, suppose a user wants to know about flood risk. The server obtains current hazard map data from the government's hazard map API, and also obtains historical rainfall and weather chart data from a weather data API. These data are preprocessed on the server, and outliers and missing values ​​are interpolated. Next, using the preprocessed data, the server uses an LSTM model to train the data using generative artificial intelligence. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database. When the user then accesses the system on their device, the device retrieves the prediction results from the server and displays them as graphs or maps. For example, a graph predicting flood risk for the next five years based on data from the past 30 years can be displayed, providing a clear picture to the user. Furthermore, an emotion engine recognizes the user's emotional state, and if the user is anxious, information that provides reassurance is presented.

[0651] Example prompt for a generative AI model:

[0652] "Train an LSTM model to predict flood risk for the next five years based on rainfall data and meteorological map data from the past 30 years. Then, recognize user-input emotions and generate messages to provide reassurance and risk information if the user is anxious."

[0653] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0654] Step 1:

[0655] The server accesses the government's hazard map API and weather data API to obtain current risk map data and weather data for the past 30 years. The data obtained in this step is stored in a database as raw data.

[0656] Input: Government hazard map API and weather data API

[0657] Output: Current risk map data and weather data from the past 30 years

[0658] Step 2:

[0659] The server preprocesses the raw data stored in the database, detecting outliers and filling in missing values ​​by estimating them from surrounding data. It also converts the data into a time series format and normalizes it, making it suitable for machine learning models.

[0660] Input: Raw data

[0661] Output: Preprocessed data

[0662] Step 3:

[0663] The server uses the preprocessed data to build and train an LSTM model. The trained LSTM model is used to predict disaster risk for several years into the future from past data. The trained model then predicts monthly disaster risk for the next five years and stores the prediction results in a database.

[0664] Input: Preprocessed data

[0665] Output: Predicted disaster risk data

[0666] Step 4:

[0667] The server uses an emotion engine to analyze the user's input text and voice data, thereby determining the user's emotional state (e.g., anxiety, relief, surprise, etc.). The emotion recognition results are also stored in a database.

[0668] Input: User-entered text and voice data

[0669] Output: User's emotional state data

[0670] Step 5:

[0671] The server adjusts the way it presents disaster risk information based on the user's emotional state, as determined by the emotion engine. Specifically, if the user expresses anxiety, the server presents the information in a reassuring tone and additionally displays specific evacuation locations and countermeasures.

[0672] Input: User emotional state data and predicted disaster risk data

[0673] Output: Tailored disaster risk information

[0674] Step 6:

[0675] The device retrieves the prediction results from the server and presents them to the user in a visually understandable format, such as graphs or maps, using visualization libraries like Matplotlib, and adjusts the tone of the message to match the user's emotional state, if necessary.

[0676] Input: Tailored disaster risk information

[0677] Output: Disaster risk information that is visually easy for users to understand

[0678] Step 7:

[0679] Users can view the visualized forecast results on their devices to understand future disaster risks. Furthermore, the emotion engine continuously monitors the user's emotional state based on their input and feedback, dynamically adjusting how information is presented.

[0680] Input: Visualized disaster risk information, user feedback

[0681] Output: Continuously updated emotional state data and adjusted risk information

[0682] The specific processing unit 290 transmits the result of the specific processing 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 a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the 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.

[0683] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0684] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0685] [Third embodiment]

[0686] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0687] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0688] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "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).

[0689] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. 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 microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0690] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0691] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0692] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0693] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0694] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[0695] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0696] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0697] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0698] The present invention relates to a system that uses a generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future, based on a current risk map and past weather data. Below, we will explain in natural language the mode for carrying out the present invention and the program processing. We will also provide specific examples.

[0699] Data retrieval by the server

[0700] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including, for example, rainfall data and weather chart data from the past 30 years. The server retrieves this data and stores it in a database.

[0701] Data preprocessing by the server

[0702] The server preprocesses the stored risk map data and weather data by imputing outliers and missing values, integrating the data, and converting it into a time series format. It also normalizes the data to make it suitable for machine learning models.

[0703] Server-based analysis using generative AI

[0704] The server uses the pre-processed data and uses generative artificial intelligence (e.g., LSTM models) to learn from past data, allowing it to extract temporal patterns and trends and predict disaster risk for several years into the future.

[0705] Server-based future prediction

[0706] The server uses the trained model to predict future disaster risks, such as flood risk over the next five years, and the results are stored in a database.

[0707] Visualization of prediction results on the device

[0708] The device receives the prediction results from the server and provides them to the user in a visual format, such as graphs and maps, in a way that is easy for the user to understand.

[0709] Specific examples

[0710] For example, suppose a user wants to know the risk of flooding. The server retrieves current hazard map data from the government's hazard map API, and also retrieves past rainfall data and weather chart data from the weather data API. These data are preprocessed on the server, and outliers and missing values ​​are filled in.

[0711] Next, using the preprocessed data, the server trains the data using a generative artificial intelligence (AI) model with an LSTM model. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database.

[0712] When a user accesses the system via a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps. For example, a graph showing flood risk forecasts for the next five years based on data from the past 30 years can be displayed in an easy-to-understand manner.

[0713] In this way, the system of the present invention effectively utilizes current risk maps and past weather data to perform highly accurate disaster predictions using generative artificial intelligence, and visualizes the results and provides them to users, allowing them to take appropriate measures against future disaster risks.

[0714] The processing flow will be explained below.

[0715] Step 1:

[0716] The server connects to external data sources to retrieve the latest risk maps and historical weather data, specifically accessing APIs from government and meteorological agencies to retrieve rainfall data and weather chart data for the past 30 years.

[0717] Step 2:

[0718] The server stores the acquired risk map data and weather data in a database, along with timestamps and related information corresponding to the acquired data.

[0719] Step 3:

[0720] The server preprocesses the data stored in the database. Specifically, it detects and complements outliers, estimates missing values ​​based on previous and subsequent data, and fills in missing values. It also organizes rainfall data and meteorological map data and converts them into a time series format.

[0721] Step 4:

[0722] The server normalizes the preprocessed data, calculating the mean and standard deviation of the data and standardizing each data point to prepare it for efficient learning by the generative AI.

[0723] Step 5:

[0724] The server uses the normalized data to train generative AI models such as LSTM models, and builds models to predict future disaster risks based on past time-series data.

[0725] Step 6:

[0726] The server uses the trained model to predict disaster risk over the next five years, generating data that quantifies monthly flood risk, for example, and storing the prediction results in a database.

[0727] Step 7:

[0728] A user accesses the system from a terminal and sends a request for disaster prediction results to the server through an application or web browser on the terminal.

[0729] Step 8:

[0730] The device receives the prediction results from the server, and the server sends the latest disaster prediction results stored in the database to the device.

[0731] Step 9:

[0732] The device visualizes the obtained prediction results, specifically using a visualization library such as Matplotlib to display the predicted disaster risk as graphs and maps.

[0733] Step 10:

[0734] Users can view the visualized forecast results on their devices, which allows them to understand future disaster risks and take appropriate measures.

[0735] Example 1

[0736] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0737] In recent years, the frequency and scale of natural disasters have increased, making it important to accurately predict future disaster risks. However, conventional systems are unable to effectively utilize past data, making it difficult to accurately predict disaster risks several years into the future. The present invention aims to solve this problem by providing a system that utilizes current risk maps and past weather data obtained from external data sources to achieve highly accurate disaster risk predictions.

[0738] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0739] In this invention, the server includes means for periodically acquiring a current risk map from an external data source, means for periodically acquiring past weather data from an external data source, means for preprocessing the current risk map and the past weather data by complementing, integrating, and normalizing outliers and missing values, means for training a generative artificial intelligence model using the preprocessed data to predict disaster risk several years from now, and means for visualizing the predicted disaster risk in the form of a graph or map. This makes it possible to effectively utilize past data to accurately predict future disaster risk and provide it to users visually.

[0740] A "risk map" is a map that geographically represents risks such as natural disasters and accidents.

[0741] "Weather data" refers to data that records past weather conditions such as rainfall, temperature, and wind speed on an hourly basis.

[0742] An "outlier" is a value that is significantly different from the rest of the data, and is a value that deviates significantly from the normal distribution.

[0743] "Missing values" refer to portions of a dataset where no values ​​are recorded.

[0744] "Preprocessing" refers to the process of preparing acquired data in a format suitable for analysis and model learning, and includes correcting outliers, filling in missing values, and integrating and normalizing data.

[0745] A "generative artificial intelligence model" is an algorithm that learns from input data and generates and predicts future data and results; specifically, it refers to neural network models such as LSTM (long short-term memory).

[0746] "Visualization" refers to displaying data and prediction results in a visual format such as a graph or map so that people can understand them intuitively.

[0747] A "data source" is an external resource or system from which data is obtained, such as an API or database.

[0748] "Integration" is the process of combining different datasets into a single dataset.

[0749] "Regularly" means to repeat at regular intervals.

[0750] "Normalization" is a process of converting each data point to fall within a certain range in order to unify the scale of the data.

[0751] MODE FOR CARRYING OUT THE INVENTION

[0752] The present invention provides a system that uses a generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future, based on a current risk map and past weather data. The following describes specific embodiments of the present invention.

[0753] Data retrieval by the server

[0754] The server periodically accesses external data sources to obtain the latest risk maps and historical weather data. For example, the server sends requests to government hazard map APIs and weather data APIs, analyzes the data received in response, and stores it in a database. This data includes rainfall data and weather map data from the past 30 years.

[0755] Data preprocessing by the server

[0756] The server preprocesses the acquired data. This includes imputing outliers and missing values, integrating risk map data with weather data, and normalizing the data. Specifically, anomalies are detected using an anomaly detection algorithm and imputed with appropriate values. Missing data is imputed with the mean or median value. The integrated dataset is then preprocessed using techniques such as z-score normalization.

[0757] Server-based analysis using generative AI

[0758] The server uses the preprocessed data to train a generative artificial intelligence (e.g., an LSTM model), which can learn patterns and trends from past data and predict future disaster risks. The server builds the LSTM model and splits the dataset into training data and validation data for training. After this process is complete, the server saves the trained model.

[0759] For example, the LSTM model predicts flood risk for the next five years based on rainfall data from the past 30 years. The server stores the prediction results in a database for future analysis and visualization.

[0760] Visualization of prediction results on the device

[0761] The terminal retrieves the forecast results from the server and provides them visually to the user. This includes displaying the forecast data in the form of graphs and maps. When a user accesses the system, the terminal sends a request to the server to retrieve the forecast results. Heat maps and graphs are then generated based on the retrieved data and displayed on the user interface. This allows the user to intuitively understand the flood risk forecast for the next five years based on data from the past 30 years.

[0762] Specific examples

[0763] For example, suppose a user wants to know the flood risk for the next five years. The server first obtains the latest risk map data from the government's hazard map API, and simultaneously collects rainfall data and weather chart data from the past 30 years from the weather data API. Next, the server preprocesses this data and uses appropriate algorithms to fill in outliers and missing values. The server then trains this data using an LSTM model to predict the flood risk for the next five years. After saving this forecast data in a database, when the user accesses the system using a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps.

[0764] Prompt Sentence Examples

[0765] "Predict flood risk over the next five years. Use rainfall data from the past 30 years and current risk map data. The generative AI used is an LSTM model."

[0766] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0767] Step 1: Data Acquisition

[0768] The server periodically retrieves current risk maps and historical weather data from external data sources (government hazard map APIs and weather data APIs). Specifically, the server sends requests to the API, analyzes the JSON or CSV format data received as a response, and stores it in a database. The input is the response data from the API, and the output is data saved in the database.

[0769] Step 2: Data Preprocessing

[0770] The server preprocesses the risk map data and weather data obtained from the database. First, it detects outliers and missing values ​​and fills them with appropriate values. Next, it integrates the risk map data and weather data and converts them into time-series data. At this time, it normalizes each data point using methods such as z-score normalization. The input is raw data obtained from the database, and the output is preprocessed, well-formatted data.

[0771] Specific working example:

[0772] An outlier detection algorithm is used to impute outliers.

[0773] Impute missing data with the mean or median.

[0774] Combine multiple datasets and transform them into time series data.

[0775] The data is scaled using a normalization algorithm.

[0776] Step 3: Model training

[0777] The server uses the preprocessed data to build a generative artificial intelligence model (such as an LSTM model) and trains it. It splits the data into training and validation data, and then trains and evaluates the model. The input is the preprocessed data, and the output is the trained LSTM model.

[0778] Specific working example:

[0779] Initialize the LSTM model and set the hyperparameters.

[0780] Split the dataset into training and validation sets.

[0781] A learning algorithm is used to train the model.

[0782] Validation data is used to evaluate the accuracy of the model and adjust the model as needed.

[0783] Step 4: Predict the future

[0784] The server uses a trained LSTM model to predict future disaster risks. The prediction period is multiple years, and risk assessments are performed for each region. The input is the latest risk map data, and the output is predicted future disaster risk data.

[0785] Specific working example:

[0786] Load a trained model.

[0787] Provide the latest risk map data as input to the model.

[0788] Predict future disaster risks using models.

[0789] The prediction results are saved in a database.

[0790] Step 5: Visualize the results

[0791] The terminal obtains the prediction results from the server and provides them visually to the user. Specifically, it generates heat maps and graphs and displays them on the user interface. The input is the prediction data obtained from the server, and the output is a display in the form of a graph or map.

[0792] Specific working example:

[0793] The terminal sends a request to the server to obtain the prediction data.

[0794] Generate graphs and heat maps based on the acquired data.

[0795] The generated graphs and heat maps are displayed in a user interface.

[0796] (Application example 1)

[0797] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0798] In recent years, the frequent occurrence of natural disasters has created a need for damage prediction and prevention. However, conventional disaster risk prediction systems do not fully utilize past data and current risk information, resulting in limitations in the accuracy of predictions and the implementation of disaster prevention measures. Furthermore, many systems only visualize disaster risks and do not adequately communicate them to users or present effective countermeasures. This presents a problem in that it is difficult for users to be aware of risks on a daily basis and take prompt and appropriate measures.

[0799] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0800] In this invention, the server includes means for acquiring a current risk map, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risk several years from now, means for visualizing the predicted disaster risk, means for sending a real-time alert to the user based on the predicted disaster risk, and means for the user to check the disaster risk around their place of residence or workplace on a map. This not only enables highly accurate disaster risk prediction, but also enables the user to be aware of disaster risk on a daily basis and take prompt and appropriate disaster prevention measures.

[0801] A "current risk map" is map information that shows the current risk of natural disasters in a specific area.

[0802] "Past weather data" refers to historical data on weather information such as rainfall and weather charts for a specific period of time.

[0803] "Preprocessing" is the process of correcting abnormal and missing values ​​in the acquired data, and integrating and normalizing it to prepare it into an analyzable format.

[0804] "Generative AI" is an artificial intelligence technology that learns from time-series data and predicts future disaster risks.

[0805] "Disaster risk several years from now" is forecast information that indicates the possibility of a natural disaster occurring in a specific area several years from now.

[0806] "Visualization" refers to the display of numerical data and forecast results in visual formats such as graphs and maps.

[0807] "Real-time alerts" is a function that instantly notifies users of warnings based on predicted disaster risks.

[0808] "Means of checking on a map" refers to a function that allows users to visually check disaster risks around their place of residence or workplace using a geographic information system.

[0809] This invention is a system that uses generative artificial intelligence to predict and visualize disaster risks several years into the future, using a current risk map and past weather data. Specific embodiments for implementing this invention are described below.

[0810] System Program

[0811] 1. Data Acquisition

[0812] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to retrieve the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years.

[0813] 2. Data Preprocessing

[0814] The server complements outliers and missing values ​​in the acquired risk map data and weather data, integrates the data, converts it into a time series format, and normalizes the data to make it suitable for machine learning models.

[0815] 3. Analysis using generative artificial intelligence

[0816] The server uses the pre-processed data and uses generative artificial intelligence (e.g., LSTM models) to learn from past data, allowing it to extract temporal patterns and trends and predict disaster risk for several years into the future.

[0817] 4. Predicting the future

[0818] The server uses the trained model to predict future disaster risks, for example, predicting flood risk over the next five years, and stores the results in a database.

[0819] 5. Visualization of prediction results

[0820] The device receives the prediction results from the server and provides them to the user in a visual format, such as graphs and maps, in a way that is easy for the user to understand.

[0821] 6. Real-time alerts

[0822] Based on the predicted disaster risk, the server sends users real-time alerts to urge them to take measures, and also provides a function that allows users to check the disaster risk around their residence or workplace on a map.

[0823] Hardware and software used

[0824] Hardware:

[0825] Server: A high-performance server handles the analysis, performing data acquisition, preprocessing, and predictive analysis.

[0826] Device: The device used by the user, such as a smartphone or tablet.

[0827] software:

[0828] API client: Uses government hazard map APIs and weather data APIs to obtain data.

[0829] Data preprocessing: Data integration, normalization, and outlier processing using the pandas library.

[0830] Generative AI: Analysis is performed using TensorFlow and LSTM models.

[0831] Visualization: Use matplotlib to visualize the prediction results.

[0832] Specific examples

[0833] For example, if a user requests the system to "predict flood risk for the next five years and display it by region," the server retrieves current hazard map data from the government's hazard map API and collects historical rainfall data and weather chart data from the weather data API. These data are preprocessed and trained using generative artificial intelligence (AI) with an LSTM model. The server then predicts flood risk for the next five years and stores the results in a database.

[0834] When a user accesses the system from their smartphone, the device retrieves the forecast results from the server and displays them in graphs and maps. For example, it provides a graph and map showing flood risk forecasts for the next five years, making it easy for users to understand.

[0835] In this way, the system of the present invention effectively utilizes current risk maps and past weather data to perform highly accurate disaster predictions using generative artificial intelligence, and visualizes the results and provides them to users, allowing them to take appropriate measures against future disaster risks.

[0836] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0837] Step 1: Data Acquisition

[0838] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to retrieve current risk map data and historical weather data. Inputs include API endpoints and parameters, and outputs JSON-formatted data. The server stores this data in a local database.

[0839] Step 2: Data Preprocessing

[0840] The server preprocesses the acquired risk map data and weather data. Specifically, it complements outliers and missing values, integrates the data, and converts it into a time series format. The acquired JSON data is the input, and the formatted data is the output. This processing uses the pandas library. The server detects outliers and complements them with appropriate values. It also merges data obtained from multiple data sources and formats it into time series data.

[0841] Step 3: Generative AI learning

[0842] The server uses the preprocessed data to learn from past data using a generative artificial intelligence (e.g., an LSTM model). The input is the preprocessed time series data, and the output is the trained model. The server uses the TensorFlow library to build the LSTM model and train the model on the time series data.

[0843] Step 4: Predict the future

[0844] The server uses the trained LSTM model to predict future disaster risks. For example, it predicts flood risks over the next five years. The inputs are the trained model and time series data, and the output is future disaster risk prediction data. The server stores this prediction data in a database.

[0845] Step 5: Visualize the prediction results

[0846] The terminal retrieves the prediction results from the server and provides them to the user in a visual format. Specifically, it displays the prediction data in the form of graphs and maps. The input is the prediction result data, and the output is visualized information. The terminal uses the matplotlib library and a geographic information system (GIS) to display the information in a visually easy-to-understand format.

[0847] Step 6: Real-time alerts

[0848] The server sends real-time alerts to users based on predicted disaster risks. The input is predicted disaster risk data, and the output is a notification message. The server sends push notifications to users' devices to provide immediate warnings. Based on the notifications received, users can immediately take appropriate disaster prevention measures.

[0849] This configuration allows users to receive highly accurate disaster prediction information based on current risk maps and past weather data, enabling them to take appropriate action in real time.

[0850] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0851] The present invention relates to a system that uses a current risk map and past weather data to predict disaster risks several years into the future using generative artificial intelligence, and also recognizes the user's emotions to adjust the way information is presented. Below, we will explain in natural language the modes for implementing the present invention and the program's processing. Specific examples will also be included.

[0852] Data retrieval by the server

[0853] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years. The server stores the obtained data in a database.

[0854] Data preprocessing by the server

[0855] The server preprocesses the stored risk map data and weather data. Specifically, it detects and complements outliers, and fills in missing values ​​by estimating them from previous and subsequent data. It also organizes rainfall data and weather map data and converts them into a time series format. It also normalizes the data and formats it in a way that is suitable for machine learning models.

[0856] Server-based analysis using generative AI

[0857] The server uses the preprocessed data to train generative AI models such as LSTM models, which build models to predict future disaster risks based on past time-series data.

[0858] Server-based future prediction

[0859] The server uses the trained model to predict disaster risk over the next five years, generating data that quantifies monthly flood risk, for example, and storing the prediction results in a database.

[0860] Emotion engine that recognizes user emotions

[0861] The server is equipped with an emotion engine that recognizes the user's emotions from their inputs, actions, images, and voice. For example, it analyzes the text and voice entered when the user accesses the system and determines the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0862] Server-adjusted presentation of information

[0863] After the emotion engine recognizes the user's emotion, the server adjusts the way disaster risk information is presented based on the recognized emotion. For example, if the user expresses strong anxiety, the server will present disaster risk information in a more reassuring tone. It will also alleviate the user's anxiety by displaying specific evacuation locations and countermeasures.

[0864] Visualization of prediction results on the device

[0865] The device retrieves the forecast results from the server and provides them to the user in a visual format. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, providing a design that is easy for users to understand.

[0866] Device and user interaction

[0867] Users can understand future disaster risks by viewing the visualized forecast results on their devices. Furthermore, based on the user's input and feedback, the emotion engine continuously monitors the user's emotional state and dynamically adjusts the way information is presented accordingly.

[0868] Specific examples

[0869] For example, suppose a user wants to know the risk of flooding. The server retrieves current hazard map data from the government's hazard map API, and also retrieves past rainfall data and weather chart data from the weather data API. These data are preprocessed on the server, and outliers and missing values ​​are filled in.

[0870] Next, the server uses the preprocessed data to train the data using a generative artificial intelligence (AI) model with an LSTM model. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database.

[0871] When a user accesses the system via a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps. For example, a graph showing flood risk forecasts for the next five years based on data from the past 30 years can be displayed in an easy-to-understand manner.

[0872] Furthermore, the emotion engine recognizes the user's emotional state and presents information that provides a sense of security if the user feels anxious. In this way, the system of the present invention effectively utilizes current risk maps and past weather data, and uses generative artificial intelligence and the emotion engine to provide highly accurate disaster predictions and appropriate information. This allows users to take more effective measures against future disaster risks.

[0873] The processing flow will be explained below.

[0874] Step 1:

[0875] The server connects to external data sources to retrieve the latest risk maps and historical weather data. Specifically, it accesses the government's hazard map API to retrieve risk map data, and weather data API to retrieve rainfall data and weather chart data for the past 30 years.

[0876] Step 2:

[0877] The server stores the acquired risk map data and weather data in a database, along with the corresponding timestamp and related meta information.

[0878] Step 3:

[0879] The server preprocesses the risk map data and weather data stored in the database. Specifically, it detects outliers and interpolates them based on the preceding and following data. Missing values ​​are also interpolated based on the preceding and following data. The server also converts the data into a time series format and normalizes it.

[0880] Step 4:

[0881] The server then uses the preprocessed data to train a generative artificial intelligence model (such as an LSTM model). Using past time-series data, the server builds a model to predict future disaster risk. This model quantifies flood risk and other natural disaster risks.

[0882] Step 5:

[0883] The server uses the trained model to predict disaster risk for the next five years. Specifically, it predicts the next five years on a monthly basis, quantifying the flood risk for each month and storing the results in a database.

[0884] Step 6:

[0885] Users access the system using a terminal and request prediction results through a web browser or application.

[0886] Step 7:

[0887] The server receives the user's request, retrieves the latest disaster risk prediction results from the database, and sends the results to the terminal.

[0888] Step 8:

[0889] The device visualizes the disaster risk prediction results it receives. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, making it easy for users to understand visually.

[0890] Step 9:

[0891] The emotion engine recognizes the user's emotions based on their input and actions. For example, it analyzes the text, voice, and facial expressions entered by the user to determine whether the user is feeling anxious or relieved.

[0892] Step 10:

[0893] The server then adjusts the presentation of disaster risk information based on the user's perceived emotions: for example, if the user feels anxious, it will present the information in a more reassuring tone and provide additional information on precautions and evacuation locations.

[0894] Step 11:

[0895] The device then revisits the adjusted information and provides it to the user, allowing them to check disaster risk information with greater peace of mind and take appropriate measures.

[0896] Example 2

[0897] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0898] Current disaster risk prediction systems predict disaster risk based on past data and current conditions, but no systems provide information that takes user emotions into consideration. This makes users more likely to feel anxious and makes it difficult for them to take appropriate disaster prevention actions. Furthermore, there are problems with insufficient data processing and imputation of outliers and missing values ​​to accurately predict disaster risk.

[0899] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0900] In this invention, the server includes means for acquiring a current risk map from an external data source, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now, means for recognizing a user's emotions and adjusting the method of presenting information, and means for visualizing the predicted disaster risks. This enables highly accurate disaster risk predictions and appropriate information provision while taking user emotions into consideration.

[0901] "External Data Source" refers to a source of data from an external source. Examples include government hazard map APIs and weather data APIs.

[0902] "Current risk maps" refer to map data showing disaster risks at the present time. These usually include risk information for earthquakes, floods, landslides, etc.

[0903] "Historical Weather Data" refers to datasets based on past weather information, including precipitation, weather maps, and temperatures.

[0904] "Preprocessing" refers to the process of cleaning, completing, and transforming data into a format suitable for machine learning models, including imputing outliers, filling in missing values, converting to time series data, and normalizing.

[0905] "Generative AI" refers to AI technology that uses past data to make future predictions, including deep learning models such as LSTM (Long Short-Term Memory).

[0906] "Disaster risk" is an indicator that shows the possibility of natural disasters such as earthquakes, floods, and landslides occurring.

[0907] "Emotion recognition" refers to the technology of analyzing and determining a user's emotional state (e.g., anxiety, relief, surprise, etc.) based on input data such as text, voice, and images.

[0908] "Adjusting the way information is presented" refers to the process of changing the means and content of information presentation based on the user's recognized emotions. For example, for a user who is feeling anxious, adding a reassuring message or specific countermeasures.

[0909] "Visualization" refers to the technology of displaying predicted disaster risks in visual formats such as graphs and maps, making it easier for users to understand the prediction results.

[0910] An "outlier" is a value that is extremely different from the other data points in a data set.

[0911] "Missing values" refers to the absence of values ​​in a dataset.

[0912] The present invention is a system that uses a current risk map and past weather data to predict disaster risks several years into the future using generative artificial intelligence, and further recognizes the user's emotions and adjusts the way information is presented. Specific embodiments for implementing the present invention are described below.

[0913] Data retrieval by the server

[0914] The server connects to external data sources (e.g., government hazard map APIs and weather data APIs) and periodically retrieves current risk map data and historical weather data (e.g., rainfall data and weather chart data from the past 30 years). The server stores this data in a database (e.g., PostgreSQL) and manages it securely.

[0915] Data preprocessing by the server

[0916] The server preprocesses the stored data. Specifically, it uses Python's Pandas library to detect and impute outliers. Missing values ​​are estimated and filled using linear interpolation and moving averages. Furthermore, rainfall data and weather map data are converted into time series data and normalized, allowing machine learning models to learn from the data efficiently.

[0917] Server-based analysis using generative AI

[0918] The server uses the preprocessed data to build and train a generative artificial intelligence model (e.g., an LSTM model) using a deep learning framework such as TensorFlow or Keras. The model is optimized over multiple epochs using a dataset split into training and test data.

[0919] Server-based future prediction

[0920] The server uses the trained model to predict disaster risk for the next five years. Specifically, it generates data quantifying monthly flood risk and stores the prediction results in a database. For example, it provides input to the generative AI model based on prompts such as, "Please predict the flood risk for the next five years."

[0921] Emotion engine that recognizes user emotions

[0922] The server is equipped with an emotion engine that recognizes the user's emotions from their input, actions, or image and audio data. Specifically, it uses Python's NLTK library and OpenCV to analyze text and audio and determine the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0923] Server-adjusted presentation of information

[0924] After the emotion engine recognizes the user's emotions, the server adjusts the way disaster risk information is presented based on the user's emotions. For example, if the user expresses strong anxiety, the server will provide information in a reassuring tone and additionally display specific evacuation locations and countermeasures to alleviate the user's anxiety.

[0925] Visualization of prediction results on the device

[0926] The terminal uses visualization libraries such as Matplotlib to display the forecast results obtained from the server in graphs and maps. For example, it can visually display flood risk forecast results for the next five years, providing a screen that is easy for users to understand.

[0927] Device and user interaction

[0928] Users can view the visualized forecast results on their devices to understand future disaster risks. Furthermore, the emotion engine continuously monitors the user's emotional state based on their input and feedback, and dynamically adjusts the way information is presented accordingly.

[0929] Specific examples

[0930] For example, if a user wants to know the "flood risk for the next five years," the server retrieves the latest hazard map data from the government's hazard map API and downloads historical rainfall data from the weather data API. These data are then interpolated for outliers and missing values ​​and converted into time-series data.

[0931] The data is then trained using an LSTM model to predict flood risk over the next five years. The results are stored in a database, and when users access the system from their devices, the forecast results are displayed in graphs and maps. The emotion engine recognizes the user's emotions and adds reassuring messages such as "There is a high risk of flooding, but there are many evacuation sites in this area and measures are in place."

[0932] This allows the system to provide highly accurate disaster risk predictions and appropriate information while taking into consideration the user's feelings.

[0933] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0934] Step 1:

[0935] The server connects to external data sources (for example, government hazard map APIs or weather data APIs) and periodically retrieves current risk map data and weather data from the past 30 years (rainfall data and weather chart data). The input is data provided through API requests, and the output is raw data that is stored in the server's database. Specifically, it uses the Python Requests library to access the API and retrieve and store the data.

[0936] Step 2:

[0937] The server preprocesses the stored data. The input is the data already retrieved in the database, and the output is the preprocessed data. Specifically, it uses the Python Pandas library to detect and impute outliers. It also estimates missing values ​​using linear interpolation and moving averages, converts them into time series data, and normalizes the data to make it suitable for machine learning models.

[0938] Step 3:

[0939] The server uses the preprocessed data to train a generative AI model, such as an LSTM model. The input is the preprocessed time series data, and the output is a trained LSTM model. Specifically, the model is defined using TensorFlow or Keras and trained through multiple epochs.

[0940] Step 4:

[0941] The server uses the trained model to predict disaster risk for the next five years. The input is the trained LSTM model, the latest risk map, and weather data, and the output is quantified disaster risk prediction data. For example, the server performs predictions based on prompt statements such as, "Please predict the flood risk for the next five years."

[0942] Step 5:

[0943] The server uses an emotion engine to recognize the user's emotional state. The input is the user's text, voice, and image data, and the output is the user's emotional state (anxiety, relief, surprise, etc.). Specifically, the server analyzes the text and voice data using Python's NLTK library and OpenCV.

[0944] Step 6:

[0945] The server adjusts the way disaster risk information is presented based on the recognized emotion. The input is the user's emotional state and disaster risk prediction data, and the output is customized information presentation. For example, if the user expresses anxiety, it displays a reassuring message along with specific evacuation locations and countermeasures.

[0946] Step 7:

[0947] The terminal obtains the forecast results from the server and provides them to the user using a visualization library such as Matplotlib. The input is disaster risk forecast data, and the output is visual information in the form of graphs and maps. Specifically, the terminal generates graphs and maps and displays them in a format that is easy for the user to understand.

[0948] Step 8:

[0949] The user checks and understands the visualized forecast results on their device. The input is the visualized disaster risk forecast data, and the output is the user's feedback and changes in emotional state. Based on the user's feedback, the emotion engine continuously monitors changes in emotions and dynamically adjusts the way information is presented.

[0950] In this way, the system takes into consideration the user's feelings while achieving highly accurate disaster risk prediction and providing appropriate information.

[0951] (Application example 2)

[0952] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0953] Currently, many disaster risk prediction systems simply predict future risks based on past data and do not take into account the user's emotional state or how that information is presented. This can lead to users feeling unnecessarily anxious or, conversely, becoming indifferent. Furthermore, particularly in electronic payment services, it is important to increase the security of transactions by providing appropriate disaster risk information. The objective of this invention is to provide a system that recognizes the user's emotions and presents appropriate disaster risk information based on those emotions, allowing users to use electronic payment services with peace of mind.

[0954] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0955] In this invention, the server includes means for acquiring a current risk map, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now, means for visualizing the predicted disaster risks, means for determining the emotional state of a user using an emotion engine, and means for adjusting the presentation method of disaster risk information based on the emotional state, thereby enabling the user to receive disaster risk information that is appropriately adjusted to match their emotional state.

[0956] A "current risk map" is data provided by governments and public institutions that shows the current risk of natural disasters in map form.

[0957] "Past weather data" refers to observational data on weather over a certain period of time in the past, including rainfall, temperature, and air pressure.

[0958] "Preprocessing" refers to processes such as organizing data, correcting outliers, and estimating missing values ​​in order to prepare raw data in a format suitable for machine learning models.

[0959] "Generative AI" is an AI technology that has the ability to generate new information from input data, and is often used to predict time series data.

[0960] An "emotion engine" is a technology that recognizes and classifies a user's emotional state based on their input, actions, images, voice, etc.

[0961] "Visualization" is a method of presenting data and prediction results in a visually easy-to-understand format, such as a graph or map.

[0962] "Adjusting presentation method" is a technology that provides information more appropriately by changing the method and content of information transmission depending on the user's current emotional state.

[0963] The present invention provides a system for predicting disaster risks in an area where a user resides and providing information that gives the user a sense of security based on the predicted disaster risks. Specific embodiments and their program processing will be described in detail below.

[0964] Data Acquisition Method

[0965] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years. The server stores the obtained data in a database.

[0966] Data preprocessing methods

[0967] The server preprocesses the stored risk map data and weather data. Specifically, it detects and complements outliers, and fills in missing values ​​by estimating them from previous and subsequent data. It also organizes rainfall data and weather map data and converts them into a time series format. It also normalizes the data and formats it in a way that is suitable for machine learning models.

[0968] A means of predictive analysis using generative artificial intelligence

[0969] The server uses the preprocessed data to train a generative artificial intelligence such as an LSTM model. Based on past time series data, it builds a model to predict future disaster risk. The server then uses the trained model to predict disaster risk over the next five years. For example, it generates data that quantifies monthly flood risk and stores the prediction results in a database.

[0970] A method for determining user emotions using an emotion engine

[0971] The server is equipped with an emotion engine that recognizes the user's emotions from their input, actions, images, and voice. For example, it analyzes the text and voice entered when the user accesses the system and determines the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[0972] A means of adjusting information presentation

[0973] After the emotion engine recognizes the user's emotion, the server adjusts the way disaster risk information is presented based on the recognized emotion. For example, if the user expresses strong anxiety, the server will present disaster risk information in a more reassuring tone. It will also alleviate the user's anxiety by displaying specific evacuation locations and countermeasures.

[0974] Visualization of prediction results

[0975] The device retrieves the forecast results from the server and provides them to the user in a visual format. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, providing a design that is easy for users to understand.

[0976] Device and user interaction

[0977] Users can understand future disaster risks by viewing the visualized forecast results on their devices. Furthermore, based on the user's input and feedback, the emotion engine continuously monitors the user's emotional state and dynamically adjusts the way information is presented accordingly.

[0978] Examples of concrete examples and prompts

[0979] For example, suppose a user wants to know about flood risk. The server obtains current hazard map data from the government's hazard map API, and also obtains historical rainfall and weather chart data from a weather data API. These data are preprocessed on the server, and outliers and missing values ​​are interpolated. Next, using the preprocessed data, the server uses an LSTM model to train the data using generative artificial intelligence. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database. When the user then accesses the system on their device, the device retrieves the prediction results from the server and displays them as graphs or maps. For example, a graph predicting flood risk for the next five years based on data from the past 30 years can be displayed, providing a clear picture to the user. Furthermore, an emotion engine recognizes the user's emotional state, and if the user is anxious, information that provides reassurance is presented.

[0980] Example prompt for a generative AI model:

[0981] "Train an LSTM model to predict flood risk for the next five years based on rainfall data and meteorological map data from the past 30 years. Then, recognize user-input emotions and generate messages to provide reassurance and risk information if the user is anxious."

[0982] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0983] Step 1:

[0984] The server accesses the government's hazard map API and weather data API to obtain current risk map data and weather data for the past 30 years. The data obtained in this step is stored in a database as raw data.

[0985] Input: Government hazard map API and weather data API

[0986] Output: Current risk map data and weather data from the past 30 years

[0987] Step 2:

[0988] The server preprocesses the raw data stored in the database, detecting outliers and filling in missing values ​​by estimating them from surrounding data. It also converts the data into a time series format and normalizes it, making it suitable for machine learning models.

[0989] Input: Raw data

[0990] Output: Preprocessed data

[0991] Step 3:

[0992] The server uses the preprocessed data to build and train an LSTM model. The trained LSTM model is used to predict disaster risk for several years into the future from past data. The trained model then predicts monthly disaster risk for the next five years and stores the prediction results in a database.

[0993] Input: Preprocessed data

[0994] Output: Predicted disaster risk data

[0995] Step 4:

[0996] The server uses an emotion engine to analyze the user's input text and voice data, thereby determining the user's emotional state (e.g., anxiety, relief, surprise, etc.). The emotion recognition results are also stored in a database.

[0997] Input: User-entered text and voice data

[0998] Output: User's emotional state data

[0999] Step 5:

[1000] The server adjusts the way it presents disaster risk information based on the user's emotional state, as determined by the emotion engine. Specifically, if the user expresses anxiety, the server presents the information in a reassuring tone and additionally displays specific evacuation locations and countermeasures.

[1001] Input: User emotional state data and predicted disaster risk data

[1002] Output: Tailored disaster risk information

[1003] Step 6:

[1004] The device retrieves the prediction results from the server and presents them to the user in a visually understandable format, such as graphs or maps, using visualization libraries like Matplotlib, and adjusts the tone of the message to match the user's emotional state, if necessary.

[1005] Input: Tailored disaster risk information

[1006] Output: Disaster risk information that is visually easy for users to understand

[1007] Step 7:

[1008] Users can view the visualized forecast results on their devices to understand future disaster risks. Furthermore, the emotion engine continuously monitors the user's emotional state based on their input and feedback, dynamically adjusting how information is presented.

[1009] Input: Visualized disaster risk information, user feedback

[1010] Output: Continuously updated emotional state data and adjusted risk information

[1011] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the 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.

[1012] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1013] 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 the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1014] [Fourth embodiment]

[1015] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1016] 7, a 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.

[1017] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "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).

[1018] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. 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 microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1019] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1020] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1021] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1022] The control object 443 includes a display device, LEDs in the eyes, and motors for driving 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 emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1023] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1024] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[1025] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1026] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1027] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1028] The present invention relates to a system that uses a generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future, based on a current risk map and past weather data. Below, we will explain in natural language the mode for carrying out the present invention and the program processing. We will also provide specific examples.

[1029] Data retrieval by the server

[1030] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including, for example, rainfall data and weather chart data from the past 30 years. The server retrieves this data and stores it in a database.

[1031] Data preprocessing by the server

[1032] The server preprocesses the stored risk map data and weather data by imputing outliers and missing values, integrating the data, and converting it into a time series format. It also normalizes the data to make it suitable for machine learning models.

[1033] Server-based analysis using generative AI

[1034] The server uses the pre-processed data and uses generative artificial intelligence (e.g., LSTM models) to learn from past data, allowing it to extract temporal patterns and trends and predict disaster risk for several years into the future.

[1035] Server-based future prediction

[1036] The server uses the trained model to predict future disaster risks, such as flood risk over the next five years, and the results are stored in a database.

[1037] Visualization of prediction results on the device

[1038] The device receives the prediction results from the server and provides them to the user in a visual format, such as graphs and maps, in a way that is easy for the user to understand.

[1039] Specific examples

[1040] For example, suppose a user wants to know the risk of flooding. The server retrieves current hazard map data from the government's hazard map API, and also retrieves past rainfall data and weather chart data from the weather data API. These data are preprocessed on the server, and outliers and missing values ​​are filled in.

[1041] Next, using the preprocessed data, the server trains the data using a generative artificial intelligence (AI) model with an LSTM model. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database.

[1042] When a user accesses the system via a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps. For example, a graph showing flood risk forecasts for the next five years based on data from the past 30 years can be displayed in an easy-to-understand manner.

[1043] In this way, the system of the present invention effectively utilizes current risk maps and past weather data to perform highly accurate disaster predictions using generative artificial intelligence, and visualizes the results and provides them to users, allowing them to take appropriate measures against future disaster risks.

[1044] The processing flow will be explained below.

[1045] Step 1:

[1046] The server connects to external data sources to retrieve the latest risk maps and historical weather data, specifically accessing APIs from government and meteorological agencies to retrieve rainfall data and weather chart data for the past 30 years.

[1047] Step 2:

[1048] The server stores the acquired risk map data and weather data in a database, along with timestamps and related information corresponding to the acquired data.

[1049] Step 3:

[1050] The server preprocesses the data stored in the database. Specifically, it detects and complements outliers, estimates missing values ​​based on previous and subsequent data, and fills in missing values. It also organizes rainfall data and meteorological map data and converts them into a time series format.

[1051] Step 4:

[1052] The server normalizes the preprocessed data, calculating the mean and standard deviation of the data and standardizing each data point to prepare it for efficient learning by the generative AI.

[1053] Step 5:

[1054] The server uses the normalized data to train generative AI models such as LSTM models, and builds models to predict future disaster risks based on past time-series data.

[1055] Step 6:

[1056] The server uses the trained model to predict disaster risk over the next five years, generating data that quantifies monthly flood risk, for example, and storing the prediction results in a database.

[1057] Step 7:

[1058] A user accesses the system from a terminal and sends a request for disaster prediction results to the server through an application or web browser on the terminal.

[1059] Step 8:

[1060] The device receives the prediction results from the server, and the server sends the latest disaster prediction results stored in the database to the device.

[1061] Step 9:

[1062] The device visualizes the obtained prediction results, specifically using a visualization library such as Matplotlib to display the predicted disaster risk as graphs and maps.

[1063] Step 10:

[1064] Users can view the visualized forecast results on their devices, which allows them to understand future disaster risks and take appropriate measures.

[1065] Example 1

[1066] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1067] In recent years, the frequency and scale of natural disasters have increased, making it important to accurately predict future disaster risks. However, conventional systems are unable to effectively utilize past data, making it difficult to accurately predict disaster risks several years into the future. The present invention aims to solve this problem by providing a system that utilizes current risk maps and past weather data obtained from external data sources to achieve highly accurate disaster risk predictions.

[1068] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1069] In this invention, the server includes means for periodically acquiring a current risk map from an external data source, means for periodically acquiring past weather data from an external data source, means for preprocessing the current risk map and the past weather data by complementing, integrating, and normalizing outliers and missing values, means for training a generative artificial intelligence model using the preprocessed data to predict disaster risk several years from now, and means for visualizing the predicted disaster risk in the form of a graph or map. This makes it possible to effectively utilize past data to accurately predict future disaster risk and provide it to users visually.

[1070] A "risk map" is a map that geographically represents risks such as natural disasters and accidents.

[1071] "Weather data" refers to data that records past weather conditions such as rainfall, temperature, and wind speed on an hourly basis.

[1072] An "outlier" is a value that is significantly different from the rest of the data, and is a value that deviates significantly from the normal distribution.

[1073] "Missing values" refer to portions of a dataset where no values ​​are recorded.

[1074] "Preprocessing" refers to the process of preparing acquired data in a format suitable for analysis and model learning, and includes correcting outliers, filling in missing values, and integrating and normalizing data.

[1075] A "generative artificial intelligence model" is an algorithm that learns from input data and generates and predicts future data and results; specifically, it refers to neural network models such as LSTM (long short-term memory).

[1076] "Visualization" refers to displaying data and prediction results in a visual format such as a graph or map so that people can understand them intuitively.

[1077] A "data source" is an external resource or system from which data is obtained, such as an API or database.

[1078] "Integration" is the process of combining different datasets into a single dataset.

[1079] "Regularly" means to repeat at regular intervals.

[1080] "Normalization" is a process of converting each data point to fall within a certain range in order to unify the scale of the data.

[1081] MODE FOR CARRYING OUT THE INVENTION

[1082] The present invention provides a system that uses a generative artificial intelligence (AI) to predict and visualize disaster risks several years into the future, based on a current risk map and past weather data. The following describes specific embodiments of the present invention.

[1083] Data retrieval by the server

[1084] The server periodically accesses external data sources to obtain the latest risk maps and historical weather data. For example, the server sends requests to government hazard map APIs and weather data APIs, analyzes the data received in response, and stores it in a database. This data includes rainfall data and weather map data from the past 30 years.

[1085] Data preprocessing by the server

[1086] The server preprocesses the acquired data. This includes imputing outliers and missing values, integrating risk map data with weather data, and normalizing the data. Specifically, anomalies are detected using an anomaly detection algorithm and imputed with appropriate values. Missing data is imputed with the mean or median value. The integrated dataset is then preprocessed using techniques such as z-score normalization.

[1087] Server-based analysis using generative AI

[1088] The server uses the preprocessed data to train a generative artificial intelligence (e.g., an LSTM model), which can learn patterns and trends from past data and predict future disaster risks. The server builds the LSTM model and splits the dataset into training data and validation data for training. After this process is complete, the server saves the trained model.

[1089] For example, the LSTM model predicts flood risk for the next five years based on rainfall data from the past 30 years. The server stores the prediction results in a database for future analysis and visualization.

[1090] Visualization of prediction results on the device

[1091] The terminal retrieves the forecast results from the server and provides them visually to the user. This includes displaying the forecast data in the form of graphs and maps. When a user accesses the system, the terminal sends a request to the server to retrieve the forecast results. Heat maps and graphs are then generated based on the retrieved data and displayed on the user interface. This allows the user to intuitively understand the flood risk forecast for the next five years based on data from the past 30 years.

[1092] Specific examples

[1093] For example, suppose a user wants to know the flood risk for the next five years. The server first obtains the latest risk map data from the government's hazard map API, and simultaneously collects rainfall data and weather chart data from the past 30 years from the weather data API. Next, the server preprocesses this data and uses appropriate algorithms to fill in outliers and missing values. The server then trains this data using an LSTM model to predict the flood risk for the next five years. After saving this forecast data in a database, when the user accesses the system using a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps.

[1094] Prompt Sentence Examples

[1095] "Predict flood risk over the next five years. Use rainfall data from the past 30 years and current risk map data. The generative AI used is an LSTM model."

[1096] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1097] Step 1: Data Acquisition

[1098] The server periodically retrieves current risk maps and historical weather data from external data sources (government hazard map APIs and weather data APIs). Specifically, the server sends requests to the API, analyzes the JSON or CSV format data received as a response, and stores it in a database. The input is the response data from the API, and the output is data saved in the database.

[1099] Step 2: Data Preprocessing

[1100] The server preprocesses the risk map data and weather data obtained from the database. First, it detects outliers and missing values ​​and fills them with appropriate values. Next, it integrates the risk map data and weather data and converts them into time-series data. At this time, it normalizes each data point using methods such as z-score normalization. The input is raw data obtained from the database, and the output is preprocessed, well-formatted data.

[1101] Specific working example:

[1102] An outlier detection algorithm is used to impute outliers.

[1103] Impute missing data with the mean or median.

[1104] Combine multiple datasets and transform them into time series data.

[1105] The data is scaled using a normalization algorithm.

[1106] Step 3: Model training

[1107] The server uses the preprocessed data to build a generative artificial intelligence model (such as an LSTM model) and trains it. It splits the data into training and validation data, and then trains and evaluates the model. The input is the preprocessed data, and the output is the trained LSTM model.

[1108] Specific working example:

[1109] Initialize the LSTM model and set the hyperparameters.

[1110] Split the dataset into training and validation sets.

[1111] A learning algorithm is used to train the model.

[1112] Validation data is used to evaluate the accuracy of the model and adjust the model as needed.

[1113] Step 4: Predict the future

[1114] The server uses a trained LSTM model to predict future disaster risks. The prediction period is multiple years, and risk assessments are performed for each region. The input is the latest risk map data, and the output is predicted future disaster risk data.

[1115] Specific working example:

[1116] Load a trained model.

[1117] Provide the latest risk map data as input to the model.

[1118] Predict future disaster risks using models.

[1119] The prediction results are saved in a database.

[1120] Step 5: Visualize the results

[1121] The terminal obtains the prediction results from the server and provides them visually to the user. Specifically, it generates heat maps and graphs and displays them on the user interface. The input is the prediction data obtained from the server, and the output is a display in the form of a graph or map.

[1122] Specific working example:

[1123] The terminal sends a request to the server to obtain the prediction data.

[1124] Generate graphs and heat maps based on the acquired data.

[1125] The generated graphs and heat maps are displayed in a user interface.

[1126] (Application example 1)

[1127] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1128] In recent years, the frequent occurrence of natural disasters has created a need for damage prediction and prevention. However, conventional disaster risk prediction systems do not fully utilize past data and current risk information, resulting in limitations in the accuracy of predictions and the implementation of disaster prevention measures. Furthermore, many systems only visualize disaster risks and do not adequately communicate them to users or present effective countermeasures. This presents a problem in that it is difficult for users to be aware of risks on a daily basis and take prompt and appropriate measures.

[1129] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1130] In this invention, the server includes means for acquiring a current risk map, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risk several years from now, means for visualizing the predicted disaster risk, means for sending a real-time alert to the user based on the predicted disaster risk, and means for the user to check the disaster risk around their place of residence or workplace on a map. This not only enables highly accurate disaster risk prediction, but also enables the user to be aware of disaster risk on a daily basis and take prompt and appropriate disaster prevention measures.

[1131] A "current risk map" is map information that shows the current risk of natural disasters in a specific area.

[1132] "Past weather data" refers to historical data on weather information such as rainfall and weather charts for a specific period of time.

[1133] "Preprocessing" is the process of correcting abnormal and missing values ​​in the acquired data, and integrating and normalizing it to prepare it into an analyzable format.

[1134] "Generative AI" is an artificial intelligence technology that learns from time-series data and predicts future disaster risks.

[1135] "Disaster risk several years from now" is forecast information that indicates the possibility of a natural disaster occurring in a specific area several years from now.

[1136] "Visualization" refers to the display of numerical data and forecast results in visual formats such as graphs and maps.

[1137] "Real-time alerts" is a function that instantly notifies users of warnings based on predicted disaster risks.

[1138] "Means of checking on a map" refers to a function that allows users to visually check disaster risks around their place of residence or workplace using a geographic information system.

[1139] This invention is a system that uses generative artificial intelligence to predict and visualize disaster risks several years into the future, using a current risk map and past weather data. Specific embodiments for implementing this invention are described below.

[1140] System Program

[1141] 1. Data Acquisition

[1142] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to retrieve the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years.

[1143] 2. Data Preprocessing

[1144] The server complements outliers and missing values ​​in the acquired risk map data and weather data, integrates the data, converts it into a time series format, and normalizes the data to make it suitable for machine learning models.

[1145] 3. Analysis using generative artificial intelligence

[1146] The server uses the pre-processed data and uses generative artificial intelligence (e.g., LSTM models) to learn from past data, allowing it to extract temporal patterns and trends and predict disaster risk for several years into the future.

[1147] 4. Predicting the future

[1148] The server uses the trained model to predict future disaster risks, for example, predicting flood risk over the next five years, and stores the results in a database.

[1149] 5. Visualization of prediction results

[1150] The device receives the prediction results from the server and provides them to the user in a visual format, such as graphs and maps, in a way that is easy for the user to understand.

[1151] 6. Real-time alerts

[1152] Based on the predicted disaster risk, the server sends users real-time alerts to urge them to take measures, and also provides a function that allows users to check the disaster risk around their residence or workplace on a map.

[1153] Hardware and software used

[1154] Hardware:

[1155] Server: A high-performance server handles the analysis, performing data acquisition, preprocessing, and predictive analysis.

[1156] Device: The device used by the user, such as a smartphone or tablet.

[1157] software:

[1158] API client: Uses government hazard map APIs and weather data APIs to obtain data.

[1159] Data preprocessing: Data integration, normalization, and outlier processing using the pandas library.

[1160] Generative AI: Analysis is performed using TensorFlow and LSTM models.

[1161] Visualization: Use matplotlib to visualize the prediction results.

[1162] Specific examples

[1163] For example, if a user requests the system to "predict flood risk for the next five years and display it by region," the server retrieves current hazard map data from the government's hazard map API and collects historical rainfall data and weather chart data from the weather data API. These data are preprocessed and trained using generative artificial intelligence (AI) with an LSTM model. The server then predicts flood risk for the next five years and stores the results in a database.

[1164] When a user accesses the system from their smartphone, the device retrieves the forecast results from the server and displays them in graphs and maps. For example, it provides a graph and map showing flood risk forecasts for the next five years, making it easy for users to understand.

[1165] In this way, the system of the present invention effectively utilizes current risk maps and past weather data to perform highly accurate disaster predictions using generative artificial intelligence, and visualizes the results and provides them to users, allowing them to take appropriate measures against future disaster risks.

[1166] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1167] Step 1: Data Acquisition

[1168] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to retrieve current risk map data and historical weather data. Inputs include API endpoints and parameters, and outputs JSON-formatted data. The server stores this data in a local database.

[1169] Step 2: Data Preprocessing

[1170] The server preprocesses the acquired risk map data and weather data. Specifically, it complements outliers and missing values, integrates the data, and converts it into a time series format. The acquired JSON data is the input, and the formatted data is the output. This processing uses the pandas library. The server detects outliers and complements them with appropriate values. It also merges data obtained from multiple data sources and formats it into time series data.

[1171] Step 3: Generative AI learning

[1172] The server uses the preprocessed data to learn from past data using a generative artificial intelligence (e.g., an LSTM model). The input is the preprocessed time series data, and the output is the trained model. The server uses the TensorFlow library to build the LSTM model and train the model on the time series data.

[1173] Step 4: Predict the future

[1174] The server uses the trained LSTM model to predict future disaster risks. For example, it predicts flood risks over the next five years. The inputs are the trained model and time series data, and the output is future disaster risk prediction data. The server stores this prediction data in a database.

[1175] Step 5: Visualize the prediction results

[1176] The terminal retrieves the prediction results from the server and provides them to the user in a visual format. Specifically, it displays the prediction data in the form of graphs and maps. The input is the prediction result data, and the output is visualized information. The terminal uses the matplotlib library and a geographic information system (GIS) to display the information in a visually easy-to-understand format.

[1177] Step 6: Real-time alerts

[1178] The server sends real-time alerts to users based on predicted disaster risks. The input is predicted disaster risk data, and the output is a notification message. The server sends push notifications to users' devices to provide immediate warnings. Based on the notifications received, users can immediately take appropriate disaster prevention measures.

[1179] This configuration allows users to receive highly accurate disaster prediction information based on current risk maps and past weather data, enabling them to take appropriate action in real time.

[1180] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1181] The present invention relates to a system that uses a current risk map and past weather data to predict disaster risks several years into the future using generative artificial intelligence, and also recognizes the user's emotions to adjust the way information is presented. Below, we will explain in natural language the modes for implementing the present invention and the program's processing. Specific examples will also be included.

[1182] Data retrieval by the server

[1183] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years. The server stores the obtained data in a database.

[1184] Data preprocessing by the server

[1185] The server preprocesses the stored risk map data and weather data. Specifically, it detects and complements outliers, and fills in missing values ​​by estimating them from previous and subsequent data. It also organizes rainfall data and weather map data and converts them into a time series format. It also normalizes the data and formats it in a way that is suitable for machine learning models.

[1186] Server-based analysis using generative AI

[1187] The server uses the preprocessed data to train generative AI models such as LSTM models, which build models to predict future disaster risks based on past time-series data.

[1188] Server-based future prediction

[1189] The server uses the trained model to predict disaster risk over the next five years, generating data that quantifies monthly flood risk, for example, and storing the prediction results in a database.

[1190] Emotion engine that recognizes user emotions

[1191] The server is equipped with an emotion engine that recognizes the user's emotions from their inputs, actions, images, and voice. For example, it analyzes the text and voice entered when the user accesses the system and determines the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[1192] Server-adjusted presentation of information

[1193] After the emotion engine recognizes the user's emotion, the server adjusts the way disaster risk information is presented based on the recognized emotion. For example, if the user expresses strong anxiety, the server will present disaster risk information in a more reassuring tone. It will also alleviate the user's anxiety by displaying specific evacuation locations and countermeasures.

[1194] Visualization of prediction results on the device

[1195] The device retrieves the forecast results from the server and provides them to the user in a visual format. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, providing a design that is easy for users to understand.

[1196] Device and user interaction

[1197] Users can understand future disaster risks by viewing the visualized forecast results on their devices. Furthermore, based on the user's input and feedback, the emotion engine continuously monitors the user's emotional state and dynamically adjusts the way information is presented accordingly.

[1198] Specific examples

[1199] For example, suppose a user wants to know the risk of flooding. The server retrieves current hazard map data from the government's hazard map API, and also retrieves past rainfall data and weather chart data from the weather data API. These data are preprocessed on the server, and outliers and missing values ​​are filled in.

[1200] Next, the server uses the preprocessed data to train the data using a generative artificial intelligence (AI) model with an LSTM model. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database.

[1201] When a user accesses the system via a terminal, the terminal retrieves the forecast results from the server and displays them in the form of graphs and maps. For example, a graph showing flood risk forecasts for the next five years based on data from the past 30 years can be displayed in an easy-to-understand manner.

[1202] Furthermore, the emotion engine recognizes the user's emotional state and presents information that provides a sense of security if the user feels anxious. In this way, the system of the present invention effectively utilizes current risk maps and past weather data, and uses generative artificial intelligence and the emotion engine to provide highly accurate disaster predictions and appropriate information. This allows users to take more effective measures against future disaster risks.

[1203] The processing flow will be explained below.

[1204] Step 1:

[1205] The server connects to external data sources to retrieve the latest risk maps and historical weather data. Specifically, it accesses the government's hazard map API to retrieve risk map data, and weather data API to retrieve rainfall data and weather chart data for the past 30 years.

[1206] Step 2:

[1207] The server stores the acquired risk map data and weather data in a database, along with the corresponding timestamp and related meta information.

[1208] Step 3:

[1209] The server preprocesses the risk map data and weather data stored in the database. Specifically, it detects outliers and interpolates them based on the preceding and following data. Missing values ​​are also interpolated based on the preceding and following data. The server also converts the data into a time series format and normalizes it.

[1210] Step 4:

[1211] The server then uses the preprocessed data to train a generative artificial intelligence model (such as an LSTM model). Using past time-series data, the server builds a model to predict future disaster risk. This model quantifies flood risk and other natural disaster risks.

[1212] Step 5:

[1213] The server uses the trained model to predict disaster risk for the next five years. Specifically, it predicts the next five years on a monthly basis, quantifying the flood risk for each month and storing the results in a database.

[1214] Step 6:

[1215] Users access the system using a terminal and request prediction results through a web browser or application.

[1216] Step 7:

[1217] The server receives the user's request, retrieves the latest disaster risk prediction results from the database, and sends the results to the terminal.

[1218] Step 8:

[1219] The device visualizes the disaster risk prediction results it receives. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, making it easy for users to understand visually.

[1220] Step 9:

[1221] The emotion engine recognizes the user's emotions based on their input and actions. For example, it analyzes the text, voice, and facial expressions entered by the user to determine whether the user is feeling anxious or relieved.

[1222] Step 10:

[1223] The server then adjusts the presentation of disaster risk information based on the user's perceived emotions: for example, if the user feels anxious, it will present the information in a more reassuring tone and provide additional information on precautions and evacuation locations.

[1224] Step 11:

[1225] The device then revisits the adjusted information and provides it to the user, allowing them to check disaster risk information with greater peace of mind and take appropriate measures.

[1226] Example 2

[1227] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1228] Current disaster risk prediction systems predict disaster risk based on past data and current conditions, but no systems provide information that takes user emotions into consideration. This makes users more likely to feel anxious and makes it difficult for them to take appropriate disaster prevention actions. Furthermore, there are problems with insufficient data processing and imputation of outliers and missing values ​​to accurately predict disaster risk.

[1229] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1230] In this invention, the server includes means for acquiring a current risk map from an external data source, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now, means for recognizing a user's emotions and adjusting the method of presenting information, and means for visualizing the predicted disaster risks. This enables highly accurate disaster risk predictions and appropriate information provision while taking user emotions into consideration.

[1231] "External Data Source" refers to a source of data from an external source. Examples include government hazard map APIs and weather data APIs.

[1232] "Current risk maps" refer to map data showing disaster risks at the present time. These usually include risk information for earthquakes, floods, landslides, etc.

[1233] "Historical Weather Data" refers to datasets based on past weather information, including precipitation, weather maps, and temperatures.

[1234] "Preprocessing" refers to the process of cleaning, completing, and transforming data into a format suitable for machine learning models, including imputing outliers, filling in missing values, converting to time series data, and normalizing.

[1235] "Generative AI" refers to AI technology that uses past data to make future predictions, including deep learning models such as LSTM (Long Short-Term Memory).

[1236] "Disaster risk" is an indicator that shows the possibility of natural disasters such as earthquakes, floods, and landslides occurring.

[1237] "Emotion recognition" refers to the technology of analyzing and determining a user's emotional state (e.g., anxiety, relief, surprise, etc.) based on input data such as text, voice, and images.

[1238] "Adjusting the way information is presented" refers to the process of changing the means and content of information presentation based on the user's recognized emotions. For example, for a user who is feeling anxious, adding a reassuring message or specific countermeasures.

[1239] "Visualization" refers to the technology of displaying predicted disaster risks in visual formats such as graphs and maps, making it easier for users to understand the prediction results.

[1240] An "outlier" is a value that is extremely different from the other data points in a data set.

[1241] "Missing values" refers to the absence of values ​​in a dataset.

[1242] The present invention is a system that uses a current risk map and past weather data to predict disaster risks several years into the future using generative artificial intelligence, and further recognizes the user's emotions and adjusts the way information is presented. Specific embodiments for implementing the present invention are described below.

[1243] Data retrieval by the server

[1244] The server connects to external data sources (e.g., government hazard map APIs and weather data APIs) and periodically retrieves current risk map data and historical weather data (e.g., rainfall data and weather chart data from the past 30 years). The server stores this data in a database (e.g., PostgreSQL) and manages it securely.

[1245] Data preprocessing by the server

[1246] The server preprocesses the stored data. Specifically, it uses Python's Pandas library to detect and impute outliers. Missing values ​​are estimated and filled using linear interpolation and moving averages. Furthermore, rainfall data and weather map data are converted into time series data and normalized, allowing machine learning models to learn from the data efficiently.

[1247] Server-based analysis using generative AI

[1248] The server uses the preprocessed data to build and train a generative artificial intelligence model (e.g., an LSTM model) using a deep learning framework such as TensorFlow or Keras. The model is optimized over multiple epochs using a dataset split into training and test data.

[1249] Server-based future prediction

[1250] The server uses the trained model to predict disaster risk for the next five years. Specifically, it generates data quantifying monthly flood risk and stores the prediction results in a database. For example, it provides input to the generative AI model based on prompts such as, "Please predict the flood risk for the next five years."

[1251] Emotion engine that recognizes user emotions

[1252] The server is equipped with an emotion engine that recognizes the user's emotions from their input, actions, or image and audio data. Specifically, it uses Python's NLTK library and OpenCV to analyze text and audio and determine the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[1253] Server-adjusted presentation of information

[1254] After the emotion engine recognizes the user's emotions, the server adjusts the way disaster risk information is presented based on the user's emotions. For example, if the user expresses strong anxiety, the server will provide information in a reassuring tone and additionally display specific evacuation locations and countermeasures to alleviate the user's anxiety.

[1255] Visualization of prediction results on the device

[1256] The terminal uses visualization libraries such as Matplotlib to display the forecast results obtained from the server in graphs and maps. For example, it can visually display flood risk forecast results for the next five years, providing a screen that is easy for users to understand.

[1257] Device and user interaction

[1258] Users can view the visualized forecast results on their devices to understand future disaster risks. Furthermore, the emotion engine continuously monitors the user's emotional state based on their input and feedback, and dynamically adjusts the way information is presented accordingly.

[1259] Specific examples

[1260] For example, if a user wants to know the "flood risk for the next five years," the server retrieves the latest hazard map data from the government's hazard map API and downloads historical rainfall data from the weather data API. These data are then interpolated for outliers and missing values ​​and converted into time-series data.

[1261] The data is then trained using an LSTM model to predict flood risk over the next five years. The results are stored in a database, and when users access the system from their devices, the forecast results are displayed in graphs and maps. The emotion engine recognizes the user's emotions and adds reassuring messages such as "There is a high risk of flooding, but there are many evacuation sites in this area and measures are in place."

[1262] This allows the system to provide highly accurate disaster risk predictions and appropriate information while taking into consideration the user's feelings.

[1263] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1264] Step 1:

[1265] The server connects to external data sources (for example, government hazard map APIs or weather data APIs) and periodically retrieves current risk map data and weather data from the past 30 years (rainfall data and weather chart data). The input is data provided through API requests, and the output is raw data that is stored in the server's database. Specifically, it uses the Python Requests library to access the API and retrieve and store the data.

[1266] Step 2:

[1267] The server preprocesses the stored data. The input is the data already retrieved in the database, and the output is the preprocessed data. Specifically, it uses the Python Pandas library to detect and impute outliers. It also estimates missing values ​​using linear interpolation and moving averages, converts them into time series data, and normalizes the data to make it suitable for machine learning models.

[1268] Step 3:

[1269] The server uses the preprocessed data to train a generative AI model, such as an LSTM model. The input is the preprocessed time series data, and the output is a trained LSTM model. Specifically, the model is defined using TensorFlow or Keras and trained through multiple epochs.

[1270] Step 4:

[1271] The server uses the trained model to predict disaster risk for the next five years. The input is the trained LSTM model, the latest risk map, and weather data, and the output is quantified disaster risk prediction data. For example, the server performs predictions based on prompt statements such as, "Please predict the flood risk for the next five years."

[1272] Step 5:

[1273] The server uses an emotion engine to recognize the user's emotional state. The input is the user's text, voice, and image data, and the output is the user's emotional state (anxiety, relief, surprise, etc.). Specifically, the server analyzes the text and voice data using Python's NLTK library and OpenCV.

[1274] Step 6:

[1275] The server adjusts the way disaster risk information is presented based on the recognized emotion. The input is the user's emotional state and disaster risk prediction data, and the output is customized information presentation. For example, if the user expresses anxiety, it displays a reassuring message along with specific evacuation locations and countermeasures.

[1276] Step 7:

[1277] The terminal obtains the forecast results from the server and provides them to the user using a visualization library such as Matplotlib. The input is disaster risk forecast data, and the output is visual information in the form of graphs and maps. Specifically, the terminal generates graphs and maps and displays them in a format that is easy for the user to understand.

[1278] Step 8:

[1279] The user checks and understands the visualized forecast results on their device. The input is the visualized disaster risk forecast data, and the output is the user's feedback and changes in emotional state. Based on the user's feedback, the emotion engine continuously monitors changes in emotions and dynamically adjusts the way information is presented.

[1280] In this way, the system takes into consideration the user's feelings while achieving highly accurate disaster risk prediction and providing appropriate information.

[1281] (Application example 2)

[1282] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1283] Currently, many disaster risk prediction systems simply predict future risks based on past data and do not take into account the user's emotional state or how that information is presented. This can lead to users feeling unnecessarily anxious or, conversely, becoming indifferent. Furthermore, particularly in electronic payment services, it is important to increase the security of transactions by providing appropriate disaster risk information. The objective of this invention is to provide a system that recognizes the user's emotions and presents appropriate disaster risk information based on those emotions, allowing users to use electronic payment services with peace of mind.

[1284] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1285] In this invention, the server includes means for acquiring a current risk map, means for acquiring past weather data, means for preprocessing the current risk map and the past weather data as data sources, means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now, means for visualizing the predicted disaster risks, means for determining the emotional state of a user using an emotion engine, and means for adjusting the presentation method of disaster risk information based on the emotional state, thereby enabling the user to receive disaster risk information that is appropriately adjusted to match their emotional state.

[1286] A "current risk map" is data provided by governments and public institutions that shows the current risk of natural disasters in map form.

[1287] "Past weather data" refers to observational data on weather over a certain period of time in the past, including rainfall, temperature, and air pressure.

[1288] "Preprocessing" refers to processes such as organizing data, correcting outliers, and estimating missing values ​​in order to prepare raw data in a format suitable for machine learning models.

[1289] "Generative AI" is an AI technology that has the ability to generate new information from input data, and is often used to predict time series data.

[1290] An "emotion engine" is a technology that recognizes and classifies a user's emotional state based on their input, actions, images, voice, etc.

[1291] "Visualization" is a method of presenting data and prediction results in a visually easy-to-understand format, such as a graph or map.

[1292] "Adjusting presentation method" is a technology that provides information more appropriately by changing the method and content of information transmission depending on the user's current emotional state.

[1293] The present invention provides a system for predicting disaster risks in an area where a user resides and providing information that gives the user a sense of security based on the predicted disaster risks. Specific embodiments and their program processing will be described in detail below.

[1294] Data Acquisition Method

[1295] The server periodically accesses external data sources (e.g., government hazard map APIs and weather data APIs) to obtain the latest risk map data and historical weather data, including rainfall data and weather chart data from the past 30 years. The server stores the obtained data in a database.

[1296] Data preprocessing methods

[1297] The server preprocesses the stored risk map data and weather data. Specifically, it detects and complements outliers, and fills in missing values ​​by estimating them from previous and subsequent data. It also organizes rainfall data and weather map data and converts them into a time series format. It also normalizes the data and formats it in a way that is suitable for machine learning models.

[1298] A means of predictive analysis using generative artificial intelligence

[1299] The server uses the preprocessed data to train a generative artificial intelligence such as an LSTM model. Based on past time series data, it builds a model to predict future disaster risk. The server then uses the trained model to predict disaster risk over the next five years. For example, it generates data that quantifies monthly flood risk and stores the prediction results in a database.

[1300] A method for determining user emotions using an emotion engine

[1301] The server is equipped with an emotion engine that recognizes the user's emotions from their input, actions, images, and voice. For example, it analyzes the text and voice entered when the user accesses the system and determines the user's emotional state (e.g., anxiety, relief, surprise, etc.).

[1302] A means of adjusting information presentation

[1303] After the emotion engine recognizes the user's emotion, the server adjusts the way disaster risk information is presented based on the recognized emotion. For example, if the user expresses strong anxiety, the server will present disaster risk information in a more reassuring tone. It will also alleviate the user's anxiety by displaying specific evacuation locations and countermeasures.

[1304] Visualization of prediction results

[1305] The device retrieves the forecast results from the server and provides them to the user in a visual format. Specifically, it uses visualization libraries such as Matplotlib to display the predicted disaster risk in graphs and maps, providing a design that is easy for users to understand.

[1306] Device and user interaction

[1307] Users can understand future disaster risks by viewing the visualized forecast results on their devices. Furthermore, based on the user's input and feedback, the emotion engine continuously monitors the user's emotional state and dynamically adjusts the way information is presented accordingly.

[1308] Examples of concrete examples and prompts

[1309] For example, suppose a user wants to know about flood risk. The server obtains current hazard map data from the government's hazard map API, and also obtains historical rainfall and weather chart data from a weather data API. These data are preprocessed on the server, and outliers and missing values ​​are interpolated. Next, using the preprocessed data, the server uses an LSTM model to train the data using generative artificial intelligence. Once trained, the model predicts flood risk for the next five years on the server and stores the results in a database. When the user then accesses the system on their device, the device retrieves the prediction results from the server and displays them as graphs or maps. For example, a graph predicting flood risk for the next five years based on data from the past 30 years can be displayed, providing a clear picture to the user. Furthermore, an emotion engine recognizes the user's emotional state, and if the user is anxious, information that provides reassurance is presented.

[1310] Example prompt for a generative AI model:

[1311] "Train an LSTM model to predict flood risk for the next five years based on rainfall data and meteorological map data from the past 30 years. Then, recognize user-input emotions and generate messages to provide reassurance and risk information if the user is anxious."

[1312] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1313] Step 1:

[1314] The server accesses the government's hazard map API and weather data API to obtain current risk map data and weather data for the past 30 years. The data obtained in this step is stored in a database as raw data.

[1315] Input: Government hazard map API and weather data API

[1316] Output: Current risk map data and weather data from the past 30 years

[1317] Step 2:

[1318] The server preprocesses the raw data stored in the database, detecting outliers and filling in missing values ​​by estimating them from surrounding data. It also converts the data into a time series format and normalizes it, making it suitable for machine learning models.

[1319] Input: Raw data

[1320] Output: Preprocessed data

[1321] Step 3:

[1322] The server uses the preprocessed data to build and train an LSTM model. The trained LSTM model is used to predict disaster risk for several years into the future from past data. The trained model then predicts monthly disaster risk for the next five years and stores the prediction results in a database.

[1323] Input: Preprocessed data

[1324] Output: Predicted disaster risk data

[1325] Step 4:

[1326] The server uses an emotion engine to analyze the user's input text and voice data, thereby determining the user's emotional state (e.g., anxiety, relief, surprise, etc.). The emotion recognition results are also stored in a database.

[1327] Input: User-entered text and voice data

[1328] Output: User's emotional state data

[1329] Step 5:

[1330] The server adjusts the way it presents disaster risk information based on the user's emotional state, as determined by the emotion engine. Specifically, if the user expresses anxiety, the server presents the information in a reassuring tone and additionally displays specific evacuation locations and countermeasures.

[1331] Input: User emotional state data and predicted disaster risk data

[1332] Output: Tailored disaster risk information

[1333] Step 6:

[1334] The device retrieves the prediction results from the server and presents them to the user in a visually understandable format, such as graphs or maps, using visualization libraries like Matplotlib, and adjusts the tone of the message to match the user's emotional state, if necessary.

[1335] Input: Tailored disaster risk information

[1336] Output: Disaster risk information that is visually easy for users to understand

[1337] Step 7:

[1338] Users can view the visualized forecast results on their devices to understand future disaster risks. Furthermore, the emotion engine continuously monitors the user's emotional state based on their input and feedback, dynamically adjusting how information is presented.

[1339] Input: Visualized disaster risk information, user feedback

[1340] Output: Continuously updated emotional state data and adjusted risk information

[1341] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the 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 voice data.

[1342] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1343] 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 the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1344] The emotion identification model 59 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 an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1345] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1346] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1347] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1348] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1349] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs 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 a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1350] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1351] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1352] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1353] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1354] 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.

[1355] It is not necessary to store all 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 all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1356] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1357] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with 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). Also, the hardware resource that executes the specific processing may be a single processor.

[1358] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1359] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1360] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1361] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1362] The following is further disclosed regarding the above embodiment.

[1363] (Claim 1)

[1364] A means of obtaining the current risk map;

[1365] a means for obtaining historical weather data;

[1366] means for preprocessing the current risk map and the historical weather data as data sources;

[1367] A means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now;

[1368] A means for visualizing the predicted disaster risk;

[1369] A system including:

[1370] (Claim 2)

[1371] 10. The system of claim 1, further comprising imputing outliers and missing values ​​to the preprocessed data.

[1372] (Claim 3)

[1373] 2. The system of claim 1, wherein the weather data includes rainfall data and weather chart data.

[1374] "Example 1"

[1375] (Claim 1)

[1376] A means of periodically obtaining a current risk map from an external data source; and

[1377] a means for periodically obtaining historical weather data from an external data source; and

[1378] means for preprocessing the current risk map and the past weather data by complementing, integrating, and normalizing outliers and missing values;

[1379] A means for training a generative artificial intelligence model using the preprocessed data to predict disaster risks several years in the future;

[1380] A means for visualizing the predicted disaster risk in the form of a graph or map;

[1381] A system including:

[1382] (Claim 2)

[1383] 10. The system of claim 1, further comprising imputing outliers and missing values ​​to the preprocessed data.

[1384] (Claim 3)

[1385] 2. The system of claim 1, wherein the weather data includes rainfall data and weather chart data.

[1386] "Application Example 1"

[1387] (Claim 1)

[1388] A means of obtaining the current risk map;

[1389] a means for obtaining historical weather data;

[1390] means for preprocessing the current risk map and the historical weather data as data sources;

[1391] A means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now;

[1392] A means for visualizing the predicted disaster risk;

[1393] means for sending real-time alerts to users based on the predicted disaster risk;

[1394] A means for the user to check disaster risks around their place of residence or workplace on a map;

[1395] A system including:

[1396] (Claim 2)

[1397] 10. The system of claim 1, further comprising imputing outliers and missing values ​​to the preprocessed data.

[1398] (Claim 3)

[1399] 2. The system of claim 1, wherein the weather data includes rainfall data and weather chart data.

[1400] "Example 2: Combining Emotion Engines"

[1401] (Claim 1)

[1402] A means of obtaining a current risk map from an external data source; and

[1403] a means for obtaining historical weather data;

[1404] means for preprocessing the current risk map and the historical weather data as data sources;

[1405] A means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now;

[1406] means for recognizing a user's emotions and adjusting the presentation of information;

[1407] A means for visualizing the predicted disaster risk;

[1408] A system including:

[1409] (Claim 2)

[1410] 10. The system of claim 1, further comprising imputing outliers and missing values ​​to the preprocessed data.

[1411] (Claim 3)

[1412] 2. The system of claim 1, wherein the weather data includes rainfall data and weather chart data.

[1413] "Application example 2 when combining emotion engines"

[1414] (Claim 1)

[1415] A means of obtaining the current risk map;

[1416] a means for obtaining historical weather data;

[1417] means for preprocessing the current risk map and the historical weather data as data sources;

[1418] A means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now;

[1419] A means for visualizing the predicted disaster risk;

[1420] means for determining an emotional state of a user using an emotion engine;

[1421] a means for adjusting a presentation method of disaster risk information based on the emotional state;

[1422] A system including:

[1423] (Claim 2)

[1424] 10. The system of claim 1, further comprising imputing outliers and missing values ​​to the preprocessed data.

[1425] (Claim 3)

[1426] 2. The system of claim 1, wherein the weather data includes rainfall data and weather chart data. [Explanation of symbols]

[1427] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of obtaining the current risk map; a means for obtaining historical weather data; means for preprocessing the current risk map and the historical weather data as data sources; A means for analyzing the preprocessed data using generative artificial intelligence to predict disaster risks several years from now; A means for visualizing the predicted disaster risk; A system including:

2. The system of claim 1 , further comprising imputing outliers and missing values ​​to the preprocessed data.

3. 2. The system of claim 1, wherein the weather data includes rainfall data and weather map data.

Images