Data processing system

Abstract

The system according to the present embodiment includes a collection unit, an extraction unit, a disclosure unit, a collaboration unit, an analysis unit, and a provision unit. The collection unit collects data. The extraction unit extracts elements from the data collected by the collection unit. The disclosure unit discloses the data extracted by the extraction unit on the Internet. The collaboration unit is responsible for collaboration with local governments. The analysis unit analyzes the data collected by the collection unit by AI. The provision unit provides the analysis result obtained by the analysis unit.

Description

Technical Field

[0001] The technology disclosed herein relates to a system. Background Technology

[0002] Patent Document 1 discloses a personalized chatbot control method executed by at least one processor, the method comprising: receiving user speech; adding the user speech to a prompt containing instructions related to a chatbot role; encoding the prompt; and inputting the encoded prompt into a language model to generate chatbot speech in response to the user speech.

[0003] Patent document 1: Japanese Patent Application Publication No. 2022-180282.

[0004] Current technologies are insufficient in digitizing and widely disclosing information such as local traditional arts, customs, and legends, leaving room for improvement. Summary of the Invention

[0005] The system described in this embodiment includes a collection unit, an extraction unit, a publishing unit, a collaboration unit, an analysis unit, and a provision unit. The collection unit collects data. The extraction unit extracts elements from the data collected by the collection unit. The publishing unit publishes the data extracted by the extraction unit on the Internet. The collaboration unit is responsible for collaboration with local governments. The analysis unit uses AI to analyze the data collected by the collection unit. The provision unit provides the analysis results obtained by the analysis unit. Attached Figure Description

[0006] Figure 1 This is a conceptual diagram illustrating an example of the configuration of a data processing system according to the first embodiment.

[0007] Figure 2 This is a conceptual diagram illustrating an example of the main functions of the data processing apparatus and smart device according to the first embodiment.

[0008] Figure 3 This is a conceptual diagram illustrating an example of the data processing system configuration in the second embodiment.

[0009] Figure 4 This is a conceptual diagram illustrating an example of the main functions of the data processing device and smart glasses according to the second embodiment.

[0010] Figure 5 This is a conceptual diagram illustrating an example of the data processing system configuration in the third embodiment.

[0011] Figure 6 This is a conceptual diagram illustrating an example of the main functions of the data processing device and head-mounted terminal according to the third embodiment.

[0012] Figure 7 This is a conceptual diagram illustrating an example of the data processing system configuration in the fourth embodiment.

[0013] Figure 8 This is a conceptual diagram illustrating an example of the functions of the main parts of the data processing device and robot according to the fourth embodiment.

[0014] Figure 9 It represents an emotion graph that maps multiple emotions.

[0015] Figure 10 It represents an emotion graph that maps multiple emotions.

[0016] Explanation of reference numerals in the attached figures Data processing systems 10, 210, 310, and 410 12 Data processing devices 14 Smart devices 214 Smart Glasses 314 Head-mounted terminal 414 Robot. Detailed Implementation

[0017] Hereinafter, an example of an implementation of the system involved in this disclosure will be described with reference to the accompanying drawings.

[0018] First, let's explain the terms used in the following description.

[0019] In the following embodiments, the processor (hereinafter referred to as "processor") can be a single computing device or a combination of multiple computing devices. Furthermore, a processor can be a single computing device or a combination of multiple computing devices. Examples of computing devices include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), or TPU (Tensor Processing Unit), etc.

[0020] In the following implementation, the labeled RAM (Random Access Memory) is a memory that temporarily stores information and is used by the processor as working memory.

[0021] In the following embodiments, the labeled memory is one or more non-volatile storage devices used to store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), disk (e.g., hard disk) or magnetic tape, etc.

[0022] In the following implementation, the labeled Communication I / F (Interface) is an interface that includes a communication processor and an antenna, etc. The Communication I / F is responsible for communication between multiple computers. Examples of communication standards applicable to the Communication I / F include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0023] In the following implementation, "A and / or B" is synonymous with "at least one of A and B". That is, "A and / or B" means that it can be only A, only B, or a combination of A and B. Furthermore, in this specification, when "and / or" connects more than three items, the same approach as "A and / or B" applies.

[0024] [First Implementation] Figure 1 An example of the configuration of the data processing system 10 according to the first embodiment is shown.

[0025] like Figure 1 As shown, the data processing system 10 includes a data processing device 12 and an intelligent device 14. An example of the data processing device 12 is a server.

[0026] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, 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. An example of the network 54 includes a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0027] The smart device 14 includes a computer 36, a receiver 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. In addition, the receiver 38, output device 40, and camera 42 are also connected to the bus 52.

[0028] The receiving device 38 includes a touchscreen 38A and a microphone 38B, etc., for receiving user input. The touchscreen 38A receives user input generated by contact with an indicator (e.g., a pen or finger). The microphone 38B receives user input generated by sound by detecting the user's voice. The control unit 46A sends data representing user input received via the touchscreen 38A and microphone 38B to the data processing unit 12. In the data processing unit 12, a specific processing unit 290 (see...) Figure 2 Get the data that represents user input.

[0029] The output device 40 includes a display 40A and a speaker 40B, etc., and presents data to the user by outputting data in a user-perceptible form (e.g., sound and / or text). The display 40A displays visual information such as text and images according to instructions from the processor 46. The speaker 40B outputs sound according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as 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.

[0030] Communication I / F 44 is connected to network 54. Communication I / F 44 and 26 are responsible for sending and receiving various information between processor 46 and processor 28 via network 54.

[0031] Figure 2 An example of the main functions of the data processing device 12 and the smart device 14 is shown.

[0032] like Figure 2 As shown, in the data processing apparatus 12, specific processing is performed by the processor 28. A specific processing program 56 is stored in the memory 32. The specific processing program 56 is an example of a "program" as understood in this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 based on the specific processing program 56 executed on the RAM 30.

[0033] The memory 32 stores a data generation model 58 and an emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290. The specific processing unit 290 can use the emotion-specific model 59 to infer the user's emotions and perform specific processing based on the user's emotions. The emotion inference function using the emotion-specific model 59 (emotion-specific function) includes various inferences and predictions related to the user's emotions, but is not limited to this example. Furthermore, emotion inference and prediction also include, for example, emotion analysis (analysis).

[0034] In the smart device 14, specific processing is performed by the processor 46. A specific processing program 60 is stored in the memory 50. The specific processing program 60 is used in conjunction with the data processing system 10. The processor 46 reads the specific processing program 60 from the memory 50 and executes the read specific processing program 60 on the RAM 48. Specific processing is implemented by the processor 46 operating as a control unit 46A based on the specific processing program 60 executed on the RAM 48. Furthermore, the smart device 14 may also have the same data generation model and emotion-specific model as the data generation model 58 and the emotion-specific model 59, and use these models to perform the same processing as the specific processing unit 290.

[0035] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains the processing results (prediction results, etc.) using the data generation model 58 by communicating with the server device that has the data generation model 58. In addition, the data processing device 12 may be a server device or a user-owned terminal device (e.g., a mobile phone, robot, home appliance, etc.). Next, an example of the processing performed by the data processing system 10 of the first embodiment will be described.

[0036] (Example) The Digital Transformation (DX) Promotion System described in this embodiment is a system built to support the digitization needed for promoting DX across all regions, leveraging the advantages of stores nationwide. The system first collects information unique to each region, such as traditional crafts, customs, and legends. This information is obtained by reading images, sounds, and videos from data sources on the internet. Next, elements are extracted from the acquired data and made public on the internet. Thus, all information in the world is digitized, improving the accuracy of AI. Furthermore, it collaborates with local governments to build a foundation for promoting regional appeal. For example, images of traditional festivals in a region are collected, detailed information about the festivals is extracted from these images, and made public on the internet. This allows people in other regions to learn about the festivals, spreading the region's appeal. In addition, the collected data is used as learning data for AI. This improves the accuracy of AI, enabling more sophisticated analysis and proposals. For example, AI analyzes images of traditional crafts in a region, learning their characteristics to compare and analyze them with traditional crafts in other regions. Through this system, all citizens can promote DX across regions and build a foundation for promoting regional appeal. Therefore, the DX Promotion System enables the entire population to promote DX across the region and builds the foundation for showcasing regional appeal. Specifically, this system possesses a large-scale data processing platform built on a distributed cloud computing environment, equipped with the ability to efficiently collect unstructured data from edge devices and local servers scattered throughout the country. The system maps the collected image data (e.g., 3D tensors composed of pixel values), audio data (e.g., amplitude vectors of time series), and text data (e.g., token ID sequences) to a high-dimensional feature space using deep learning models based on convolutional neural networks (CNN) or Transformer architectures. The system then uses these features to quantify the region's inherent cultural elements (e.g., movement patterns in festivals, rhythmic features of dialects, visual textures of traditional crafts) and stores them as a knowledge graph in a structured database. In the AI ​​learning process of this system, collected multimodal data is used as input. By employing self-supervised learning or contrastive learning, the system can reduce annotation costs while obtaining potential representations that can accurately identify similarities or differences between regions. For example, using image data of Kagura from a certain region as input (a sequence of image tensors for each frame), the action recognition model analyzes the dance patterns and outputs action labels such as "rotation" and "jump" along with their probability scores, thereby enabling quantitative comparisons with Kagura from other regions.Through this technical structure, the system is not limited to simple information collection, but can also digitize regional cultural assets in a computer-understandable form, achieving high-level information processing effects such as correlation analysis between heterogeneous data or discovery of new tourism resources.

[0037] The DX promotion system implemented includes a collection unit, an extraction unit, a publishing unit, a collaboration unit, a parsing unit, and a provision unit. The collection unit collects data. For example, it collects data from data sources on the internet. The collection unit can collect data from news websites, social networking sites (SNS), blogs, etc. It can also collect data from Yahoo! or instant messaging applications. The extraction unit extracts elements from the collected data. For example, it extracts elements such as keywords, feature quantities, and patterns from the collected data. The extraction unit can extract keywords from text data using natural language processing techniques. The publishing unit publishes the extracted data on the internet. The publishing unit can publish the data using platforms such as websites, SNS, and databases. The publishing unit can publish the data on a website, making it accessible to anyone. The collaboration unit is responsible for collaborating with local governments. For example, it implements joint projects with local governments, collecting and publishing data for promoting regional appeal. The collaboration unit can collaborate with local governments via APIs to share data. The parsing unit uses AI to parse the collected data. The parsing unit uses techniques such as machine learning, deep learning, and natural language processing to parse the data. The parsing unit, for example, can use AI to parse the collected data and extract its features. The providing unit provides the parsing results. The providing unit provides the parsing results, for example, in the form of reports, charts, statistics, etc. The providing unit can publish the parsing results on a website, making them accessible to anyone. Thus, the DX facilitation system according to the implementation method can efficiently collect, extract, publish, collaborate, parse, and provide data. Specifically, each part of this system is installed as an independent container based on a microservice architecture and managed by orchestration tools such as Kubernetes. The collection unit consists of a web crawler and an API connector group, which has the function of issuing HTTP requests in parallel to obtain HTML documents or JSON objects and putting them into a distributed message queue (such as Kafka). The extraction unit includes a natural language processing (NLP) engine and an image recognition engine. It performs named entity recognition (NER) on the input text data to identify entities such as "place names" and "event names", and applies object detection algorithms (such as YOLO or EfficientDet) to the image data to identify the main objects in the image and output their bounding box coordinates and class probabilities. This analysis unit comprises an inference server running on a GPU cluster. Using extracted feature vectors as input, it calculates regional vitality scores or tourism demand predictions through a regression model. This collaboration unit connects to the autonomous system's backbone system via a secure API gateway, facilitating data exchange through encrypted communication using mutual authentication (mTLS). This provisioning unit features a dashboard generation engine for visualizing the analysis results, dynamically generating and publishing chart rendering data (SVG or Canvas instructions) optimized for the user's terminal resolution or communication environment.Through these structures, this system achieves the technical effect of maximizing throughput and minimizing latency in the processing of large and diverse amounts of data.

[0038] The data collection department is capable of gathering data from data sources on the internet. For example, it can collect data from news websites, social networking sites (SNS), blogs, etc. It can also collect data from Yahoo! or instant messaging applications. For example, it can collect the latest news reports from news websites and save them to a database. For example, it can collect user submissions from SNS and save them to a database. For example, it can collect articles from blogs and save them to a database. Thus, the data collection department can obtain a wide range of information from data sources on the internet. Some or all of the above processing in the data collection department can be performed using AI, or it can be performed without AI. For example, the data collection department can input the data collected from data sources on the internet into a generating AI, allowing the generating AI to perform data parsing. Specifically, this data collection department has a crawling module that receives a list of target URLs or search queries as input and performs high-speed crawling using asynchronous I / O. This data collection department inputs the acquired unstructured text data, such as HTML, into a large language model (LLM). Here, the input to the LLM is a sequence of tokens containing the main text and metadata of the web page. The output from the LLM is a structured JSON object containing a summary of the report, main topics, sentiment polarity (positive / negative / neutral), and related entities (place names, organization names). This collection unit uses this structured data to filter out only useful information from noisy web data, reducing the processing load in subsequent steps. Furthermore, this collection unit is equipped with a patrol scheduler using reinforcement learning, which observes the update frequency or importance of each website as a state and determines the next URL to be visited and the timing as an action, thereby optimizing network bandwidth utilization while maintaining the freshness of information.

[0039] The extraction unit is capable of extracting elements from collected data. For example, it extracts keywords, feature values, patterns, and other elements from the collected data. For example, it can use natural language processing techniques to extract keywords from text data. For example, it can use image recognition techniques to extract feature values ​​from image data. For example, it can use speech recognition techniques to extract patterns from sound data. Thus, by extracting elements from collected data, the extraction unit facilitates data processing. Some or all of the above processing in the extraction unit can be performed using AI, or it can be performed without AI. For example, the extraction unit can input the collected data into a generative AI, allowing the AI ​​to perform element extraction. Specifically, this extraction unit utilizes a multimodal deep learning model to map heterogeneous data to a common semantic space. For text data, this extraction unit uses a Transformer-based encoder (e.g., BERT) to convert the input article into a fixed-length embedding vector (e.g., 768-dimensional). For image data, this extraction unit uses a convolutional neural network (e.g., ResNet) to generate visual feature maps from the input image (e.g., a 224x224 pixel RGB image) and pools them to obtain feature vectors. This extraction unit uses a speech recognition model (e.g., Wav2Vec) to convert waveform data into sequences of phonemes or words, and extracts vectors representing the speaker's emotional or prosodic features. These vectors are then fed as input to a generative AI (multimodal LLM), which outputs descriptive text such as "This image depicts a vibrant festival," or tag information such as "festival," "portable shrine," or "summer." Through this processing, the extraction unit effectively provides semantic indexes to unstructured data, dramatically improving the accuracy and speed of database retrieval.

[0040] The public disclosure department can publicly release extracted data on the internet. This can be done through platforms such as websites, social networking sites (SNS), and databases. For example, the department can publish the data on websites, making it accessible to anyone. It can also submit the data to social networking sites for widespread sharing. Furthermore, it can store the data in a database, making it searchable. Thus, by publicly releasing extracted data on the internet, the public disclosure department promotes information sharing. Some or all of the above processes in the public disclosure department can be performed using AI, or they can be performed without AI. For example, the public disclosure department can input the extracted data into a generating AI, allowing the AI ​​to perform the data disclosure. Specifically, this public disclosure department has a content generation engine that automatically generates content based on the characteristics of the platform where the data is to be published, using the extracted structured data and feature vectors as input. The public disclosure department specifies the extracted keywords or summary text, as well as prompts for the target audience (e.g., "for tourists," "for researchers"), as input to the generating AI (LLM). The generating AI outputs SEO (Search Engine Optimization) blog post titles, body text, hashtags, and meta descriptions. Furthermore, this publication utilizes image generation AI to generate eye-catching images that match the article's content, or automatically performs processing such as blurring faces for privacy protection. This publication automatically submits generated content via the API of its CMS (Content Management System) and simultaneously updates the cache on its CDN (Content Delivery Network), thereby achieving low-latency information delivery to users worldwide.

[0041] The Collaboration Department is responsible for collaborating with local governments. For example, the Collaboration Department can implement joint projects with local governments, collecting and publishing data for promoting regional appeal. For example, the Collaboration Department can collaborate with local governments via APIs to share data. For example, the Collaboration Department can co-organize events with local governments, collecting and publishing data on those events. For example, the Collaboration Department can collaborate with local governments to collect and publish data on regional traditional skills, customs, legends, etc. Thus, by collaborating with local governments, the Collaboration Department builds the foundation for promoting regional appeal. Some or all of the above processes within the Collaboration Department can be performed using AI, or not. For example, the Collaboration Department can input data about collaboration with local governments into a generative AI, allowing the AI ​​to propose collaborative methods. Specifically, this Collaboration Department possesses a data integration AI that converts heterogeneous data formats (e.g., CSV, PDF, Excel, etc.) held by different local governments into a unified data model. This Collaboration Department uses raw data provided by the local governments as input, employs the generative AI to perform semantic parsing of the data, and outputs structured data such as standardized JSON-LD format. Furthermore, this collaboration department vectorizes the regional issues (e.g., "population decline," "insufficient tourists") or resources ("specialty products," "natural landscapes") of each autonomous region. By calculating similarity in the vector space, it generates AI-driven proposals that match autonomous regions with potential synergistic effects or optimal joint project solutions (e.g., "mutual promotion of specialty products"). The input to the AI ​​is attribute data from both autonomous regions and a database of past success stories. The output is a draft project plan, predicted economic effects, and a proposed implementation schedule. Thus, this collaboration department delivers a technological advantage that promotes data-driven and efficient inter-autonomous collaboration without relying on human resources.

[0042] The analysis unit is capable of using AI to analyze collected data. For example, the analysis unit may use techniques such as machine learning, deep learning, and natural language processing to analyze the data. For example, the analysis unit may use AI to analyze the collected data and extract its features. For example, the analysis unit may use machine learning algorithms to classify or cluster the data. For example, the analysis unit may use deep learning models to identify patterns in the data. For example, the analysis unit may use natural language processing techniques to analyze text data. Thus, by using AI to analyze the collected data, the analysis unit can efficiently perform data analysis. Some or all of the above processing in the analysis unit may be performed using generative AI, or it may not. For example, the analysis unit may input the collected data into generative AI, allowing the generative AI to perform data analysis. Specifically, this analysis unit is equipped with a prediction model and a classification model using multi-layer neural networks, receiving collected time-series data (such as the trend of visitor numbers, mentions on SNS) as input. This analysis unit uses sequence transformation models such as LSTM (Long Short-Term Memory) or Transformer to predict future trends or demand, outputting predicted values ​​(such as a prediction of the number of visitors next month) and their confidence intervals. Furthermore, this analysis unit uses unsupervised learning algorithms (such as K-means or DBSCAN) to cluster the collected regional data in a high-dimensional feature space, automatically discovering regional groups with similar characteristics. Then, this analysis unit takes the numerical data or clustering results as input to the Generating AI (LLM) and outputs a report in natural language describing the insights implied by those insights (e.g., text such as "SNS marketing is effective because both Region A and Region B have seen a surge in young people's attention"). Through this processing, this analysis unit provides interpretable information that is not merely a simple list of numbers, but is directly related to meaningful decisions.

[0043] The provisioning department can provide analysis results. This provisioning department can provide analysis results in the form of reports, charts, statistical data, etc. The provisioning department can publish the analysis results on a website, making them accessible to anyone. The provisioning department can submit the analysis results to social networking sites for widespread sharing. The provisioning department can save the analysis results in a database, making them searchable. Thus, by providing analysis results, the provisioning department enables users to utilize the analysis results. Some or all of the above processing in the provisioning department can be performed using AI, or it can be performed without AI. For example, the provisioning department can input the analysis results into a generating AI, allowing the generating AI to provide the results. Specifically, this provisioning department takes the multidimensional analysis result data (tensors or vectors) output by the analysis department as input and has a rendering engine that determines the optimal visualization format based on the user's attributes or browsing device. This provisioning department uses a generating AI to convert complex statistical data into summary explanations based on easy-to-understand text. For example, given structured data such as "Tourist increase rate: 15%, main factor: festival activities" as input, the AI ​​outputs the natural language text "Due to the festival, the number of tourists increased by 15%", which is then displayed along with a chart. Furthermore, this provisioning unit executes recommendation algorithms (such as collaborative filtering or matrix factorization) that dynamically change the priority of information provided based on the user's past browsing history or interest vectors. Thus, this provisioning unit achieves the technical effect of accurately highlighting truly valuable information for the user from a large volume of parsed results, reducing cognitive load caused by information overload.

[0044] The data collection unit can analyze the user's emotions and adjust the timing of data collection based on the analyzed emotions. For example, the collection unit can immediately collect data when the user is relaxed. It can postpone data collection when the user is busy. It can pause data collection when the user is excited. Thus, by adjusting the timing of data collection according to the user's emotions, the collection unit can collect data at more appropriate times. Emotion inference can be achieved, for example, using emotion inference functions such as emotion engines or generative AI. Generative AI can be, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the collection unit can be performed using AI, or not. For example, the collection unit can input the user's emotional data into the generative AI, allowing the generative AI to adjust the timing of data collection. Specifically, this collection unit has a multimodal emotion recognition model that takes biometric data (facial expressions, voice waveforms, heart rate variability data, etc.) acquired from the user's terminal's camera, microphone, or wearable sensors as input. This model employs an architecture combining convolutional neural networks (CNNs) or recurrent neural networks (RNNs) to output probability vectors representing emotional states from input data (e.g., [joy: 0.1, anger: 0.0, relaxation: 0.8, anxiety: 0.1]). The data collection unit executes decision logic that compares the output emotional score with a predetermined threshold. For example, only when the "relaxation" score exceeds the threshold is a control signal generated to trigger a data collection request (questionnaire display or location information sending request). Conversely, in cases where the "anxiety" or "anger" scores are high, backoff control is implemented to suppress interruption processing and prevent a decline in user experience (UX). Through this emotion-adaptive scheduling, the system maximizes the collection rate of high-quality data while minimizing the user's psychological burden.

[0045] The collection department can analyze users' past data provision history and select appropriate collection methods. For example, the collection department may prioritize collecting the same types of data provided by users in the past. It may also adjust the collection frequency based on the amount of data provided by users in the past. Furthermore, it may optimize the collection method based on the quality of the data provided by users in the past. Thus, by analyzing users' past data provision history, the collection department can select the optimal collection method. Some or all of the above processing in the collection department may be performed using AI, or it may not. For example, the collection department can input users' past data provision history into a generative AI, allowing the AI ​​to select the collection method. Specifically, this collection department stores each user's interaction logs (provided date and time, data category, response time, and amount of data provided) as time-series data in a database, using this as input for a reinforcement learning model (e.g., DQN or PPO) to learn. In this model, State is defined as a user's historical feature vector, Action is defined as the choice of collection method (push notifications, emails, in-app pop-ups, etc.), and Reward is defined as the success or failure of data acquisition or a data quality score. This data collection department uses a learned model to infer and execute the collection action that maximizes the expected reward based on the current user state. Furthermore, when using generative AI, this department embeds user behavior logs as text into prompts (e.g., "This user tends to provide more photo data on weekend evenings"), allowing the generative AI to output the optimal collection plan (date and time, request text, incentive). This achieves personalized, adaptive data collection control tailored to individual user characteristics, which is impossible with a generic rule base.

[0046] The data collection department can filter data based on the user's current areas of interest. For example, it may prioritize collecting data related to topics of current interest. It may also filter data based on keywords recently searched by the user. Furthermore, it may collect data based on topics from online communities the user participates in. Thus, by filtering data based on the user's current areas of interest, the collection department can collect highly relevant data. Some or all of the above processing in the collection department can be performed using AI, or it can be performed without AI. For example, the collection department can input data about the user's areas of interest into a generative AI, allowing the AI ​​to perform data filtering. Specifically, this collection department analyzes the user's search history, browsing history, and activity data such as "likes" on social media in real time to generate user profile vectors representing the user's interests. Simultaneously, for data that becomes the collection target (web articles, images, etc.), it also uses natural language processing models (such as BERT or Sentence-BERT) to generate content vectors. In a high-dimensional vector space, this collection department calculates the cosine similarity between the user profile vector and the content vector, performing filtering processing that selectively collects and saves only data with similarity exceeding a specified threshold. When using generative AI, this data collection department takes the user's list of keywords of interest and a summary of candidate data as input, requests the generative AI to output a binary classification of "relevance" or a relevance score from 0 to 1, and selects data based on the result. Through this vector-based matching process, high-precision filtering that takes into account semantic relevance can be performed, achieving savings in storage capacity and increased information density.

[0047] The data collection unit can analyze user emotions and determine the priority of data collection based on the analyzed emotions. For example, when the user is relaxed, the collection unit prioritizes collecting data of high importance. When the user is busy, the collection unit postpones collecting data of low importance. When the user is excited, the collection unit prioritizes collecting emotion-related data. Thus, by prioritizing data collection based on user emotions, the collection unit can prioritize collecting important data. Emotion inference is achieved, for example, using emotion inference functions such as emotion engines or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the collection unit can be performed using AI, or not, for example. For example, the collection unit can input user emotion data into the generative AI, allowing the generative AI to determine the data priority. Specifically, this collection unit has a priority queue for managing collection tasks and a scoring engine that dynamically calculates the priority score for each task. This scoring engine receives, as input, the original importance of the data (static score) and the user's current emotion vector (dynamic coefficient) output from the emotion recognition model. For example, when a user is in a "relaxed" state (high coefficient), the engine amplifies the scores of high-importance tasks and places them at the beginning of the queue. Conversely, when a user is in a "stressed" state (low coefficient), the scores of high-load tasks are reduced, prioritizing low-load tasks or highly entertaining data collection tasks. When using generative AI, this collection unit inputs the emotional state and a task list, allowing the generative AI to output an optimal execution order list, and then controls the task scheduler according to this order. This enables the collection of important data at moments when the user's psychological receptivity is high, improving the completion rate and quality of data collection.

[0048] When collecting data, the data collection unit can prioritize collecting highly relevant data by considering the user's geographic location information. For example, the collection unit may prioritize collecting data related to the user's current location, locations the user has previously visited, or locations the user plans to visit in the future. Thus, by considering the user's geographic location information, the collection unit can prioritize collecting highly relevant data. Some or all of the above processing in the collection unit may be performed using AI, or it may not. For example, the collection unit can input the user's geographic location information into a generating AI, allowing the AI ​​to perform data collection. Specifically, this collection unit takes the user's latitude and longitude information and movement history data obtained from GPS, Wi-Fi access points, or beacons as input. This collection unit uses geospatial indexes (such as R-Tree or Quadtree) to quickly retrieve data sources (store information, tourist attractions, local news) existing around the user's current location (within the geofence). Furthermore, this collection unit uses sequence prediction models such as recurrent neural networks (RNNs) to analyze the user's movement trajectory and predict highly probable destinations. This data collection department uses the current location and predicted destination location information as query input to generate AI or a search engine, prioritizing the crawling and caching of local culture, activities, transportation information, and other unique data. This creates a situation where the necessary data is already locally available at the time the user arrives at the location, providing a zero-latency experience that improves application responsiveness.

[0049] The data collection department analyzes users' social media activities and collects relevant data during data collection. For example, it collects data related to users' recent posts. It also collects data based on posts from accounts followed by the user. Furthermore, it collects data related to topics in groups the user participates in. Thus, by analyzing users' social media activities, the collection department can collect relevant data. Some or all of the above processing in the collection department can be performed using AI, or it can be performed without AI. For example, the collection department can input data about users' social media activities into a generative AI, allowing the AI ​​to perform data collection. Specifically, this collection department uses graph neural networks (GNNs) to model the user's social graph, representing the user's relationships with other accounts, topics, and communities as embedding vectors. This collection department not only extracts the user's own post text but also extracts features from posts by followed opinion leaders or images "liked," determining the user's potential interest clusters. When using generative AI, this data collection department uses the user's recent SNS activity logs (posted text, hashtags, reactions) as input prompts, allowing the generative AI to infer "topics the user might be interested in next" or "keywords for relevant news to be collected," and uses the output keywords to control the web crawler. This allows information matching the user's potential needs to be collected in advance, before any explicit search behavior is performed.

[0050] The extraction unit can analyze the user's emotions and determine the priority of extracted elements based on the analyzed emotions. For example, when the user is relaxed, the extraction unit prioritizes extracting high-importance elements. When the user is busy, the extraction unit postpones extracting low-importance elements. When the user is excited, the extraction unit prioritizes extracting emotion-related elements. Thus, by prioritizing the extraction of elements based on the user's emotions, the extraction unit can prioritize extracting important elements. Emotion inference is achieved, for example, using emotion inference functions such as emotion engines or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the extraction unit can be performed using AI, or not, without AI. For example, the extraction unit can input the user's emotion data into the generative AI, allowing the generative AI to perform element extraction. Specifically, in the data processing pipeline, this extraction unit manages a task list that defines the processing cost and importance for each feature quantity (entity, image object, voice keyword, etc.) that is the extraction target. This extraction unit takes the user's emotional state (e.g., arousal and valence) output from the emotion recognition model as input and executes a dynamic weighted algorithm. For example, if the user is determined to have high arousal and seek immediacy, this extraction unit increases the priority of "keyword extraction," which has low computational cost, and decreases the priority of "deep semantic parsing," which has high computational cost. When using generative AI, this extraction unit provides the input data (text or image) and the user's emotional tags as prompts, allowing the generative AI to determine "the element that most resonates with the current user's emotion (e.g., positive words or bright colors when feeling happy)," and selectively extracts and outputs only that element. Thus, by optimally allocating limited computational resources according to the user's current psychological state, the overall responsiveness of the system and user satisfaction are improved.

[0051] During extraction, the extraction unit can adjust the level of detail based on the importance of the data. For example, the extraction unit performs detailed extraction on data with high importance. For example, the extraction unit performs simplified extraction on data with low importance. The extraction unit can also adjust the level of detail in stages according to importance. Thus, by adjusting the level of detail based on the importance of the data, the extraction unit enables efficient extraction. Some or all of the above processing in the extraction unit can be performed using AI, or not. For example, the extraction unit can input the importance of the data into the generating AI, allowing the generating AI to adjust the level of detail of the extraction. Specifically, this extraction unit has a scoring model that calculates a data importance score (0.0 to 1.0) based on the metadata (reliability of the source, number of visits, update date and time) or a summary of the content of the input data. This extraction unit has an adaptive processing mechanism that switches the complexity of the applied extraction model according to this score. For example, when the importance score is high (above the threshold), this extraction unit uses a high-precision but computationally expensive large-scale model (e.g., a GPT-4 level LLM or a high-resolution image resolution model) to extract details down to subtle differences or small objects. On the other hand, when the importance score is low, this extraction unit uses a lightweight model (such as DistilBERT or MobileNet) to extract only the main keywords or approximate categories. When using generative AI, this extraction unit dynamically changes instructions such as "detailed analysis" or "concise summary" within the prompts based on their importance, controlling the number of output tokens from the generative AI. Through this process, computational costs commensurate with the value of the information can be invested, optimizing the overall throughput and cost efficiency of the system.

[0052] During extraction, the extraction unit can apply different extraction algorithms based on the data category. For example, the extraction unit applies an image recognition algorithm to image data. For example, the extraction unit applies a speech recognition algorithm to audio data. For example, the extraction unit applies a video parsing algorithm to video data. Thus, by applying different extraction algorithms based on the data category, the extraction unit enables appropriate extraction. Some or all of the above processing in the extraction unit can be performed using AI, or it can be performed without AI. For example, the extraction unit can input the data category into the generating AI, allowing the generating AI to execute the application of the extraction algorithm. Specifically, this extraction unit has a distributor that parses the MIME type or binary header of the input data to determine the data category (text, image, audio, video). In the case of image data, this extraction unit loads object detection using a convolutional neural network (CNN) or a Vision Transformer (ViT). The segmentation model outputs object regions and category labels from pixel data. In the case of audio data, this extraction unit uses an Automatic Speech Recognition (ASR) model (e.g., Whisper) to convert the speech waveform into text and further extracts acoustic features (pitch, tone). In the case of text data, this extraction unit uses a Transformer-based language model for syntactic analysis or semantic role labeling. When using generative AI, this extraction unit passes the input data to a single, massive model corresponding to the multimodal input, instructing it to perform tasks appropriate to the data format by providing prompts such as "list objects in the image" or "summarize the content of the sound." Thus, for various data formats, a dedicated, optimized processing pipeline is automatically applied to achieve high-precision information extraction.

[0053] The extraction unit can analyze the user's emotions and adjust the display method of the extracted elements based on the analyzed emotions. For example, the extraction unit provides a detailed display method when the user is relaxed. For example, the extraction unit provides a concise display method when the user is busy. For example, the extraction unit provides a visually stimulating display method when the user is excited. Thus, by adjusting the display method of the extracted elements according to the user's emotions, the extraction unit makes appropriate display possible. Emotion inference is achieved, for example, using emotion inference functions such as emotion engines or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the extraction unit can be performed using AI, or not, for example, without AI. For example, the extraction unit can input the user's emotion data into the generative AI, allowing the generative AI to perform the adjustment of the display method. Specifically, this extraction unit has the function of assigning metadata (font size, color, layout category, summary level) for UI display to the extracted data elements (text, images, numerical values). This extraction unit takes the user's emotion vector (e.g., [relaxation level, arousal level]) obtained from an emotion recognition model as input and executes rule-based or learning-based logic to determine the values ​​of this metadata. For example, if the user is determined to be "busy" (high arousal level, low relaxation level), this extraction unit summarizes the text data to its limit and establishes markers for display in an enumerated form. On the other hand, if the user is "relaxed," metadata is assigned for displaying detailed explanatory text or high-resolution images. When using generative AI, this extraction unit takes the extracted raw data and user emotion as input, gives the generative AI instructions such as "because the user is relaxed, generate a story-style explanation," and transforms the content to be displayed in accordance with the emotion before outputting it. As a result, information prompts that match the user's psychological state can be provided, which has the effect of improving the acceptability and comprehension of information.

[0054] During extraction, the extraction unit can determine the extraction priority based on the data submission time. For example, the extraction unit may prioritize extracting the most recently submitted data. It may also postpone extracting older data. Furthermore, the extraction unit may adjust the extraction priority in stages based on the submission time. Thus, by determining the extraction priority based on the data submission time, the extraction unit enables efficient extraction. Some or all of the above processing in the extraction unit may be performed using AI, or it may not. For example, the extraction unit can input the data submission time into the generating AI, allowing the AI ​​to determine the extraction priority. Specifically, this extraction unit has a scheduler that manages the data processing queue, referencing the timestamp information of each data packet. This extraction unit calculates the difference between the current time and the data's timestamp (elapsed time), using a decay function that assigns higher priority scores to data with shorter elapsed times (newer) data. For example, using an exponential decay model, the score of the most recent data requiring real-time processing is maintained higher, while data that has elapsed for a certain period is moved to a lower priority queue for batch processing. When using AI-generated data, this extraction unit takes the data's metadata (creation date and time, validity period) as input, allowing the AI ​​to determine "Is this data highly urgent?", and dynamically rearranges the processing order based on the result (High / Medium / Low). This achieves the technical effect of processing news or disaster information requiring rapid reporting without delay, ensuring the freshness of the information.

[0055] During extraction, the extraction unit can adjust the extraction order based on the relevance of the data. For example, the extraction unit prioritizes extracting data with high relevance. Alternatively, it may postpone extracting data with low relevance. Furthermore, the extraction unit may adjust the extraction order in stages based on relevance. Thus, by adjusting the extraction order based on data relevance, the extraction unit can perform efficient extraction. Some or all of the above processing in the extraction unit can be performed using AI, or it can be performed without AI. For example, the extraction unit can input the relevance of the data into the generating AI, allowing the AI ​​to adjust the extraction order. Specifically, this extraction unit vectorizes the currently processed task or the context (sense) that the user is interested in, and calculates the similarity between this vector and the vectors of the waiting data group. This extraction unit calculates the inner product or cosine similarity between the context vector and the data vector, and determines data with high similarity as "highly relevant" and moves it to the beginning of the processing queue. For example, if processing data about "festivals" in a certain region, the priority of data about "traffic information" or "weather" in the same region is dynamically increased. When using generative AI, this extraction unit takes the current context and data list as input, allowing the AI ​​to infer "which piece of data with the highest relevance should be processed next," and controls the processing flow based on its output. This enables consistent information extraction that fits the context, eliminating inefficiencies caused by fragmented data processing.

[0056] The public display department can analyze user emotions and adjust the display method of the public data based on the analyzed user emotions. For example, when the user is relaxed, the public display department provides a detailed display method. For example, when the user is busy, the public display department provides a concise display method. For example, when the user is excited, the public display department provides a visually stimulating display method. Thus, by adjusting the display method of the public data according to the user's emotions, the public display department can provide appropriate display. Emotion inference is achieved using emotion inference functions, such as using an emotion engine or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the public display department can be performed using AI, or not. For example, the public display department can input the user's emotion data into the generative AI, allowing the generative AI to perform the adjustment of the display method. Specifically, this public display department has a rendering engine that dynamically generates HTML / CSS or UI components that constitute web pages or application screens. This disclosure takes user emotional parameters (e.g., stress level, concentration) received from the emotion recognition module as input and modifies CSS variables (font size, line spacing, color theme) or DOM structure (number or order of displayed chapters). For example, when the user is in an "excited" state, a dynamic layout using highly saturated colors and large images is generated; when in a "relaxed" state, a static layout using soft colors and a legible Song typeface is generated. When using a generative AI, this disclosure takes raw data and user emotions as input, instructs the generative AI to "create eye-catching headlines and impactful summaries for excited users," and displays the generated text combined with style information. This provides an interface synchronized with the user's psychological state, effectively increasing engagement.

[0057] When publishing data, the disclosure department can adjust the level of detail based on the importance of the data. For example, the disclosure department can publish high-importance data in detail. Alternatively, it can publish low-importance data briefly. Or, it can adjust the level of detail periodically based on importance. Thus, by adjusting the level of detail based on the importance of the data, the disclosure department can achieve efficient publishing. Some or all of the above processing in the disclosure department can be performed using AI, or it can be performed without AI. For example, the disclosure department can input the importance of the data into the AI ​​generator, allowing the AI ​​to adjust the level of detail. Specifically, this disclosure department has filtering logic that controls the level of detail (LOD) of information based on the importance score of the data (based on a value predicted by access frequency or social impact). For high-importance data (e.g., disaster alerts, major tourist events), this disclosure department publishes a complete set of information including full text, high-resolution images, and related links. On the other hand, for low-importance data (e.g., daily logs, minor updates), only the title and a one-line summary are displayed, hiding the detailed content behind a "view more" button, or omitting the disclosure itself. When using AI-generated content, this publication takes raw data and importance scores as input, provides the AI ​​with prompts such as "because of high importance, a detailed explanatory report will be generated" or "because of low importance, a summary in 30 words will be used," and uses the output text as the content to be published. This achieves the technical effect of focusing the user's attention on important information and preventing information overload.

[0058] When publishing, the disclosure section can apply different publishing algorithms based on the data category. For example, the disclosure section applies an image display algorithm to image data. For example, the disclosure section applies an audio playback algorithm to audio data. For example, the disclosure section applies a video playback algorithm to video data. Thus, by applying different publishing algorithms according to the data category, the disclosure section can make appropriate disclosures. Some or all of the above processing in the disclosure section can be performed using AI, or it can be performed without AI. For example, the disclosure section can input the data category into the generating AI, allowing the generating AI to execute the application of the publishing algorithm. Specifically, this disclosure section has a selector function that parses the MIME type or metadata of the content and selects the best player or viewer component. In the case of image data, this disclosure section performs lazy loading or automatic conversion to highly compressed formats such as WebP, and applies algorithms for displaying in slideshow or grid format. In the case of video data, this disclosure section uses adaptive streaming media (HLS or DASH) technology that automatically adjusts the bitrate according to the user's bandwidth to control preview playback or automatic playback. When using generative AI, this disclosure takes the category and content of the data as input and instructs the generative AI to "generate HTML / JavaScript code to most effectively display the data." For example, if it is 3D model data, it outputs interactive viewer code using WebGL and embeds it into the page. In this way, a display environment that always provides the best user experience (UX) is automatically built for a wide variety of media formats.

[0059] The disclosure department can analyze user sentiment and determine the priority of published data based on the analyzed sentiment. For example, when a user is relaxed, the disclosure department prioritizes publishing data of high importance. When a user is busy, the disclosure department postpones publishing data of low importance. When a user is excited, the disclosure department prioritizes publishing data related to their sentiment. Thus, by prioritizing the publication of data based on user sentiment, the disclosure department can prioritize the publication of important data. Sentiment inference can be achieved using sentiment inference functions, such as sentiment engines or generative AI. Generative AI can be text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the disclosure department can be performed using AI, or not. For example, the disclosure department can input user sentiment data into generative AI, allowing the generative AI to determine the data priority. Specifically, this disclosure department incorporates the user's sentiment state as a feature in the learning-to-rank algorithm that determines the display order of news sources or timelines. When calculating the ratings of various content, this publication increases the weight of longer articles or in-depth analytical reports (high importance) and displays them at the top if the user is in a "relaxed" state. Conversely, if the user is in a "busy (anxious)" state, it increases the weight of quick-to-consume content or data that only states the conclusion. When using AI-generated content, this publication takes a list of candidate content and the user's sentiment vector as input, instructs the AI ​​to "reorder according to the order that best suits the current user's mood," and obtains the reordered list as output. This allows for information delivery that aligns with the user's acceptance level, maximizing the click-through rate (CTR) or dwell time.

[0060] When publishing data, the disclosure department can determine the priority of publication based on the data's submission date. For example, the disclosure department prioritizes publishing the most recently submitted data. Alternatively, it may postpone older data. Furthermore, the disclosure department may periodically adjust the publication priority based on the submission date. Thus, by determining the publication priority based on the data's submission date, the disclosure department can perform efficient publication. Some or all of the above processing in the disclosure department can be performed using AI, or it can be performed without AI. For example, the disclosure department can input the data's submission date into the generating AI, allowing the AI ​​to determine the publication priority. Specifically, this disclosure department collaborates with real-time stream processing infrastructures (such as Apache Flink or Spark Streaming) to manage the data's event time and processing time. This disclosure department sets time windows, prioritizing the push and distribution of data included in the latest window to the user interface (using WebSocket, etc.). For older data, it is moved to an archive repository and switched to a pull-based distribution that only displays data when a search query is issued. When using generative AI, this disclosure takes the timestamps and content of multiple data points as input, allowing the AI ​​to determine "what is the most popular topic to display now, considering the time-series context," and dynamically replaces the content of the headline news section based on this determination. This achieves the technical effect of consistently conveying the latest information to users in use cases such as news websites or stock price information where information freshness is paramount.

[0061] When publishing data, the publishing department can adjust the order of publication based on the relevance of the data. For example, the publishing department prioritizes publishing data with high relevance. For example, the publishing department postpones publishing data with low relevance. For example, the publishing department adjusts the order of publication in stages according to relevance. Thus, by adjusting the order of publication based on the relevance of the data, the publishing department can perform efficient publication. Some or all of the above processing in the publishing department can be performed using AI, or it can be performed without AI. For example, the publishing department can input the relevance of the data into the generating AI, allowing the generating AI to perform the adjustment of the publication order. Specifically, this publishing department is equipped with a hybrid recommendation engine that combines collaborative filtering and content-based filtering. This publishing department calculates the similarity (cosine similarity, etc.) between the feature vector of the content the user is currently browsing and the feature vectors of other content in the database, and lists them as "related reports" or "recommendations" in descending order of similarity. Furthermore, this disclosure inputs the user's behavioral history (click-throughs) within the drawing into an RNN (Recurrent Neural Network) to predict the category of information the user wants next and improves the display ranking of data belonging to that category. When using generative AI, this disclosure inputs a list of currently viewed articles and candidate articles, allowing the generative AI to infer "what articles people who read this article want to read next," and dynamically generates navigation links based on this result. This increases website browsing rates without interrupting the user's interest.

[0062] The Collaboration Department can analyze the sentiment of local governments and adjust collaboration methods based on the analyzed sentiment. For example, when a local government is relaxed, the Collaboration Department proposes detailed collaboration methods. When a local government is busy, the Collaboration Department proposes concise collaboration methods. When a local government is excited, the Collaboration Department proposes positive collaboration methods. Thus, by adjusting collaboration methods according to the sentiment of local governments, the Collaboration Department can conduct appropriate collaboration. Sentiment inference is achieved using sentiment inference functions, such as using a sentiment engine or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the Collaboration Department can be performed using AI, or not using AI. For example, the Collaboration Department can input the sentiment data of local governments into the generative AI, allowing the generative AI to perform adjustments to the collaboration methods. Specifically, this Collaboration Department takes voice data from emails, chats, or online meetings with the heads of local governments as input and performs sentiment analysis or voice sentiment parsing based on Natural Language Processing (NLP). This collaboration department infers the recipient's organizational "emotional state (atmosphere)" from emotional words such as "thank you," "dissatisfaction," and "anxiety" in the text, or from the tone and speed of the voice. For example, when judged as "busy (anxious)" at the end of the year, the department automatically sends a "quick start guide" that simplifies API collaboration steps, or proposes scripts to automate procedures. Conversely, when judged as "relaxed (have spare time)," proposals for long-term collaborative research or large-scale data integration projects are generated and sent. When using AI-generated proposals, the department uses the recipient's emotional state and the content of the proposal as input to generate a "polite and effective proposal email that considers the recipient's situation" and sends it. This achieves smooth communication and improves the completion rate and success rate of collaborative projects.

[0063] When collaborating, the Collaboration Department can select appropriate collaboration methods by referring to the past collaboration history of local governments. For example, the Collaboration Department proposes the best collaboration method based on the successful collaboration methods of local governments in the past. For example, the Collaboration Department avoids collaboration methods that have failed in the past. For example, the Collaboration Department analyzes the past collaboration history of local governments to select the best collaboration method. Thus, the Collaboration Department can select the best collaboration method by referring to the past collaboration history of local governments. Some or all of the above processes in the Collaboration Department can be performed using AI, or they can be performed without AI. For example, the Collaboration Department can input the past collaboration history of local governments into the AI ​​generator, allowing the AI ​​generator to select the collaboration method. Specifically, this Collaboration Department structures and accumulates data from all past collaboration projects (implementation content, period, budget, performance indicators (KPIs), and evaluations of responsible persons) in a database (case library). This Collaboration Department uses case-based reasoning (CBR) algorithms or similarity search to search for successful examples of local governments with similar characteristics (population size, industrial structure, past history) to the current candidate local governments. This collaboration department uses successful examples (best practices) found through searches as templates to develop collaboration plans tailored to the current autonomous system. When using generative AI, this department takes the target autonomous system's historical data and a list of success / failure examples as input, allowing the generative AI to propose plans that "avoid past failure patterns (e.g., interruptions due to insufficient budget) while adopting successful patterns (e.g., collaboration with tourism activities)." This enables data-driven decision-making that does not rely on rules of thumb, minimizing collaboration risk and maximizing effectiveness.

[0064] The Collaboration Department can analyze the sentiment of local governments and determine the priority of collaborations based on the analyzed sentiment. For example, when a local government is relaxed, the Collaboration Department prioritizes collaborations with high importance. Conversely, when a local government is busy, the Collaboration Department postpones collaborations with low importance. And when a local government is excited, the Collaboration Department prioritizes collaborations related to sentiment. Thus, by prioritizing collaborations based on the sentiment of local governments, the Collaboration Department can prioritize important collaborations. Sentiment inference can be achieved using sentiment inference functions, such as sentiment engines or generative AI. Generative AI can be, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the Collaboration Department can be performed using AI, or not. For example, the Collaboration Department can input the sentiment data of local governments into the generative AI, allowing the AI ​​to determine the priority of collaborations. Specifically, this Collaboration Department has a scoring engine that integrates with a CRM (Customer Relationship Management) system to dynamically manage the priority of tasks or projects for each local government. This collaboration department uses the content or frequency of inquiries from municipalities and the sentiment scores of those in charge (sentiment analysis results) as input variables to calculate the "urgency" and "importance" of each collaborative case. For example, when the other party shows "excitement (positive enthusiasm)," to capitalize on this momentum, the relevant projects are prioritized to the highest level, with resources (person in charge or computer resources) concentrated. Conversely, when the other party shows "confusion" or "anger," the priority of new proposals is lowered, and support or problem-solving tasks are prioritized instead. When using generative AI, this collaboration department inputs the status and sentiment data of all cases, allowing the generative AI to create a "ranking of municipalities to contact this week," seeking to optimize business or support activities.

[0065] When collaborating, the Collaboration Department can select the optimal collaboration method by considering the geographical location information of local governments. For example, the Collaboration Department proposes the optimal collaboration method based on the geographical location of local governments. For example, the Collaboration Department adjusts the collaboration method by considering the geographical characteristics of local governments. For example, the Collaboration Department selects the optimal collaboration method based on the geographical location information of local governments. Thus, the Collaboration Department can select the optimal collaboration method by considering the geographical location information of local governments. Some or all of the above processes in the Collaboration Department can be performed using AI, or they can be performed without AI. For example, the Collaboration Department can input the geographical location information of local governments into the generating AI, allowing the generating AI to perform the selection of collaboration methods. Specifically, this Collaboration Department uses GIS (Geographic Information System) data to vectorize and manage the location information (latitude and longitude), topographic data, transportation networks, climate zones, etc., of each government. This Collaboration Department uses spatial clustering algorithms (such as DBSCAN) to identify geographically adjacent groups of government agencies with common topics (e.g., a group of government agencies in a blizzard zone, a group of tourist government agencies in a coastal area). This Collaboration Department proposes wide-area collaboration projects for these groups (e.g., data sharing of wide-area tourist routes, mutual integration system of snow removal resources). When using generative AI, this collaboration department takes the location information of the target autonomous region and data from surrounding autonomous regions as input, and outputs ideas for "creating synergies through collaboration with neighboring autonomous regions," such as proposing specific measures like "developing a universal stamp rally application in the adjacent City A and Town B." This enables effective regional collaboration by leveraging geographical constraints or characteristics.

[0066] The parsing unit can analyze user emotions and adjust its analysis methods based on the analyzed emotions. For example, when the user is relaxed, the parsing unit performs detailed analysis. When the user is busy, the parsing unit performs brief analysis. When the user is excited, the parsing unit performs emotion-related analysis. Thus, by adjusting its analysis methods according to the user's emotions, the parsing unit can perform appropriate analysis. Emotion inference is achieved using emotion inference functions, such as by using an emotion engine or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the parsing unit can be performed using AI, or not. For example, the parsing unit can input the user's emotion data into the generative AI, allowing the generative AI to adjust the parsing method. Specifically, this parsing unit has an orchestrator that dynamically changes the structure of the data analysis pipeline. This parsing unit receives the user's emotional state (e.g., high curiosity, busy) as input parameters and selects the analysis module to be executed. When users are in a "relaxed" (curious) state, this analysis department not only runs basic statistics but also performs computationally intensive detailed analysis modules at full speed, such as correlation analysis, causal inference, and time series forecasting, yielding multi-faceted insights. Conversely, when users are "busy," this analysis department selects a lightweight analysis path that only calculates the main KPIs (Key Performance Indicators) and immediately returns the results. When using generative AI, this analysis department takes the dataset and user sentiment as input, allowing the AI ​​to suggest "entry points for in-depth analysis that the user might be interested in," and automatically generates and executes SQL queries or Python code that match those entry points (e.g., trend analysis by era). Thus, analysis results are provided with depth and speed that meet the user's psychological needs.

[0067] During parsing, the parsing unit can adjust the level of detail based on the importance of the data. For example, the parsing unit may analyze highly important data in detail, or analyze less important data briefly. Alternatively, the parsing unit may adjust the level of detail in stages based on importance. Thus, by adjusting the level of detail based on the importance of the data, the parsing unit can perform efficient parsing. Some or all of the above processing in the parsing unit can be performed using AI, or it can be performed without AI. For example, the parsing unit can input the importance of the data into the generating AI, allowing the AI ​​to adjust the level of detail of the parsing. Specifically, this parsing unit calculates an importance score based on the metadata or preliminary scan results of the input data. For highly important data (e.g., data containing outliers, core data related to business decisions), this parsing unit applies high-precision models using deep learning (e.g., Transformer-based time series prediction models) or computationally intensive algorithms such as Monte Carlo simulations to obtain parsing results with fewer errors. For low-importance data, lightweight algorithms such as linear regression or moving averages are applied to save computational resources (GPU / CPU time). When using generative AI, this analysis unit takes the data summary and importance as input, allowing the AI ​​to determine "what is the best statistical method to apply to this data," and then calls functions from analysis libraries (Pandas, Scikit-learn, etc.) based on this determination. This optimizes the allocation of limited computing resources, maximizing the overall cost-effectiveness of the system.

[0068] During parsing, the parsing unit can apply different parsing algorithms based on the data category. For example, the parsing unit applies image parsing algorithms to image data. For example, the parsing unit applies audio parsing algorithms to audio data. For example, the parsing unit applies video parsing algorithms to video data. Thus, by applying different parsing algorithms according to the data category, the parsing unit can perform appropriate parsing. Some or all of the above processing in the parsing unit can be performed using AI, or it can be performed without AI. For example, the parsing unit can input the data category into the generating AI, allowing the generating AI to execute the application of the parsing algorithm. Specifically, this parsing unit manages a model zoo, dynamically loading the best learned models according to the data modality (image, text, speech, tabular data). In the case of image data, this parsing unit uses CNNs (such as ResNet) for feature extraction or object recognition. In the case of text data, this parsing unit uses LLMs such as BERT or GPT for sentiment analysis or summarization. In the case of time-series numerical data, this parsing unit uses ARIMA or LSTM for trend prediction. When using generative AI, this analysis department employs an AutoML (Automated Machine Learning) approach. It inputs sample and category information of the data into the generative AI, which then generates Python code (using PyTorch or TensorFlow) to analyze the data. This code is then executed in a sandbox environment, allowing for flexible parsing of even unknown data formats. This enables the comprehensive analysis of diverse data on a single platform.

[0069] The parsing unit can analyze user emotions and determine the parsing priority based on the analyzed emotions. For example, when the user is relaxed, the parsing unit prioritizes high-importance parsing. When the user is busy, the parsing unit postpones low-importance parsing. When the user is excited, the parsing unit prioritizes emotion-related parsing. Thus, by determining the parsing priority based on the user's emotions, the parsing unit can prioritize important parsing. Emotion inference is achieved using emotion inference functions, such as an emotion engine or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the parsing unit can be performed using AI, or not. For example, the parsing unit can input the user's emotion data into the generative AI, allowing the generative AI to determine the parsing priority. Specifically, this parsing unit collaborates with a job scheduler that manages parsing jobs to dynamically control the execution priority of each job. This analysis department monitors users' emotional states in real time (e.g., anxiety, anticipation). When a user is eager for results (anxiety), the priority of the analysis job requested by that user is set to the highest level, other background jobs are suspended, and resources are allocated (preemption control). On the other hand, if the user is in a relaxed waiting state, adjustments are made, such as shifting the job to nighttime batch processing. When using generative AI, this analysis department uses the list of pending jobs and the user's emotional state as input to have the generative AI create a "job execution schedule that maximizes user satisfaction," and then executes the processing according to that schedule. This balances system resource leveling with improved user experience.

[0070] During parsing, the parsing department can determine the parsing priority based on the data's submission date. For example, the parsing department prioritizes parsing the most recently submitted data. Alternatively, it may postpone parsing older data. Furthermore, the parsing department can periodically adjust the parsing priority based on the submission date. Thus, by determining the parsing priority based on the data's submission date, the parsing department can perform efficient parsing. Some or all of the above processing in the parsing department can be performed using AI, or it can be performed without AI. For example, the parsing department can input the data's submission date into the generation AI, allowing the generation AI to determine the parsing priority. Specifically, this parsing department incorporates priority control logic that uses data freshness as the evaluation function. The parsing department assigns higher weight to data that has been recently generated and feeds it into the real-time parsing pipeline (stream processing). Data that has been stored for a certain period is accumulated and then processed by the batch parsing pipeline (Hadoop / Spark, etc.). When using generative AI, this parsing unit takes the data's timestamp and content variability as input, allowing the AI ​​to determine whether the data needs to be parsed immediately or can be parsed later, and allocates a processing path (hot path / cold path) based on the result. This effectively prevents parsing delays for data with real-time value, such as market prices or disaster information.

[0071] During parsing, the parsing unit can adjust the parsing order based on the data's relevance. For example, it prioritizes parsing data with high relevance, postpones parsing data with low relevance, or adjusts the parsing order in stages based on relevance. Thus, by adjusting the parsing order based on data relevance, the parsing unit can perform efficient parsing. Some or all of the above processing in the parsing unit can be performed using AI, or it can be performed without AI. For example, the parsing unit can input the data's relevance into the generating AI, allowing the AI ​​to adjust the parsing order. Specifically, this parsing unit manages the dependencies or semantic relationships between data as a dependency graph. When the parsing result of data A is necessary for the parsing of data B, or when data A and data B are related to the same topic (e.g., a specific event), this parsing unit groups them and schedules them for sequential parsing. Using algorithms such as topological sorting, it resolves dependencies while processing data in order to improve cache hit rate. When using generative AI, this parsing department takes the metadata of the unprocessed dataset as input, instructs the generative AI to "cluster data with deep contextual relevance and propose an efficient processing order," and reconstructs the parsing queue based on this proposal. This reduces the overhead of context switching and improves parsing efficiency.

[0072] The information delivery unit can analyze the user's emotions and adjust the display method of the provided information based on the analyzed emotions. For example, when the user is relaxed, the information delivery unit provides a detailed display method. When the user is busy, the information delivery unit provides a concise display method. When the user is excited, the information delivery unit provides a visually stimulating display method. Thus, by adjusting the display method of the provided information according to the user's emotions, the information delivery unit can provide appropriate display. Emotion inference is achieved using emotion inference functions, such as using an emotion engine or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the information delivery unit can be performed using AI, or not, for example. For example, the information delivery unit can input the user's emotion data into the generative AI, allowing the generative AI to adjust the display method. Specifically, this information delivery unit has an adaptive user interface (UI) generation system that changes the attributes of screen components (widgets) in real time according to the user's emotional state. When the user is "relaxed," this information delivery unit uses a grid layout to display information in a comprehensive manner and reduces the font size to increase information density. When the user is "busy," this department uses a card-based layout, switching to a "summary mode" that displays only important values ​​or conclusions in large sizes. When the user is "excited," dynamic presentations with animation effects or vibrant colors are added. When using AI-generated content, this department takes the data to be displayed and the user's emotions as input, allowing the AI ​​to generate "HTML / CSS code for a dashboard that matches the user's mood," which is then rendered in the browser. This achieves information delivery optimized for the user's psychological characteristics.

[0073] The provision department can adjust the level of detail provided based on the importance of the information. For example, it can provide detailed information about high importance, or concise information about low importance. It can also adjust the level of detail in stages based on importance. Thus, by adjusting the level of detail based on the importance of the information, the provision department can provide information efficiently. Some or all of the above processing in the provision department can be performed using AI, or it can be performed without AI. For example, the provision department can input the importance of the information into the AI ​​generator, allowing the AI ​​to adjust the level of detail provided. Specifically, this provision department has the function of automatically generating a hierarchical structure (drill-down structure) of information based on its importance score. For high-importance information (e.g., emergency alerts, unmet KPIs), the provision department directly displays detailed charts or analytical comments on the homepage. For low-importance information, it is configured behind a collapsed menu or link, and detailed content is only displayed when the user performs an explicit action (progressive disclosure). When using AI-generated text, this provider takes the raw data and importance as input, and instructs the AI ​​to generate "summary texts based on importance (100 words, 300 words, 1000 words)" for display. The appropriate text length is then selected based on the available space or importance on the UI. This prevents information overload on the screen and reduces the risk of users missing important information.

[0074] When providing information, the providing unit can apply different providing algorithms based on the category of the information. For example, the providing unit applies an image display algorithm to image information. For example, the providing unit applies a voice playback algorithm to voice information. For example, the providing unit applies a video playback algorithm to video information. Thus, by applying different providing algorithms according to the category of information, the providing unit can provide appropriate information. Some or all of the above processing in the providing unit can be performed using AI, or it can be performed without AI. For example, the providing unit can input the category of information into the generating AI, allowing the generating AI to execute the application of the providing algorithm. Specifically, this providing unit has a multimodal output generation engine that selects the optimal representation layer according to the form of the data. In the case of text information, this providing unit uses Natural Language Generation (NLG) technology to shape it into an easy-to-read article and uses a Text-to-Speech (TTS) engine as needed for voice provision. In the case of numerical data, this providing unit analyzes the characteristics of the data (time series, distribution, comparison), automatically selects the optimal chart type (line chart, bar chart, scatter plot, heatmap) and draws it. When using generative AI, this provider can also drive text-to-video (TTO) AI with text data as input to automatically generate and provide explanatory videos. Thus, users can enjoy information in the most intuitive and understandable form without needing to be aware of the data's format.

[0075] The delivery unit analyzes the user's emotions and determines the priority of information delivered based on the analyzed emotions. For example, when the user is relaxed, the delivery unit prioritizes information of high importance. When the user is busy, the delivery unit postpones information of low importance. When the user is excited, the delivery unit prioritizes information related to emotions. Thus, by prioritizing information based on the user's emotions, the delivery unit can prioritize important information. Emotion inference is achieved using emotion inference functions, such as using an emotion engine or generative AI. Generative AI is, for example, text generation AI (e.g., LLM) or multimodal generation AI, but is not limited to this example. Some or all of the above processing in the delivery unit can be performed using AI, or not, without AI. For example, the delivery unit can input the user's emotion data into the generative AI, allowing the generative AI to determine the priority of information. Specifically, this delivery unit has a notification manager that controls push notifications, email distribution, and in-app feed display. This manager takes the user's emotional state as input and optimizes the timing and order of information distribution. When the user is "relaxed," this delivery unit notifies the user of long reports or analysis results (of high importance) that should be read carefully. When a user is "busy," this department retains notifications (in silent mode) except for high-urgency alerts, and performs a "summary distribution" of notifications later. When using AI-generated messages, this department takes the unread message list and user sentiment as input, allowing the AI ​​to select the "3 best messages to convey to the current user," and then sends notifications based on these selections. This achieves intelligent information delivery that takes into account the user's tolerance for interruptions.

[0076] When providing information, the delivery department can determine the priority of delivery based on the submission time of the information. For example, the delivery department may prioritize the delivery of the most recently submitted information. It may also postpone older information. Furthermore, the delivery department may adjust the delivery priority period according to the submission time. Thus, by determining the delivery priority based on the submission time of the information, the delivery department can provide information efficiently. Some or all of the above processing in the delivery department may be performed using AI, or it may not. For example, the delivery department can input the submission time of the information into the generation AI, allowing the generation AI to determine the delivery priority. Specifically, this delivery department uses an algorithm that scores the "freshness" of information to place new information at the top in a timeline-style UI. Based on the difference between the information's generation time and the current time, this delivery department performs visual weighting, such as lowering the display position or reducing the display size over time. In addition, this delivery department also has the following logic: even if the information is old, if its relevance to the current event reappears (resurrection), the priority is recalculated and it is displayed at the top. When using generative AI, this provision department takes newly arrived information and past archives as input, allowing the generative AI to generate "an explanation of the latest news based on past context," integrating and providing new and old information. This enables the provision of information that goes beyond a simple chronological order, providing contextualized information.

[0077] When providing information, the delivery department can adjust the order of information delivery based on the relevance of the information. For example, the delivery department may prioritize providing highly relevant information. For example, it may postpone providing less relevant information. For example, it may adjust the delivery order in stages based on relevance. Thus, by adjusting the delivery order based on the relevance of information, the delivery department can provide information efficiently. Some or all of the above processing in the delivery department may be performed using AI, or it may not. For example, the delivery department can input the relevance of information into the generating AI, allowing the generating AI to adjust the delivery order. Specifically, this delivery department is equipped with a "next action prediction model" that predicts the information the user should view next based on the content the user is currently viewing or their recent browsing history. This delivery department uses semantic similarity between content (distance in vector space) or the browsing migration probability of other users (Markov chains, etc.) to list highly relevant information as "content to read together" suggestions in the sidebar or at the bottom of the article. When using generating AI, this delivery department takes the currently displayed content and candidate content in the database as input, allowing the generating AI to infer "which material is the most suitable to supplement this information," and generates a recommendation bar based on the result. This ensures that the user's thought process is not interrupted and supports smooth information collection.

[0078] The system described in this embodiment is not limited to the examples above and can be modified in various ways, as described below. Specifically, the system architecture is designed as a loosely coupled design independent of specific hardware or software structures, making it easily migrateable to hybrid architectures combining local servers, public clouds (AWS, Azure, GCP, etc.), and edge devices (IoT sensors, smartphones), or serverless architectures. Furthermore, since each functional unit (collection, extraction, disclosure, collaboration, parsing, provision) is installed as an independent microservice, specific functions can be replaced simply with third-party APIs or the latest AI models (e.g., higher-performance LLM or image generation models). Moreover, this system has the flexibility to accommodate future technological expansions, such as incorporating blockchain technology to ensure data authenticity or utilizing quantum computers to accelerate optimization computations.

[0079] The data collection unit monitors the user's health status and can adjust the timing of data collection based on this status. For example, data collection can be postponed when the user is fatigued, and collected immediately when the user is healthy. Thus, by adjusting the timing of data collection according to the user's health status, the collection unit can reduce the user's burden. Specifically, this collection unit collaborates with wearable devices such as smartwatches or fitness trackers via API to acquire vital sign data such as heart rate, blood pressure, sleep duration, activity level (steps), and electrical skin activity (EDA) in real time. This collection unit has a health status estimation model (e.g., a random forest or LSTM model that regresses to predict fatigue or stress levels) using this vital sign data as input, quantifying the user's status. When the estimated fatigue level exceeds a specified threshold, this collection unit suppresses proactive data collection actions such as push notifications or questionnaire requests, switching only to passive log collection in the background. Conversely, when it is determined that the user is sufficiently rested and in good condition, this collection unit delegates data provision with complex operations. This achieves a health-conscious system action that considers the user's physical and physiological state.

[0080] The extraction department evaluates the reliability of data and can also determine the priority of extracted elements based on reliability. For example, important elements can be extracted first from data with high reliability, while data with low reliability can be extracted later. Thus, by prioritizing extraction based on data reliability, the extraction department can provide more accurate information. Specifically, this extraction department has a reliability evaluation engine that calculates a credibility score for each data point based on the domain authority of the data source, the author's reputation, the diffusion pattern on social media, and the results of comparison with fact-checking databases. For data with high credibility scores (such as publications from public institutions or peer-reviewed papers), the extraction department prioritizes detailed information extraction processing (full-text analysis, structuring of numerical data) and integrates it into a knowledge graph. On the other hand, for data with low credibility scores (such as blogs with unclear sources or rumor-based social media submissions), the processing priority is reduced, or it is isolated by assigning a "verification required" flag. When using generative AI, the extraction department takes the content of the report as input and allows the generative AI to analyze "the logical coherence and evidence of the claims contained in the report," determining the risk of illusion or fake news. This serves as a technical guarantee for the quality and reliability of the information provided by the system.

[0081] The visual appeal of the data evaluated by the disclosure department can also be used to adjust the display method of the disclosed data based on its appeal. For example, visually appealing data is displayed in detail, while less appealing data is displayed concisely. Thus, by adjusting the display method based on the visual appeal of the data, the disclosure department can attract user interest. Specifically, this disclosure department applies an aesthetic assessment model (e.g., NIMA: Neural Image Assessment) to image or video data, quantifying the "Visual Appeal Score" from the perspectives of composition, color, lighting, and sharpness. For images with high scores (e.g., photos of beautiful landscapes or delicious-looking food), this disclosure department automatically selects a layout that increases the screen's occupancy and displays the image at a high resolution. For images with low scores (e.g., shaky photos or dark images), they are displayed as thumbnails or after applying image correction AI (super-resolution, brightness adjustment). When using generative AI, this disclosure department uses text content as input to allow the generative AI to generate "eye-catching slogans" or "attractive and eye-catching images," enhancing the appeal of the content. This improves user visual satisfaction, contributing to longer dwell time or higher conversion rates.

[0082] The Collaboration Department analyzes the economic situation of local governments and can adjust collaboration methods accordingly. For example, it proposes proactive collaboration methods for economically well-off local governments and cost-effective methods for those facing economic difficulties. Thus, by adjusting collaboration methods based on the economic situation of local governments, the Collaboration Department can achieve effective collaboration. Specifically, the department possesses a financial analysis model that collects and analyzes publicly available financial statements (revenue, expenditure, fiscal strength index, actual public debt ratio, etc.) and open data from each local government to estimate its fiscal soundness and investment capacity. For local governments with high fiscal strength indices and investment capacity, the department proposes high-value-added collaboration plans that include the latest AI implementation or large-scale infrastructure development. On the other hand, for local governments facing financial difficulties, the department proposes low-cost and efficient collaboration plans that include the use of open-source software, pay-as-you-go cloud services, or support for subsidy applications. When using generative AI, the department takes the local government's financial report as input and allows the AI ​​to determine the "optimal DX implementation roadmap within budget constraints," automatically generating proposals with high feasibility. Therefore, we can build collaborative relationships that are in line with the actual situation of the autonomous region.

[0083] The parsing department considers the data update frequency and can determine the parsing priority based on it. For example, frequently updated data is parsed first, while data with low update frequency can be parsed later. Thus, by prioritizing parsing based on data update frequency, the parsing department can quickly provide the latest information. Specifically, this parsing department has an update prediction engine that statistically analyzes the update history of each data source and models the probability distribution of update intervals (e.g., Poisson distribution). For data sources with high update frequency (e.g., stock prices, traffic information, SNS trends), the parsing department sets a short polling interval or keeps a listener running to receive push notifications via WebSocket, starting the parsing pipeline (event-driven processing) as soon as the data arrives. For data sources with low update frequency (e.g., demographics, annual reports), it uses periodic batch processing (e.g., once a day). When using generative AI, the parsing department takes the data change patterns as input, allowing the generative AI to predict "when the next major change will occur," and allocates resources accordingly. This ensures real-time information while suppressing the consumption of useless computing resources.

[0084] The data collection department analyzes user emotions and can determine the types of data to collect based on these emotions. For example, when a user is relaxed, data of interest can be collected. When a user is busy, only the minimum necessary data can be collected. Thus, by determining the types of data to collect based on user emotions, the collection department can reduce the user's burden. Specifically, this collection department maintains a correlation matrix between the output of the emotion recognition model (emotion vector) and data categories (entertainment, news, business contacts, advertising, etc.). When the user is in a "relaxed" state, the collection department increases the collection ratio of "exploratory data" that evokes user interest, such as highly entertaining video data or blog posts related to their interests. When the user is in a "busy (high-stress)" state, the collection department only filters "essential data" required for business execution or urgent contacts, stopping the collection of information that becomes noise. When using generative AI, the collection department uses the user's emotions and the current context as input, allowing the generative AI to infer "the type of content that is comfortable for the current user," exploring and collecting data from the Web that matches that type. This achieves information filtering that aligns with the user's psychological receptivity.

[0085] The extraction unit analyzes user sentiment and can adjust the level of detail in the extracted data based on this sentiment. For example, detailed data can be extracted when the user is relaxed, while concise data can be extracted when the user is busy. Thus, by adjusting the level of detail in the extracted data according to the user's sentiment, the extraction unit can provide information that responds to the user's needs. Specifically, this extraction unit dynamically changes the parameters controlling the granularity of information in the extraction process based on the user's sentiment score. When the user is judged to be "relaxed" and has a high information absorption capacity, the extraction unit sets a low compression rate for the text summary (long text summary), extracting many tags from the image, including background details. When the user is "busy," the extraction unit compresses the text down to "3 items," extracting only the main objects from the image. When using generative AI, the extraction unit takes metadata and user sentiment as input, instructs the generative AI to "consider the user's cognitive load and summarize with the optimal amount of information," and uses the output text. This allows the extraction unit to provide information with the optimal information density in response to the user's situation.

[0086] The disclosure department analyzes user emotions and can adjust the timing of data disclosure based on these emotions. For example, data can be disclosed immediately when the user is relaxed, and later when the user is busy. Thus, by adjusting the timing of disclosure according to user emotions, the disclosure department improves user convenience. Specifically, this disclosure department has an intelligent distribution scheduler that controls the timing of notifications or data display to users. This scheduler monitors the user's emotional state in real time, detecting the moment the user enters a "relaxed" state (e.g., when work ends or during rest time after returning home), and then compiles and discloses accumulated data (news, recommendations, etc.). During periods when the user is in a "focused" or "busy" state, data is buffered (temporarily stored) to avoid interruption. When using generative AI, this disclosure department takes the user's lifestyle and emotional history as input, allowing the generative AI to predict the "time period when the user is most receptive to information," and schedules distribution accordingly. This achieves distribution control that does not hinder user experience while improving information delivery rates.

[0087] The Collaboration Department analyzes the sentiment within local governments and can determine the types of collaborative projects based on this analysis. For example, when local governments are relaxed, long-term projects can be proposed. When local governments are busy, short-term projects can be proposed. Thus, by determining project types based on the sentiment of local governments, the Collaboration Department can achieve effective collaboration. Specifically, this Collaboration Department possesses an organizational sentiment analysis model that infers the "atmosphere" or "morale" of the local government organization from textual analysis of publicly available meeting minutes or communication logs with responsible persons. When the organization is determined to be in a "relaxed (stable, positive)" state, this Collaboration Department proposes large-scale projects requiring resources and time, such as smart city initiatives or long-term tourism branding. When the organization is in a "busy (chaotic, exhausted)" state, this Collaboration Department proposes small-scale projects (Quick Wins) that yield results in a short period, such as business efficiency improvements through RPA (Robotic Process Automation) or immediate event customer acquisition support. When using generative AI, this Collaboration Department takes the current issues and organizational sentiment of the local government as input, allowing the generative AI to generate "project solutions that can be executed without strain and are effective under the current organizational structure." This enables realistic collaboration that matches the autonomous body's capacity to accept new ideas.

[0088] The analysis department analyzes user emotions and adjusts the delivery method of the analysis results based on the analyzed user emotions. For example, detailed analysis results can be provided when the user is relaxed, while concise analysis results can be provided when the user is busy. Thus, by adjusting the delivery method of the analysis results according to the user's emotions, the analysis department can provide information that responds to the user's needs. Specifically, this analysis department passes the user's emotional parameters to the formatter that determines the output format of the analysis results. When the user is "relaxed," this analysis department provides a detailed CSV file containing the raw data, multi-angle charts, and a complete package containing a detailed investigation report. When the user is "busy," this analysis department provides a single slide extracting only the key insights, or a short text message simply stating the conclusions. When using generative AI, this analysis department takes complex analysis results (numerical data) and user emotions as input, instructs the generative AI to "create an explanatory text that matches the user's psychological state in tone and style (e.g., an encouraging tone, a business-like tone)," and presents this output to the user. Thus, in addition to the value of the data itself, by optimizing the way it is communicated to the user, the department aims to promote the effective use of data.

[0089] The following is a brief description of the processing flow in this implementation. Specifically, the data processing pipeline in this system is defined as a workflow consisting of a series of sequential and parallel processing steps, each executed as a stateless function or microservice. Overall control is orchestrated by a workflow engine (such as Apache Airflow or AWS Step Functions), and the completion status, error handling, and retry handling of each step are rigorously managed. The steps shown below represent the logical data flow; in an actual installation, the aim is to increase throughput through pipelined processing or reduce latency through distributed processing.

[0090] Step 1: The Collection Department collects data. This department collects data from sources on the internet, such as news websites, social networking sites (SNS), blogs, Yahoo!, or messenger applications. Specifically, web crawlers send HTTP requests based on a list of target URLs to obtain HTML or JSON data and save it to a data lake (S3, etc.). Step 2: The Extraction Department extracts elements from the collected data. This department extracts elements such as keywords, feature values, and patterns from the collected data. It uses natural language processing (NLP) techniques to extract keywords from text data. Specifically, it reads the saved raw data, applies NLP or image recognition models to generate structured metadata (entities, categories, feature vectors), and stores it in a database. Step 3: The Publication Department publishes the extracted data on the internet. This department publishes the data using platforms such as websites, social networking sites (SNS), and databases. It publishes the data on websites, making it accessible to everyone. Specifically, it distributes structured data as web pages or JSON responses to external clients via CMS or API servers. Step 4: The Collaboration Department is responsible for collaborating with local governments. The Collaboration Department, for example, implements joint projects with local governments to collect and publish data for showcasing the region's appeal. The Collaboration Department can collaborate with local governments via APIs to share data. Specifically, it synchronizes data with the local government's system through a secure data exchange platform, establishing a mutually usable state. Step 5: The Analysis Department analyzes the collected data using AI. The Analysis Department, for example, uses technologies such as machine learning, deep learning, and natural language processing to analyze the data. The Analysis Department can use AI to analyze the collected data and extract its features. Specifically, it performs batch or streaming analysis on accumulated big data to calculate trend predictions or clustering results. Step 6: The Provision Department provides the analysis results. The Provision Department provides the analysis results, for example, in the form of reports, charts, and statistical data. The Provision Department can publish the analysis results on a website, making them accessible to everyone. Specifically, it visualizes the analysis results through BI tools or dashboards, providing prompts in a way that supports user decision-making.

[0091] The specific processing unit 290 sends 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 voice representing the user's input to the result of the specific processing. The control unit 46A sends the voice data representing the user's 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 voice data.

[0092] Data generation model 58 is what is known as generative AI (Artificial Intelligence). An example of data generation model 58 includes ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> Generative AI, such as data generation model 58, is obtained by deep learning through a neural network. The data generation model 58 is input with a prompt containing instructions, and with inference data such as speech data representing speech, text data representing text, and image data representing images (e.g., data of still images or data of moving images). The data generation model 58 infers the inference data based on the instructions shown in the prompt and outputs the inference result in one or more data forms such as speech data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the above-described specific processing while using the data generation model 58. The data generation model 58 can also be a fine-tuned model to output inference results from prompts without instructions; in this case, the data generation model 58 can output inference results from prompts without instructions. The data processing device 12, etc., includes multiple data generation models 58, including AI other than generative AI. AI other than generative AI includes, but is not limited to, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes. Furthermore, AI can also act as an AI agent. Additionally, when the processing described above is performed by AI, this processing may be partially or entirely performed by AI, but is not limited to this example. Moreover, processing performed by AI, including generative AI, can be replaced by rule-based processing, and vice versa.

[0093] Furthermore, the processing performed by the aforementioned data processing system 10 is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but it can also be executed jointly by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the smart device 14 or external devices, and the smart device 14 acquires or collects information required for processing from the data processing device 12 or external devices.

[0094] Each of the aforementioned elements, including a collection unit, extraction unit, disclosure unit, collaboration unit, analysis unit, and provision unit, is implemented, for example, in at least one of the smart device 14 and the data processing device 12. For example, the collection unit collects data from a data source on the Internet via the communication I / F 44 of the smart device 14. The extraction unit extracts elements from the collected data, for example, via a specific processing unit 290 of the data processing device 12. The disclosure unit discloses the extracted data on the Internet, for example, via the control unit 46A of the smart device 14. The collaboration unit, for example, is responsible for collaboration with local governments, via the specific processing unit 290 of the data processing device 12. The analysis unit, for example, uses AI to analyze the collected data via the specific processing unit 290 of the data processing device 12. The provision unit, for example, provides the analysis results via the control unit 46A of the smart device 14. The correspondence between each unit and the device or control unit is not limited to the above example and can be varied.

[0095] [Second Implementation] Figure 3 An example of the configuration of the data processing system 210 according to the second embodiment is shown.

[0096] like Figure 3 As shown, the data processing system 210 includes a data processing device 12 and smart glasses 214. One example of the data processing device 12 is a server.

[0097] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, 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. An example of the network 54 includes a WAN and / or LAN, etc.

[0098] 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, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0099] Microphone 238 receives user commands by receiving the user's voice. Microphone 238 captures the user's voice, converts the captured sound into speech data, and outputs it to processor 46. Speaker 240 outputs sound according to commands from processor 46.

[0100] Camera 42 is a small digital camera equipped with an optical system such as 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, used to capture the user's surroundings (e.g., the shooting range defined by an angle of view equivalent to the field of vision of an average healthy person).

[0101] Communication I / F 44 is connected to network 54. Communication I / F 44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F 44 and 26 is performed in a secure state.

[0102] Figure 4 An example of the main functions of the data processing device 12 and the smart glasses 214 is shown. Figure 4 As shown, in the data processing device 12, specific processing is performed by the processor 28. The memory 32 stores a specific processing program 56.

[0103] The processor 28 reads a specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is implemented by the processor 28 operating as a specific processing unit 290 based on the specific processing program 56 executed on the RAM 30.

[0104] The memory 32 stores a data generation model 58 and an emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290. The specific processing unit 290 can use the emotion-specific model 59 to infer the user's emotions and perform specific processing based on the user's emotions. The emotion inference function using the emotion-specific model 59 (emotion-specific function) includes various inferences and predictions related to the user's emotions, but is not limited to this example. Furthermore, emotion inference and prediction also include, for example, emotion analysis (analysis).

[0105] In the smart glasses 214, specific processing is performed by the processor 46. A specific processing program 60 is stored in the memory 50. The processor 46 reads the specific processing program 60 from the memory 50 and executes the read specific processing program 60 on the RAM 48. Specific processing is implemented by the processor 46 operating as a control unit 46A based on the specific processing program 60 executed on the RAM 48. Furthermore, the smart glasses 214 may also have the same data generation model and emotion-specific model as the data generation model 58 and the emotion-specific model 59, and use these models to perform the same processing as the specific processing unit 290.

[0106] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains the processing results (prediction results, etc.) using the data generation model 58 by communicating with the server device that has the data generation model 58. In addition, the data processing device 12 may be a server device or a user-owned terminal device (e.g., a mobile phone, robot, home appliance, etc.).

[0107] The specific processing unit 290 sends 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 voice input representing the user's input to the specific processing result. The control unit 46A sends the voice data representing the user's 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.

[0108] The data generation model 58 is a so-called generative AI. An example of the data generation model 58 includes generative AIs such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with a prompt containing instructions, and also with inference data such as speech data representing speech, text data representing text, and image data representing images (e.g., data of still images or data of moving images). The data generation model 58 infers the inference data based on the instructions shown in the prompt and outputs the inference result in one or more data forms such as speech data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the aforementioned specific processing while using the data generation model 58. The data generation model 58 can also be a fine-tuned model to output inference results from prompts that do not contain instructions; in this case, the data generation model 58 is able to output inference results from prompts that do not contain instructions. The data processing device 12, etc., includes various data generation models 58, which include AI other than generative AI. These AIs include, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and are capable of various processing methods, but are not limited to these examples. Furthermore, AI can also be an AI agent. Additionally, when the processing described above is performed by AI, this processing may be partially or entirely performed by AI, but is not limited to these examples. Moreover, processing performed by AI including generative AI can be replaced by rule-based processing, and vice versa.

[0109] The data processing system 210 of the second embodiment performs the same processing as the data processing system 10 of the first embodiment. The processing performed by the data processing system 210 is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but it can also be executed jointly by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the smart glasses 214 or external devices, and the smart glasses 214 acquires or collects information required for processing from the data processing device 12 or external devices.

[0110] Each of the aforementioned elements, including a collection unit, extraction unit, disclosure unit, collaboration unit, analysis unit, and provision unit, is implemented, for example, in at least one of the smart glasses 214 and the data processing device 12. For example, the collection unit collects data from a data source on the Internet via the communication I / F 44 of the smart glasses 214. The extraction unit extracts elements from the collected data, for example, via a specific processing unit 290 of the data processing device 12. The disclosure unit discloses the extracted data on the Internet, for example, via the control unit 46A of the smart glasses 214. The collaboration unit, for example, is responsible for collaboration with local governments, via the specific processing unit 290 of the data processing device 12. The analysis unit, for example, uses AI to analyze the collected data via the specific processing unit 290 of the data processing device 12. The provision unit, for example, provides the analysis results via the control unit 46A of the smart glasses 214. The correspondence between each unit and the device or control unit is not limited to the above example and can be varied.

[0111] [Third Implementation] Figure 5 An example of the configuration of the data processing system 310 according to the third embodiment is shown.

[0112] like Figure 5 As shown, the data processing system 310 includes a data processing device 12 and a head-mounted terminal 314. One example of the data processing device 12 is a server.

[0113] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, 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. An example of the network 54 includes a WAN and / or LAN, etc.

[0114] The head-mounted 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, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0115] Microphone 238 receives user commands by receiving the user's voice. Microphone 238 captures the user's voice, converts the captured sound into speech data, and outputs it to processor 46. Speaker 240 outputs sound according to commands from processor 46.

[0116] Camera 42 is a small digital camera equipped with an optical system such as 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, used to capture the user's surroundings (e.g., the shooting range defined by an angle of view equivalent to the field of vision of an average healthy person).

[0117] Communication I / F 44 is connected to network 54. Communication I / F 44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F 44 and 26 is performed in a secure state.

[0118] Figure 6 An example of the main functions of the data processing device 12 and the head-mounted terminal 314 is shown. Figure 6 As shown, in the data processing device 12, specific processing is performed by the processor 28. The memory 32 stores a specific processing program 56.

[0119] The processor 28 reads a specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is implemented by the processor 28 operating as a specific processing unit 290 based on the specific processing program 56 executed on the RAM 30.

[0120] The memory 32 stores a data generation model 58 and an emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290. The specific processing unit 290 can use the emotion-specific model 59 to infer the user's emotions and perform specific processing based on the user's emotions. The emotion inference function using the emotion-specific model 59 (emotion-specific function) includes various inferences and predictions related to the user's emotions, but is not limited to this example. Furthermore, emotion inference and prediction also include, for example, emotion analysis (analysis).

[0121] In the head-mounted terminal 314, specific processing is performed by the processor 46. A specific program 60 is stored in the memory 50. The processor 46 reads the specific program 60 from the memory 50 and executes the read specific program 60 on the RAM 48. Specific processing is implemented by the processor 46 operating as a control unit 46A based on the specific program 60 executed on the RAM 48. Furthermore, the head-mounted terminal 314 may also have the same data generation model and emotion-specific model as the data generation model 58 and the emotion-specific model 59, and use these models to perform the same processing as the specific processing unit 290.

[0122] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains the processing results (prediction results, etc.) using the data generation model 58 by communicating with the server device that has the data generation model 58. In addition, the data processing device 12 may be a server device or a user-owned terminal device (e.g., a mobile phone, robot, home appliance, etc.).

[0123] The specific processing unit 290 sends the result of the specific processing to the head-mounted terminal 314. In the head-mounted terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires voice representing the user's input to the specific processing result. The control unit 46A sends the voice data representing the user's 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.

[0124] The data generation model 58 is a so-called generative AI. An example of the data generation model 58 includes generative AIs such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with a prompt containing instructions, and also with inference data such as speech data representing speech, text data representing text, and image data representing images (e.g., data of still images or data of moving images). The data generation model 58 infers the inference data based on the instructions shown in the prompt and outputs the inference result in one or more data forms such as speech data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the aforementioned specific processing while using the data generation model 58. The data generation model 58 can also be a fine-tuned model to output inference results from prompts that do not contain instructions; in this case, the data generation model 58 is able to output inference results from prompts that do not contain instructions. The data processing device 12, etc., includes various data generation models 58, which include AI other than generative AI. These AIs include, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and are capable of various processing methods, but are not limited to these examples. Furthermore, AI can also be an AI agent. Additionally, when the processing described above is performed by AI, this processing may be partially or entirely performed by AI, but is not limited to these examples. Moreover, processing performed by AI including generative AI can be replaced by rule-based processing, and vice versa.

[0125] The data processing system 310 of the third embodiment performs the same processing as the data processing system 10 of the first embodiment. The processing performed by the data processing system 310 is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the head-mounted terminal 314, but it can also be executed jointly by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the head-mounted terminal 314. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the head-mounted terminal 314 or external devices, and the head-mounted terminal 314 acquires or collects information required for processing from the data processing device 12 or external devices.

[0126] Each of the aforementioned elements, including a collection unit, an extraction unit, a publication unit, a collaboration unit, an analysis unit, and a provision unit, is implemented, for example, in at least one of the head-mounted terminal 314 and the data processing device 12. For example, the collection unit collects data from a data source on the Internet via the communication I / F 44 of the head-mounted terminal 314. The extraction unit extracts elements from the collected data, for example, via a specific processing unit 290 of the data processing device 12. The publication unit publishes the extracted data on the Internet, for example, via the control unit 46A of the head-mounted terminal 314. The collaboration unit, for example, is responsible for collaboration with local governments, via the specific processing unit 290 of the data processing device 12. The analysis unit, for example, uses AI to analyze the collected data via the specific processing unit 290 of the data processing device 12. The provision unit, for example, provides the analysis results via the control unit 46A of the head-mounted terminal 314. The correspondence between each unit and the device or control unit is not limited to the above example and can be varied.

[0127] [Fourth Implementation] Figure 7 An example of the configuration of the data processing system 410 according to the fourth embodiment is shown.

[0128] like Figure 7 As shown, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[0129] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, 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. An example of the network 54 includes a WAN and / or LAN, etc.

[0130] Robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control object 443. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, camera 42, and control object 443 are also connected to the bus 52.

[0131] Microphone 238 receives user commands by receiving the user's voice. Microphone 238 captures the user's voice, converts the captured sound into speech data, and outputs it to processor 46. Speaker 240 outputs sound according to commands from processor 46.

[0132] The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture and shutter, and an imaging element such as a CMOS image sensor or a CCD image sensor, used to photograph the user's surroundings (e.g., the shooting range defined by an angle of view equivalent to the field of vision of an average healthy person).

[0133] Communication I / F 44 is connected to network 54. Communication I / F 44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F 44 and 26 is performed in a secure state.

[0134] The controlled object 443 includes a display device, LEDs for the eyes, and motors for driving the arms, hands, and feet. The posture and movements of the robot 414 are controlled by controlling the motors for the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. In addition, facial expressions of the robot 414 can also be expressed by controlling the illumination state of the LEDs for the robot 414's eyes.

[0135] Figure 8 An example of the main functions of the data processing device 12 and the robot 414 is shown. For example... Figure 8 As shown, in the data processing device 12, specific processing is performed by the processor 28. The memory 32 stores a specific processing program 56.

[0136] The processor 28 reads a specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is implemented by the processor 28 operating as a specific processing unit 290 based on the specific processing program 56 executed on the RAM 30.

[0137] The memory 32 stores a data generation model 58 and an emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290. The specific processing unit 290 can use the emotion-specific model 59 to infer the user's emotions and perform specific processing based on the user's emotions. The emotion inference function using the emotion-specific model 59 (emotion-specific function) includes various inferences and predictions related to the user's emotions, but is not limited to this example. Furthermore, emotion inference and prediction also include, for example, emotion analysis (analysis).

[0138] In robot 414, specific processing is performed by processor 46. A specific program 60 is stored in memory 50. Processor 46 reads the specific program 60 from memory 50 and executes the read specific program 60 on RAM 48. Specific processing is achieved by processor 46 acting as control unit 46A based on the specific program 60 executed on RAM 48. Furthermore, robot 414 may also have the same data generation model and emotion-specific model as data generation model 58 and emotion-specific model 59, and use these models to perform the same processing as specific processing unit 290.

[0139] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains the processing results (prediction results, etc.) using the data generation model 58 by communicating with the server device that has the data generation model 58. In addition, the data processing device 12 may be a server device or a user-owned terminal device (e.g., a mobile phone, robot, home appliance, etc.).

[0140] The specific processing unit 290 sends the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires voice representing the user's input regarding the result of the specific processing. The control unit 46A sends the voice data representing 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.

[0141] The data generation model 58 is a so-called generative AI. An example of the data generation model 58 includes generative AIs such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with a prompt containing instructions, and also with inference data such as speech data representing speech, text data representing text, and image data representing images (e.g., data of still images or data of moving images). The data generation model 58 infers the inference data based on the instructions shown in the prompt and outputs the inference result in one or more data forms such as speech data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the aforementioned specific processing while using the data generation model 58. The data generation model 58 can also be a fine-tuned model to output inference results from prompts that do not contain instructions; in this case, the data generation model 58 is able to output inference results from prompts that do not contain instructions. The data processing device 12, etc., includes various data generation models 58, which include AI other than generative AI. These AIs include, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and are capable of various processing methods, but are not limited to these examples. Furthermore, AI can also be an AI agent. Additionally, when the processing described above is performed by AI, this processing may be partially or entirely performed by AI, but is not limited to these examples. Moreover, processing performed by AI including generative AI can be replaced by rule-based processing, and vice versa.

[0142] The data processing system 410 of the fourth embodiment performs the same processing as the data processing system 10 of the first embodiment. The processing performed by the data processing system 410 is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but it can also be executed jointly by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the robot 414 or external devices, and the robot 414 acquires or collects information required for processing from the data processing device 12 or external devices.

[0143] Each of the aforementioned elements, including a collection unit, extraction unit, disclosure unit, collaboration unit, analysis unit, and provision unit, is implemented, for example, in at least one of the robot 414 and the data processing device 12. For example, the collection unit collects data from a data source on the Internet via the communication I / F 44 of the robot 414. The extraction unit extracts elements from the collected data, for example, via a specific processing unit 290 of the data processing device 12. The disclosure unit discloses the extracted data on the Internet, for example, via the control unit 46A of the robot 414. The collaboration unit, for example, is responsible for collaboration with local governments, via the specific processing unit 290 of the data processing device 12. The analysis unit, for example, uses AI to analyze the collected data via the specific processing unit 290 of the data processing device 12. The provision unit, for example, provides the analysis results via the control unit 46A of the robot 414. The correspondence between the various units and the device or control unit is not limited to the above example and can be varied.

[0144] Furthermore, the emotion-specific model 59, serving as an emotion engine, can determine the user's emotion based on a specific mapping. Specifically, the emotion-specific model 59 can determine the user's emotion based on an emotion graph that serves as a specific mapping (see...). Figure 9 The system determines the user's emotions. In addition, the emotion-specific model 59 can also determine the robot's emotions in the same way, and the specific processing unit 290 can also perform specific processing using the robot's emotions.

[0145] Figure 9 This is a diagram representing an emotion map 400 that maps various emotions. In the emotion map 400, emotions are arranged radially from the center in concentric circles. The closer to the center of the concentric circles, the more primitive the emotion is. Further out on the concentric circles, emotions are arranged representing states or actions arising from mood. Emotion is a concept that includes both feelings and mental states. To the left of the concentric circles, emotions generated by reactions occurring in the brain are arranged roughly. To the right of the concentric circles, emotions guided by situational judgments are arranged roughly. Above and below the concentric circles, emotions generated by reactions occurring in the brain and guided by situational judgments are arranged roughly. Furthermore, the emotion of "pleasure" is arranged above the concentric circles, and the emotion of "unpleasantness" is arranged below. Thus, in the emotion map 400, various emotions are mapped according to the structure of emotion generation, while easily generated emotions are mapped nearby.

[0146] These emotions are distributed at the 3 o'clock position on the Emotion Chart 400, and usually fluctuate between peace and unease. In the right half of the Emotion Chart 400, because situational awareness is more dominant than internal feelings, it gives a sense of calm.

[0147] The inner side of the emotion diagram 400 represents the mind, and the outer side of the emotion diagram 400 represents actions. Therefore, the further you go to the outer side of the emotion diagram 400, the more the emotion can be seen (manifested in actions).

[0148] Here, human emotions are based on a balance of various factors such as posture and blood sugar levels. When these balances deviate from the ideal, it indicates unhappiness; when they approach the ideal, it indicates pleasure. In robots, cars, and motorcycles, emotions can also be created based on a balance of factors such as posture and remaining battery power. When these balances deviate from the ideal, it indicates unhappiness; when they approach the ideal, it indicates pleasure. Emotion maps can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Speech Emotion Recognition and Brain Physiological Signal Analysis Systems for Emotion, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). In the left half of the emotion map, emotions belonging to the "Reaction" domain, where sensation is dominant, are arranged. Furthermore, in the right half of the emotion map, emotions belonging to the "Situation" domain, where situational cognition is dominant, are arranged.

[0149] The emotion map defines two types of emotions that promote learning. One is a negative emotion located near the middle of "repentance" or "reflection" on the situation side. That is, when the robot experiences negative emotions such as "I never want to feel this way again" or "I never want to be scolded again." The other is a positive emotion located near "desire" on the response side. That is, when the robot experiences positive feelings such as "wanting more" or "wanting to know more."

[0150] The emotion-specific model 59 feeds user input into a pre-learned neural network to obtain emotion values ​​representing each emotion shown in the emotion graph 400, and determines the user's emotion. This neural network is pre-learned based on multiple learning data sets that combine user input with emotion values ​​representing each emotion shown in the emotion graph 400. Furthermore, this neural network is learned to... Figure 10 As shown in sentiment graph 900, sentiment values ​​in nearby configurations are similar to each other. Figure 10 Examples show how emotions such as "peace of mind," "stability," and "reassurance" can result in similar emotional values.

[0151] In the above embodiments, a specific processing is described by a single computer 22, but the technology disclosed herein is not limited to this, and distributed processing by multiple computers, including computer 22, is also possible.

[0152] In the above embodiments, an example of storing a specific processing program 56 in memory 32 is illustrated, but the technology disclosed herein is not limited thereto. For example, the specific processing program 56 may also 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 performs specific processing according to the specific processing program 56.

[0153] Alternatively, the specific processing program 56 can be stored in a storage device such as a server connected to the data processing device 12 via a network 54, and the specific processing program 56 can be downloaded and installed into the computer 22 upon request from the data processing device 12.

[0154] Furthermore, it is not necessary to store the entire specific process 56 in a storage device such as a server connected to the data processing device 12 via the network 54, nor is it necessary to store the entire specific process 56 in the memory 32; a portion of the specific process 56 may also be stored.

[0155] As a hardware resource for performing specific processing, various processors can be used. For example, a CPU is a general-purpose processor that functions as a hardware resource for performing specific processing by executing software, i.e., programs. Additionally, a dedicated circuit can be listed as a processor; it is a processor with a circuit structure specifically designed for performing specific processing, such as a FPGA (Field-Programmable Gate Array), PLD (Programmable Logic Device), or ASIC (Application Specific Integrated Circuit). Every processor has built-in or connected memory, and every processor executes specific processing by using that memory.

[0156] The hardware resources for performing a specific process can consist of one of these various processors, or a combination of two or more processors of the same or different types (e.g., a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Furthermore, the hardware resources for performing a specific process can also be a single processor.

[0157] As an example of a single processor, the first type consists of a combination of one or more CPUs and software, which functions as a hardware resource to perform specific processing. The second type uses a processor, such as a System-on-a-chip (SoC), which implements the entire system functionality, including multiple hardware resources performing specific processing, using a single IC chip. In this case, the specific processing is implemented using one or more of the aforementioned processors that serve as hardware resources.

[0158] Furthermore, as the hardware architecture of these various processors, more specifically, circuits combining semiconductor elements and other circuit components can be used. Moreover, the specific process described above is merely an example. Therefore, it goes without saying that, without departing from the main point, unnecessary steps can be removed, new steps can be added, or the processing order can be changed.

[0159] Furthermore, although the above examples have been described in terms of first to fourth embodiments, some or all of these embodiments can also be combined. Additionally, the smart device 14, smart glasses 214, head-mounted terminal 314, and robot 414 are just examples; they can be combined separately or are other devices.

[0160] The foregoing descriptions and illustrations are detailed explanations of the parts covered by this disclosure and are merely one example of this disclosure. For instance, the descriptions of the above-described structure, function, role, and effect are just one example of the structure, function, role, and effect of the parts covered by this disclosure. Therefore, it goes without saying that, without departing from the spirit of this disclosure, unnecessary parts can be deleted, new elements can be added, or replacements can be made to the foregoing descriptions and illustrations. Furthermore, to avoid confusion and facilitate understanding of the parts covered by this disclosure, explanations of technical common sense that does not require special explanation for implementing this disclosure have been omitted from the foregoing descriptions and illustrations.

[0161] All documents, patent applications and technical standards described in this specification are incorporated herein by reference as if they were specifically and individually described as incorporated by reference.

[0162] (Note 1) A system, characterized in that it comprises: The data collection department is responsible for collecting data. An extraction unit is used to extract elements from the data collected by the collection unit; The disclosure section is used to publicly disclose the data extracted by the extraction section on the Internet; The Cooperation Department is responsible for cooperation with local governments. The analysis unit, comprised of AI, analyzes the data collected by the collection unit; and A providing unit is used to provide the parsing results obtained by the parsing unit.

[0163] (Note 2) The system as described in Appendix 1 is characterized in that, The collection unit collects data from data sources on the Internet.

[0164] (Note 3) The system as described in Appendix 1 is characterized in that, The extraction unit extracts elements from the collected data.

[0165] (Note 4) The system as described in Appendix 1 is characterized in that, The publicly disclosed department will publish the extracted data on the Internet.

[0166] (Note 5) The system as described in Appendix 1 is characterized in that, The cooperation department is responsible for cooperation with local governments.

[0167] (Note 6) The system as described in Appendix 1 is characterized in that, The analysis unit uses AI to analyze the collected data.

[0168] (Note 7) The system as described in Appendix 1 is characterized in that, The providing department provides the analysis results.

[0169] (Note 8) The system as described in Appendix 1 is characterized in that, The data collection unit analyzes the user's emotions and adjusts the timing of data collection based on the analyzed emotions.

[0170] (Note 9) The system as described in Appendix 1 is characterized in that, The collection unit analyzes the user's past data to provide history and selects an appropriate collection method.

[0171] (Postscript 10) The system as described in Appendix 1 is characterized in that, The data collection unit filters data based on the user's current area of ​​interest during data collection.

[0172] (Postscript 11) The system as described in Appendix 1 is characterized in that, The collection unit analyzes the user's emotions and determines the priority of the collected data based on the analyzed user emotions.

[0173] (Postscript 12) The system as described in Appendix 1 is characterized in that, When collecting data, the collection unit prioritizes collecting highly relevant data, taking into account the user's geographic location information.

[0174] (Postscript 13) The system as described in Appendix 1 is characterized in that, The data collection unit analyzes users' social media activities and collects relevant data during the data collection process.

[0175] (Postscript 14) The system as described in Appendix 1 is characterized in that, The extraction unit analyzes the user's emotions and determines the priority order of the extracted elements based on the analyzed user emotions.

[0176] (Postscript 15) The system as described in Appendix 1 is characterized in that, During extraction, the extraction unit adjusts the level of detail based on the importance of the data.

[0177] (Postscript 16) The system as described in Appendix 1 is characterized in that, During the extraction process, the extraction unit applies different extraction algorithms based on the category of the data.

[0178] (Postscript 17) The system as described in Appendix 1 is characterized in that, The extraction unit analyzes the user's emotions and adjusts the display method of the extracted elements based on the analyzed user emotions.

[0179] (Postscript 18) The system as described in Appendix 1 is characterized in that, When extracting data, the extraction unit determines the extraction priority based on the data submission time.

[0180] (Postscript 19) The system as described in Appendix 1 is characterized in that, During extraction, the extraction unit adjusts the extraction order based on the correlation of the data.

[0181] (Postscript 20) The system as described in Appendix 1 is characterized in that, The public part analyzes the user's emotions and adjusts the display method of the public data based on the analyzed user emotions.

[0182] (Postscript 21) The system as described in Appendix 1 is characterized in that, When making a disclosure, the disclosure department adjusts the level of detail based on the importance of the data.

[0183] (Postscript 22) The system as described in Appendix 1 is characterized in that, When making data public, the disclosure department applies different disclosure algorithms based on the category of the data.

[0184] (Postscript 23) The system as described in Appendix 1 is characterized in that, The disclosure unit analyzes users' emotions and determines the priority of the disclosed data based on the analyzed user emotions.

[0185] (Postscript 24) The system as described in Appendix 1 is characterized in that, When making public data, the priority of disclosure is determined based on the data submission period.

[0186] (Postscript 25) The system as described in Appendix 1 is characterized in that, When making disclosures, the disclosure department adjusts the order of disclosure based on the relevance of the data.

[0187] (Postscript 26) The system as described in Appendix 1 is characterized in that, The collaboration department analyzes the emotions of local autonomous regions and adjusts the collaboration method based on the analyzed emotions.

[0188] (Postscript 27) The system as described in Appendix 1 is characterized in that, When cooperating, the cooperation department selects appropriate cooperation methods by referring to the past cooperation history of the local autonomous region.

[0189] (Postscript 28) The system as described in Appendix 1 is characterized in that, The collaboration department analyzes the sentiments of local autonomous regions and determines the priority of collaboration based on the analyzed sentiments.

[0190] (Postscript 29) The system as described in Appendix 1 is characterized in that, When collaborating, the collaboration department takes into account the geographical location information of the local autonomous region to select the best collaboration method.

[0191] (Note 30) The system as described in Appendix 1 is characterized in that, The parsing unit analyzes the user's emotions and adjusts the parsing method based on the analyzed user emotions.

[0192] (Postscript 31) The system as described in Appendix 1 is characterized in that, During the parsing process, the parsing unit adjusts the level of detail based on the importance of the data.

[0193] (Note 32) The system as described in Appendix 1 is characterized in that, During the parsing process, the parsing unit applies different parsing algorithms based on the category of the data.

[0194] (Postscript 33) The system as described in Appendix 1 is characterized in that, The parsing unit analyzes the user's emotions and determines the parsing priority based on the analyzed user emotions.

[0195] (Postscript 34) The system as described in Appendix 1 is characterized in that, During parsing, the parsing unit determines the parsing priority based on the data submission time.

[0196] (Postscript 35) The system as described in Appendix 1 is characterized in that, During the parsing process, the parsing unit adjusts the parsing order based on the correlation of the data.

[0197] (Postscript 36) The system as described in Appendix 1 is characterized in that, The provider analyzes the user's emotions and adjusts the display method of the provided information based on the analyzed user emotions.

[0198] (Postscript 37) The system as described in Appendix 1 is characterized in that, When providing information, the providing department adjusts the level of detail based on the importance of the information.

[0199] (Postscript 38) The system as described in Appendix 1 is characterized in that, When providing information, the providing unit applies different providing algorithms based on the category of the information.

[0200] (Postscript 39) The system as described in Appendix 1 is characterized in that, The provider analyzes the user's emotions and determines the priority of the information provided based on the analyzed emotions.

[0201] (Postscript 40) The system as described in Appendix 1 is characterized in that, When providing information, the providing department determines the priority order of provision based on the submission time of the information.

[0202] (Postscript 41) The system as described in Appendix 1 is characterized in that, When providing information, the providing department adjusts the order of provision based on the relevance of the information.

Claims

1. A system, characterized in that, include: The data collection department is responsible for collecting data. An extraction unit is used to extract elements from the data collected by the collection unit; The disclosure section is used to publicly disclose the data extracted by the extraction section on the Internet; The Cooperation Department is responsible for cooperation with local governments. The analysis unit, comprised of AI, analyzes the data collected by the collection unit; and A providing unit is used to provide the parsing results obtained by the parsing unit.

2. The system as described in claim 1, characterized in that, The collection unit collects data from data sources on the Internet.

3. The system as described in claim 1, characterized in that, The extraction unit extracts elements from the collected data.

4. The system as described in claim 1, characterized in that, The publicly disclosed department will publish the extracted data on the Internet.

5. The system as described in claim 1, characterized in that, The cooperation department is responsible for cooperation with local governments.

6. The system as described in claim 1, characterized in that, The analysis unit uses AI to analyze the collected data.

7. The system as described in claim 1, characterized in that, The providing department provides the analysis results.

8. The system as described in claim 1, characterized in that, The data collection unit analyzes the user's emotions and adjusts the timing of data collection based on the analyzed emotions.

Images