system

Abstract

We provide the system. [Solution] A means of autonomously monitoring the performance of other AI agents, A means for identifying necessary improvements based on the aforementioned performance, Based on the identified areas for improvement, a means to retrain other AI agents, A means of providing feedback through natural language dialogue, A means for learning from the aforementioned feedback and improving the performance of the learning agent, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Many AI agents are optimized for tasks in specific environments, but it is difficult to flexibly adapt to newly changing environments and data. Therefore, the performance of AI agents may decline. In addition, retraining requires advanced expertise and costs, which is a burden for many companies and organizations. As a result, there are problems in that the limits of AI agents occur and effective operation is hindered.

Means for Solving the Problems

[0005] This invention autonomously monitors the performance of an AI agent and identifies areas for improvement. Furthermore, it retrains the AI ​​agent based on these improvements and provides feedback through natural language dialogue. Using a reinforcement learning algorithm, the AI ​​agent adapts to changes in the environment and automatically updates its retraining content in response to new data and changes. As a result, the AI ​​agent can always maintain an optimal state and perform tasks efficiently.

[0006] An "AI agent" is an artificial intelligence program designed to perform specific tasks and possesses the ability to make autonomous decisions based on data.

[0007] "Means of autonomous monitoring" refers to a function that tracks and records the performance and actions of AI agents in real time without human intervention.

[0008] "Methods for identifying areas for improvement" refer to methods for analyzing the operational data of an AI agent to find areas that need optimization or areas where performance can be improved.

[0009] "Retraining" refers to a function that performs a process in which an AI agent improves its performance by learning again based on new data and environments.

[0010] "Natural language dialogue" refers to a method of communication in which information is exchanged with AI using human language, and includes the ability of a program to engage in text-based dialogue.

[0011] "Means of providing feedback" refers to methods of communicating information that allows an AI agent to evaluate its own performance and learning results based on its operational results, and to reflect this in its next actions.

[0012] A "reinforcement learning algorithm" is a machine learning technique that learns strategies to maximize rewards through trial and error, and is an algorithm designed to improve adaptability in an environment.

[0013] "Changes in data and environment" refers to updates to the factors that the AI ​​agent must consider due to newly acquired information or fluctuations in external factors. [Brief explanation of the drawing]

[0014] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, which incorporates an emotion engine. [Figure 14]It is a sequence diagram showing the processing flow of a data processing system in Application Example 2 when a sentiment engine is combined.

Embodiments for Carrying Out the Invention

[0015] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0016] First, the terms used in the following description will be explained.

[0017] In the following embodiments, a processor with a reference number (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), etc.

[0018] In the following embodiments, a RAM (Random Access Memory) with a reference number is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0019] In the following embodiments, a storage with a reference number is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, etc.

[0020] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0021] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0022] [First Embodiment]

[0023] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0024] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0025] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0026] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0027] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0028] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio 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 image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0029] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0030] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0031] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0032] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0033] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0034] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0035] This invention is an advanced system for educating and training AI agents, and is implemented as follows: The system mainly consists of a server, terminals, and users.

[0036] 1. Server Role

[0037] The server autonomously monitors the performance of AI agents and collects log data. For example, in the case of customer service agents, it records the content of customer interactions, response times, and accurate responses. The server analyzes the collected data to identify areas for improvement to enhance performance.

[0038] 2. Feedback and Retraining

[0039] The server sends identified areas for improvement as feedback to the learning agent. The learning agent uses the feedback to retrain itself using a reinforcement learning algorithm. For example, if the price of a product changes, the agent updates its response based on new information. The reinforcement learning algorithm adapts to environmental changes in real time and performs further optimizations.

[0040] 3. User Interaction

[0041] Users can access the system via their devices to visualize the AI ​​agent's current performance and retraining progress. If necessary, users can also provide additional data to the system and guide the learning process. For example, they can provide new sales materials in response to changes in sales strategy.

[0042] 4. Dialogue using natural language

[0043] On the device, an AI agent can interact with the user in natural language. Based on user queries, the AI ​​provides optimal answers and accumulates feedback. Through this interaction, the system understands the user's intent and uses this data to guide better results.

[0044] These components work together to enable the AI ​​agent to continuously evolve, gaining high flexibility and adaptability to meet the needs of businesses. The entire system provides a robust foundation for the AI ​​agent to respond quickly to environmental changes and perform tasks efficiently.

[0045] The following describes the processing flow.

[0046] Step 1:

[0047] The server collects AI agent operation data in real time and stores it as logs. This includes AI response time, accuracy, and user interaction.

[0048] Step 2:

[0049] The server analyzes the collected data to identify areas where the AI ​​agent's performance needs improvement. This analysis may involve heuristic evaluation or machine learning models.

[0050] Step 3:

[0051] The server sends identified areas for improvement as feedback to the learning agent. This feedback includes details of instances where the response was insufficient and instances where it was successful.

[0052] Step 4:

[0053] The device then initiates retraining of the learning agent based on feedback. Reinforcement learning algorithms set new parameters, and the agent improves itself.

[0054] Step 5:

[0055] Users can use their devices to check the retraining status and improvements made to the AI ​​agent. They can also provide additional data and instructions to the server as needed.

[0056] Step 6:

[0057] The device manages user interactions through natural language processing capabilities, supporting the AI ​​agent in providing optimal responses. This conversational data is also sent back to the server for further analysis.

[0058] Step 7:

[0059] The server monitors the retrained AI agent's new performance and evaluates whether the expected improvements are being seen. If further optimization is deemed necessary, the feedback loop is repeated.

[0060] (Example 1)

[0061] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0062] Traditional artificial intelligence (AI) faces challenges such as its inability to quickly adapt to environmental changes and new data, and the difficulty in continuously improving its performance. Furthermore, it often fails to effectively utilize feedback through natural language interaction with users, and the process of identifying areas for improvement and retraining is frequently done manually, making it inefficient. A system is needed to solve these problems and enhance the flexibility and adaptability of AI.

[0063] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0064] In this invention, the server includes means for autonomously monitoring other artificial intelligences, means for collecting and analyzing performance data of the artificial intelligences, and means for identifying areas for improvement based on the analysis results. This makes it possible to efficiently improve the performance of the artificial intelligences and to quickly adapt to new data and environmental changes.

[0065] "Artificial intelligence" is a program or system that learns autonomously based on data and makes decisions.

[0066] "Monitoring" is the act of continuously checking the status and activity of a system or process and collecting necessary information.

[0067] "Performance data" refers to quantitative or qualitative information used to evaluate the operation and effectiveness of artificial intelligence.

[0068] "Analysis" is the process of analyzing collected data to find meaning and derive specific results or conclusions.

[0069] "Improvements" refer to changes or modifications necessary to enhance the capabilities and performance of artificial intelligence.

[0070] "Retraining" is an additional learning process that uses newly collected information and feedback to further improve the performance of artificial intelligence.

[0071] "Natural language" refers to the language that humans use in their daily lives, and which has a specific grammar and structure.

[0072] "Dialogue" is a form of communication in which information and opinions are exchanged between different individuals or systems.

[0073] "Feedback" refers to evaluations or opinions provided based on specific actions or outcomes, serving as guidance for improvement or adjustment.

[0074] Reinforcement learning is a type of machine learning algorithm in which an agent learns the optimal action by receiving rewards through trial and error.

[0075] "Adaptation" is a response that flexibly responds to changes in the environment and new information, thereby improving behavior and abilities.

[0076] This invention improves the performance of artificial intelligence through a process of collecting, analyzing, and improving various data within an artificial intelligence-based system. The system mainly consists of a server, terminals, and users, and each element works in coordination.

[0077] The server is responsible for autonomously monitoring the performance of the artificial intelligence and collecting log data. Specifically, it uses software used as a log collection tool to collect data such as the content of the AI's dialogue, response time, and response accuracy, and stores it in a database. The server then analyzes the collected data using data analysis software such as Python or R to identify areas for improvement to enhance performance.

[0078] The terminal provides an interface for users to access the system and visualize the current performance and retraining progress of the artificial intelligence. A dedicated dashboard application allows users to check the system status in real time. Furthermore, the artificial intelligence interacts with the user in natural language via the terminal, accumulating the feedback received. SpaCy and Transformers are used as natural language processing frameworks.

[0079] Users can provide additional data to the artificial intelligence (AI) via the system's dashboard, and use that data to guide the AI's retraining. This allows users to provide new sales materials in response to changes in sales strategies. Users can also send queries to the AI ​​using prompts and receive responses. For example, they can expect to receive new information by using a prompt such as, "Please tell me the latest information on new products."

[0080] In this way, through the cooperation of servers, terminals, and users, artificial intelligence can adapt to environmental changes in real time and achieve efficient task processing. The entire system combines a generative AI model with natural language dialogue using prompt sentences to achieve efficient and flexible artificial intelligence operation.

[0081] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0082] Step 1:

[0083] The server collects log data from AI-generated conversations. It receives conversation text data, response times, and response accuracy as input. The server aggregates this data using a log collection tool and stores it in a database. Specifically, periodic data imports are performed into the database. The output obtained at this stage is structured log data that can be used for subsequent analysis.

[0084] Step 2:

[0085] The server analyzes the collected log data. It uses the structured log data obtained in Step 1 as input. The server uses a data analysis program and applies machine learning algorithms to calculate metrics related to the AI's performance and identify areas for improvement. Specifically, data cleaning and feature extraction are performed, and analysis results are generated. The output is a list of specific areas for improvement to enhance the AI's performance.

[0086] Step 3:

[0087] The server generates feedback to the artificial intelligence based on the analysis results and triggers retraining. The list of improvements from Step 2 is used as input. The server uses reinforcement learning to prepare a retraining protocol to update the AI's parameters based on the input data. Specifically, this involves preparing the retraining dataset and updating the model. The output is the updated artificial intelligence model.

[0088] Step 4:

[0089] Users access the system through a terminal to check current performance and retraining progress. The input is the latest AI status information sent from the server. The terminal visualizes this information via a dedicated dashboard and presents it to the user. Specifically, the dashboard displays graphs and metrics that are updated in real time. The output is visualized information for the user to use.

[0090] Step 5:

[0091] The device facilitates interaction between the user and artificial intelligence using natural language. It receives prompt text from the user as input. The device uses a generative AI model to generate the optimal response based on the prompt. Specifically, it analyzes the user's intent through natural language processing and generates a response. Therefore, the output is a natural language response provided to the user.

[0092] (Application Example 1)

[0093] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0094] Current content delivery systems struggle to quickly respond to users' diverse viewing needs and rapidly changing preferences, often resulting in insufficient generation of personalized recommendations. This leads to problems such as the inability to appropriately recommend content that viewers are interested in, and thus the inability to improve the user experience.

[0095] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0096] In this invention, the server includes means for collecting viewing history and user evaluation information, means for generating personalized recommendations based on the collected information, and means for updating the generated recommendations in real time. This makes it possible to leverage user feedback to adjust personalized recommendations in a timely manner and increase user satisfaction.

[0097] "Viewing history" refers to data such as records, duration, and frequency of content viewing by a user.

[0098] "Rating information" refers to information such as feedback and rating scores that users have given to content.

[0099] "Personalized recommendations" refer to a list of content customized based on each user's viewing history and rating information.

[0100] "Real-time updates" means instantly reflecting users' new viewing history and rating information, and changing recommendations on the spot.

[0101] "Natural language dialogue" is a form of communication in which users interact with computer systems using human language.

[0102] "Feedback" refers to opinions and impressions that users provide about their experience using content or a system.

[0103] A "recommendation algorithm" is a computational method for automatically selecting content that is suitable for the user.

[0104] To implement this invention, consider the following system configuration. First, the server functions as the main data processing unit, collecting user viewing history and evaluation information. Specifically, it uses log data obtained from various content platforms and stores it in a database. The server is equipped with a machine learning framework such as TENSORFLOW®, and serves as a foundation for data analysis and reinforcement learning.

[0105] The server uses this data to periodically update its recommendation algorithm, which suggests the most suitable content for each user. To enable real-time updates, a Python script is used to ensure the algorithm always reflects the latest information. Furthermore, user feedback is obtained through the device in a natural language dialogue format, and this information is also used to optimize the recommendation algorithm. The devices used are smartphones and tablets equipped with user-friendly interfaces that accept natural language input.

[0106] For example, if a user provides feedback such as "I enjoyed it" after watching a movie, the server analyzes it and adds other works of similar genres or by similar directors to the recommendation list. Another example of a prompt message is, "I watched some new content. What should I watch next?" In this way, it is possible to increase user engagement and provide a more personalized viewing experience.

[0107] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0108] Step 1:

[0109] The server receives user viewing history and rating information from the content platform. The received data is stored in the database. At this point, the input is viewing history and rating data, and the output is the stored data. The server performs format conversion and integrity checks on the log data and inserts it into the database in a formatted state.

[0110] Step 2:

[0111] The server runs a recommendation algorithm using stored data to generate personalized content recommendations. The input for this step is user data retrieved from a database, and the output is a personalized recommendation list. The server leverages a machine learning framework to extract significant patterns from the data, shaping the most relevant content for each user into a list.

[0112] Step 3:

[0113] The terminal receives natural language feedback from the user and sends it to the server. The input is the user's natural language feedback, and the output is the parsed data sent to the server. The terminal has speech recognition or text input capabilities to format the user's input feedback without ambiguity.

[0114] Step 4:

[0115] The server analyzes the received feedback and readjusts the recommended algorithm. The input to this process is the feedback data, and the output is the updated algorithm parameters. The server uses the feedback to perform reinforcement learning, learning in real time to improve the parameters of the recommended model.

[0116] Step 5:

[0117] The server sends the updated recommendation list back to the terminal, displaying the new recommendations when the user accesses it. The input is the latest recommendation data, and the output is the display list on the user's terminal. The server quickly transfers the data to the terminal via a communication protocol, creating an environment where users can smoothly check the new recommendations.

[0118] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0119] This invention is a system that combines an emotion engine with an AI agent to achieve more advanced interaction and effective learning. This system mainly consists of a server, terminals, and users.

[0120] 1. Server processing

[0121] The server monitors the performance of the AI ​​agent and collects user emotion data along with performance data. Emotion data is obtained using methods such as voice tone, facial expressions, and text analysis. For example, in a customer support system, the emotional state of a user can be analyzed from their voice tone and logged in real time.

[0122] 2. Integration of the Emotional Engine

[0123] The server incorporates an emotion engine, and the AI ​​agent analyzes the user's emotions through this engine. This data influences the AI ​​agent's response, which is then adjusted to suit the user's emotions. For example, if the user is dissatisfied, the AI ​​agent will provide a more empathetic response.

[0124] 3. Feedback and Retraining

[0125] The server sends feedback, including the user's emotional state, to the learning agent. Based on this feedback, the AI ​​agent retrains and adjusts its learning algorithm. This allows the AI ​​agent to strengthen its emotion-based response strategy and improve the accuracy of its responses.

[0126] 4. User Interaction

[0127] Users interact with an AI agent through their device, and their emotional state is recognized by the system. Users can feel that the AI's responses are appropriate and considerate of their emotions. For example, in educational services, the AI ​​can assess a student's emotional state and provide optimal learning support.

[0128] By incorporating an emotion engine in this way, AI agents can achieve interactions that respond to the user's emotions, providing a more personalized experience. This system provides a foundation for AI agents to perform their tasks while flexibly responding to changes in the environment and the user's emotional state.

[0129] The following describes the processing flow.

[0130] Step 1:

[0131] The server initiates interaction with the user, receiving voice input and text data. This data is immediately sent to the emotion engine for analysis.

[0132] Step 2:

[0133] The emotion engine analyzes the tone of voice, vocabulary, and context of the text to identify the user's emotional state. This analysis is returned to the server as a quantified emotion index.

[0134] Step 3:

[0135] The server adjusts the AI ​​agent's response based on sentiment indicators. For example, if a user expresses dissatisfaction, the AI ​​agent is instructed to apologize and offer quick solutions to resolve the problem.

[0136] Step 4:

[0137] On the device, the AI ​​agent provides the user with tailored responses. This involves using natural speech or text to engage in emotionally conscious dialogue.

[0138] Step 5:

[0139] Through interaction with the AI ​​agent, users receive responses tailored to their needs. Emotionally sensitive responses make users feel more supported.

[0140] Step 6:

[0141] The server sends user feedback and sentiment data to the learning agent, prompting retraining. Based on this data, the learning agent performs reinforcement learning to develop more refined sentiment recognition and response strategies.

[0142] Step 7:

[0143] The device releases an updated AI agent based on the retraining content, preparing for future user interactions. This process allows the AI ​​agent to continuously evolve and adapt more closely to the user's emotions.

[0144] (Example 2)

[0145] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0146] Traditional AI systems have struggled to provide responses that take into account the user's emotional state, resulting in shortcomings in the user experience. Furthermore, the lack of means to analyze emotions in real time and optimize responses based on that analysis meant that user feedback could not be effectively incorporated, limiting the improvement of the accuracy of the learning models.

[0147] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0148] In this invention, the server includes means for acquiring user data based on an information processing device for analyzing emotional states, means for identifying the user's emotional state by analyzing the user data, and means for adjusting the AI ​​system's response according to the emotional state. This enables the AI ​​system to provide interactions that respond to the user's emotions, improving the quality of the user experience and simultaneously improving the accuracy of the learning model that reflects the feedback.

[0149] An "information processing device" is a device that collects and analyzes user data to identify emotional states.

[0150] "User data" refers to information such as voice, facial expressions, and text acquired during interactions with users.

[0151] "Emotional state" refers to the results of analysis to identify the user's emotions, and serves as fundamental information for adjusting the AI ​​system's response.

[0152] "AI system response" refers to the content of the conversation with the AI ​​agent, which is adjusted based on the user's emotional state.

[0153] "Feedback information" refers to information that users provide regarding their evaluations and suggestions for improvement in response to the AI ​​system's actions.

[0154] A "learning model" is a mathematical model used to improve the response accuracy of an AI system based on data analysis.

[0155] The system of this invention aims to adjust the response of an AI agent based on the user's emotions. The system mainly consists of a server, a terminal, and a user, and provides the environment for each function to operate.

[0156] The server houses the information processing unit that forms the core of the AI ​​agent, and it collects and analyzes user data. Various data collection methods are used here, such as speech recognition software and facial expression recognition tools. For example, general speech analysis service software is used for speech recognition, processing changes in speech tone and other data as digital data. For facial expression recognition, facial expression analysis software is used to quantify the user's emotional state in real time and send the analysis data to the server. Based on this analysis data, the AI ​​system's response is dynamically optimized.

[0157] The terminal functions as a means of providing a user interface and acts as a border for information exchange between the user and the AI ​​agent. Specifically, a chat application or voice interface runs on the terminal, allowing the user to interact with the AI ​​agent in natural language. When the user provides voice input to the terminal, the voice is analyzed on a server, and an appropriate response based on the user's emotional state is returned to the terminal. This allows the user to feel that the AI's responses are considerate of their emotions.

[0158] The feedback information that users provide to the AI ​​agent is collected by the server and used to retrain the learning model. This feedback is important data for improving the AI's performance, and the generative AI model is used to improve response accuracy. The generative AI model used at this time is made to generate appropriate responses using prompt statements.

[0159] A concrete example of a prompt might be, "In the following educational scenario, how should the AI ​​agent respond if a student is struggling to understand the learning material?" By using this prompt to obtain a response from the generative AI model, the AI ​​agent can provide more empathetic and effective support to the user.

[0160] In this way, the present invention provides a system that flexibly responds to user emotions, achieving improved user experience and enhanced performance of AI agents.

[0161] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0162] Step 1:

[0163] The server begins collecting user data. When a user uses a terminal to input voice or text, the terminal sends this data to the server. The input data is analyzed by speech recognition software and text analysis tools. Specifically, in the case of voice input, it is digitized and converted into text format through natural language processing. The output of this process is the text data to be analyzed.

[0164] Step 2:

[0165] The server identifies emotional data based on the collected text data. The emotion analysis engine within the server receives this text data and analyzes the user's emotional state by comprehensively considering the results of voice tone and facial recognition. This analysis is performed by an emotion analysis algorithm, and the output is an emotion category (e.g., joy, anger, sadness).

[0166] Step 3:

[0167] The server inputs the obtained emotion categories into a generative AI model to generate a response appropriate for the user. By providing the emotion categories along with prompts to the model, the AI ​​generates an appropriate response that takes emotions into consideration. Specifically, an example prompt such as "What is the appropriate response if the user is angry?" is set, and the generative AI model is used to generate a response to the user. As a result, the AI ​​agent receives a customized response to the user as output.

[0168] Step 4:

[0169] The server forwards the generated response to the terminal. The terminal receives this response and displays it to the user through the user interface. Specifically, if it is in text format, it is displayed in a chat window on the screen, and if it is a voice response, the text is converted into speech using speech synthesis technology. The user interacts with the AI ​​agent through this response.

[0170] Step 5:

[0171] The user provides feedback on the AI ​​agent's responses. The user sends feedback input to the server via their device, indicating satisfaction and areas for improvement. The server accumulates this feedback and incorporates it into its learning algorithm. The AI ​​model's performance is retrained using this data to improve future response accuracy. The output is an improved, trained model.

[0172] (Application Example 2)

[0173] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0174] Conventional AI agents struggle to dynamically adjust their responses based on user emotions, and appropriate dialogue is especially crucial in emotionally sensitive environments such as caregiving settings. However, existing systems are unable to adequately analyze users' emotions and provide adaptive responses instantly, making it difficult to provide personalized services that take emotions into consideration.

[0175] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0176] In this invention, the server includes means for autonomously monitoring the performance of other AI agents, means for analyzing the user's emotions, and means for the AI ​​agent to adjust its response based on the emotion analysis. This makes it possible to instantly analyze the user's emotions in settings such as nursing care and provide a response that is in harmony with them, thereby providing a more personalized experience.

[0177] "Autonomously monitoring the performance of other AI agents" means that an AI agent independently observes the activities of other AI agents and continuously evaluates their performance.

[0178] "Identifying necessary improvements" means analyzing data on the AI ​​agent's performance to determine the specific elements and areas that need improvement.

[0179] "Retraining" means updating the AI ​​agent's learning model based on identified areas for improvement and running the learning process again to enhance its performance.

[0180] "Emotional analysis," which analyzes the emotions of users, is a process that collects emotional data from the user's facial expressions, tone of voice, text, etc., and clarifies their emotional state.

[0181] "Adjusting responses" means modifying the output of the AI ​​agent based on the results of sentiment analysis to provide users with more emotionally sensitive responses.

[0182] "Natural language dialogue" refers to the communication process between a user and an AI agent using the language that humans use in everyday life.

[0183] A "reinforcement learning algorithm" is a learning process that improves the efficiency of an AI agent by selecting actions based on rewards from the environment and repeatedly refining those selections.

[0184] "Providing feedback" means sending back to the system the user's reactions and interaction results obtained by the AI ​​agent as useful information for performance improvement.

[0185] This invention aims to enhance individualized care by introducing AI agents into small-group living environments such as caregiving settings. The specific elements for implementing this invention are described below.

[0186] The server autonomously monitors the performance of other AI agents and runs software for emotion analysis. For example, it uses the Google® Cloud Speech-to-Text API to analyze tone from speech and the Microsoft® Face API for facial expression analysis. The server collects this data and evaluates the user's emotional state in real time.

[0187] The terminals used are smartphones and caregiving robots, which acquire audio and video data through sensor devices such as microphones and cameras. The acquired data is sent to a server, where an AI agent generates an appropriate response based on the analysis results. The generated response is provided to the user in natural language through a dialogue module. The AI ​​agent receives feedback within a certain range periodically and improves its learning algorithm with new data.

[0188] Users interact with the AI ​​agent through everyday conversations, evaluating the emotional adaptability of its responses. For example, if the AI ​​agent determines that the user is feeling down, it may suggest prompts such as, "Did anything fun happen today? Shall I put on some relaxing music?" to facilitate a more personalized experience.

[0189] In this way, the entire system adjusts according to the user's emotional changes and provides more adaptive help, aiming to improve efficiency and quality in care settings.

[0190] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0191] Step 1:

[0192] The device uses a microphone and camera to acquire the user's voice and video data in real time. Voice data includes tone, volume, and speaking speed, while video data captures the user's facial expressions and movements. This data is then transferred directly to the server.

[0193] Step 2:

[0194] The server converts the received audio data into text using the Google Cloud Speech-to-Text API and then performs emotional analysis on the voice. It analyzes variations in volume and speaking speed from the audio to profile the emotional state. As a result, it obtains a numerical evaluation of the user's emotions (e.g., excitement, joy, sadness, anger).

[0195] Step 3:

[0196] The server analyzes the video data and uses facial expression analysis tools such as the Microsoft Face API to determine the user's emotions from their facial expressions. Specifically, it analyzes facial feature points (eyes, eyebrows, mouth angle, etc.) and maps changes in facial expressions in the video to emotion categories (e.g., smile, confusion, surprise).

[0197] Step 4:

[0198] By combining this sentiment analysis data, the server uses an AI model to determine the user's current emotional state. Based on this determination, it generates a response from the AI ​​agent and creates a prompt. For example, if it determines that the user is feeling anxious, it prepares a response that will provide reassurance.

[0199] Step 5:

[0200] The AI ​​agent sends the generated prompt message back to the terminal, which then presents it to the user as voice or text. For example, it can engage in a more human-like conversation, such as, "Did anything fun happen today? Shall I put on some relaxing music?"

[0201] Step 6:

[0202] After receiving a response from the AI ​​agent, the user inputs that feedback as a separate dialogue. The device sends this feedback to the server as new input data. The server uses the received feedback as training data for the AI ​​agent and retrains it to improve future responses.

[0203] This entire process enables the entire system to continuously provide a flexible and adaptive experience that responds to the user's changing emotions.

[0204] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0205] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0206] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0207] [Second Embodiment]

[0208] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0209] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0210] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0211] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0212] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0213] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0214] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0215] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0216] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0217] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0218] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0219] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0220] This invention is an advanced system for educating and training AI agents, and is implemented as follows: The system mainly consists of a server, terminals, and users.

[0221] 1. Server Role

[0222] The server autonomously monitors the performance of AI agents and collects log data. For example, in the case of customer service agents, it records the content of customer interactions, response times, and accurate responses. The server analyzes the collected data to identify areas for improvement to enhance performance.

[0223] 2. Feedback and Retraining

[0224] The server sends identified areas for improvement as feedback to the learning agent. The learning agent uses the feedback to retrain itself using a reinforcement learning algorithm. For example, if the price of a product changes, the agent updates its response based on new information. The reinforcement learning algorithm adapts to environmental changes in real time and performs further optimizations.

[0225] 3. User Interaction

[0226] Users can access the system via their devices to visualize the AI ​​agent's current performance and retraining progress. If necessary, users can also provide additional data to the system and guide the learning process. For example, they can provide new sales materials in response to changes in sales strategy.

[0227] 4. Dialogue using natural language

[0228] On the device, an AI agent can interact with the user in natural language. Based on user queries, the AI ​​provides optimal answers and accumulates feedback. Through this interaction, the system understands the user's intent and uses this data to guide better results.

[0229] These components work together to enable the AI ​​agent to continuously evolve, gaining high flexibility and adaptability to meet the needs of businesses. The entire system provides a robust foundation for the AI ​​agent to respond quickly to environmental changes and perform tasks efficiently.

[0230] The following describes the processing flow.

[0231] Step 1:

[0232] The server collects AI agent operation data in real time and stores it as logs. This includes AI response time, accuracy, and user interaction.

[0233] Step 2:

[0234] The server analyzes the collected data to identify areas where the AI ​​agent's performance needs improvement. This analysis may involve heuristic evaluation or machine learning models.

[0235] Step 3:

[0236] The server sends identified areas for improvement as feedback to the learning agent. This feedback includes details of instances where the response was insufficient and instances where it was successful.

[0237] Step 4:

[0238] The device then initiates retraining of the learning agent based on feedback. Reinforcement learning algorithms set new parameters, and the agent improves itself.

[0239] Step 5:

[0240] Users can use their devices to check the retraining status and improvements made to the AI ​​agent. They can also provide additional data and instructions to the server as needed.

[0241] Step 6:

[0242] The device manages user interactions through natural language processing capabilities, supporting the AI ​​agent in providing optimal responses. This conversational data is also sent back to the server for further analysis.

[0243] Step 7:

[0244] The server monitors the retrained AI agent's new performance and evaluates whether the expected improvements are being seen. If further optimization is deemed necessary, the feedback loop is repeated.

[0245] (Example 1)

[0246] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0247] Traditional artificial intelligence (AI) faces challenges such as its inability to quickly adapt to environmental changes and new data, and the difficulty in continuously improving its performance. Furthermore, it often fails to effectively utilize feedback through natural language interaction with users, and the process of identifying areas for improvement and retraining is frequently done manually, making it inefficient. A system is needed to solve these problems and enhance the flexibility and adaptability of AI.

[0248] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0249] In this invention, the server includes means for autonomously monitoring other artificial intelligences, means for collecting and analyzing performance data of the artificial intelligences, and means for identifying areas for improvement based on the analysis results. This makes it possible to efficiently improve the performance of the artificial intelligences and to quickly adapt to new data and environmental changes.

[0250] "Artificial intelligence" is a program or system that learns autonomously based on data and makes decisions.

[0251] "Monitoring" is the act of continuously checking the status and activity of a system or process and collecting necessary information.

[0252] "Performance data" refers to quantitative or qualitative information used to evaluate the operation and effectiveness of artificial intelligence.

[0253] "Analysis" is the process of analyzing collected data to find meaning and derive specific results or conclusions.

[0254] "Improvements" refer to changes or modifications necessary to enhance the capabilities and performance of artificial intelligence.

[0255] "Retraining" is an additional learning process that uses newly collected information and feedback to further improve the performance of artificial intelligence.

[0256] "Natural language" refers to the language that humans use in their daily lives, and which has a specific grammar and structure.

[0257] "Dialogue" is a form of communication in which information and opinions are exchanged between different individuals or systems.

[0258] "Feedback" refers to evaluations or opinions provided based on specific actions or outcomes, serving as guidance for improvement or adjustment.

[0259] Reinforcement learning is a type of machine learning algorithm in which an agent learns the optimal action by receiving rewards through trial and error.

[0260] "Adaptation" is a response that flexibly responds to changes in the environment and new information, thereby improving behavior and abilities.

[0261] This invention improves the performance of artificial intelligence through a process of collecting, analyzing, and improving various data within an artificial intelligence-based system. The system mainly consists of a server, terminals, and users, and each element works in coordination.

[0262] The server is responsible for autonomously monitoring the performance of the artificial intelligence and collecting log data. Specifically, it uses software used as a log collection tool to collect data such as the content of the AI's dialogue, response time, and response accuracy, and stores it in a database. The server then analyzes the collected data using data analysis software such as Python or R to identify areas for improvement to enhance performance.

[0263] The terminal provides an interface for users to access the system and visualize the current performance and retraining progress of the artificial intelligence. A dedicated dashboard application allows users to check the system status in real time. Furthermore, the artificial intelligence interacts with the user in natural language via the terminal, accumulating the feedback received. SpaCy and Transformers are used as natural language processing frameworks.

[0264] Users can provide additional data to the artificial intelligence (AI) via the system's dashboard, and use that data to guide the AI's retraining. This allows users to provide new sales materials in response to changes in sales strategies. Users can also send queries to the AI ​​using prompts and receive responses. For example, they can expect to receive new information by using a prompt such as, "Please tell me the latest information on new products."

[0265] In this way, through the cooperation of servers, terminals, and users, artificial intelligence can adapt to environmental changes in real time and achieve efficient task processing. The entire system combines a generative AI model with natural language dialogue using prompt sentences to achieve efficient and flexible artificial intelligence operation.

[0266] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0267] Step 1:

[0268] The server collects log data from AI-generated conversations. It receives conversation text data, response times, and response accuracy as input. The server aggregates this data using a log collection tool and stores it in a database. Specifically, periodic data imports are performed into the database. The output obtained at this stage is structured log data that can be used for subsequent analysis.

[0269] Step 2:

[0270] The server analyzes the collected log data. It uses the structured log data obtained in Step 1 as input. The server uses a data analysis program and applies machine learning algorithms to calculate metrics related to the AI's performance and identify areas for improvement. Specifically, data cleaning and feature extraction are performed, and analysis results are generated. The output is a list of specific areas for improvement to enhance the AI's performance.

[0271] Step 3:

[0272] The server generates feedback to the artificial intelligence based on the analysis results and triggers retraining. The list of improvements from Step 2 is used as input. The server uses reinforcement learning to prepare a retraining protocol to update the AI's parameters based on the input data. Specifically, this involves preparing the retraining dataset and updating the model. The output is the updated artificial intelligence model.

[0273] Step 4:

[0274] Users access the system through a terminal to check current performance and retraining progress. The input is the latest AI status information sent from the server. The terminal visualizes this information via a dedicated dashboard and presents it to the user. Specifically, the dashboard displays graphs and metrics that are updated in real time. The output is visualized information for the user to use.

[0275] Step 5:

[0276] The device facilitates interaction between the user and artificial intelligence using natural language. It receives prompt text from the user as input. The device uses a generative AI model to generate the optimal response based on the prompt. Specifically, it analyzes the user's intent through natural language processing and generates a response. Therefore, the output is a natural language response provided to the user.

[0277] (Application Example 1)

[0278] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0279] Current content delivery systems struggle to quickly respond to users' diverse viewing needs and rapidly changing preferences, often resulting in insufficient generation of personalized recommendations. This leads to problems such as the inability to appropriately recommend content that viewers are interested in, and thus the inability to improve the user experience.

[0280] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0281] In this invention, the server includes means for collecting viewing history and user evaluation information, means for generating personalized recommended content based on the collected information, and means for updating the generated recommended content in real time. This enables the utilization of user feedback to timely adjust the personalized recommended content and enhance user satisfaction.

[0282] The "viewing history" refers to data such as records, times, and frequencies of users' viewing of content.

[0283] The "evaluation information" refers to information such as user feedback and evaluation scores on content.

[0284] The "personalized recommended content" is a list of customized content based on each user's viewing history and evaluation information.

[0285] "Updating in real time" means immediately reflecting new user viewing history and evaluation information and changing the recommended content on the spot.

[0286] "Natural language dialogue" is a form of communication in which users use human language with a computer system.

[0287] "Feedback" refers to opinions and feelings provided by users regarding the use experience of content or systems.

[0288] The "recommendation algorithm" is a calculation method for automatically selecting content suitable for users.

[0289] To implement this invention, consider the following system configuration. First, the server functions as the main data processing unit, collecting user viewing history and evaluation information. Specifically, it uses log data obtained from various content platforms and stores it in a database. The server is equipped with a machine learning framework such as TensorFlow, and serves as a foundation for data analysis and reinforcement learning.

[0290] The server uses this data to periodically update its recommendation algorithm, which suggests the most suitable content for each user. To enable real-time updates, a Python script is used to ensure the algorithm always reflects the latest information. Furthermore, user feedback is obtained through the device in a natural language dialogue format, and this information is also used to optimize the recommendation algorithm. The devices used are smartphones and tablets equipped with user-friendly interfaces that accept natural language input.

[0291] For example, if a user provides feedback such as "I enjoyed it" after watching a movie, the server analyzes it and adds other works of similar genres or by similar directors to the recommendation list. Another example of a prompt message is, "I watched some new content. What should I watch next?" In this way, it is possible to increase user engagement and provide a more personalized viewing experience.

[0292] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0293] Step 1:

[0294] The server receives user viewing history and rating information from the content platform. The received data is stored in the database. At this point, the input is viewing history and rating data, and the output is the stored data. The server performs format conversion and integrity checks on the log data and inserts it into the database in a formatted state.

[0295] Step 2:

[0296] The server runs a recommendation algorithm using stored data to generate personalized content recommendations. The input for this step is user data retrieved from a database, and the output is a personalized recommendation list. The server leverages a machine learning framework to extract significant patterns from the data, shaping the most relevant content for each user into a list.

[0297] Step 3:

[0298] The terminal receives natural language feedback from the user and sends it to the server. The input is the user's natural language feedback, and the output is the parsed data sent to the server. The terminal has speech recognition or text input capabilities to format the user's input feedback without ambiguity.

[0299] Step 4:

[0300] The server analyzes the received feedback and readjusts the recommended algorithm. The input to this process is the feedback data, and the output is the updated algorithm parameters. The server uses the feedback to perform reinforcement learning, learning in real time to improve the parameters of the recommended model.

[0301] Step 5:

[0302] The server sends the updated recommended list back to the terminal and displays new recommended content when the user accesses it. The input is the latest recommended data, and the output is the display list on the user terminal. The server transfers data to the terminal quickly via the communication protocol, creating an environment where the user can smoothly check new recommendations.

[0303] Furthermore, an emotion engine for estimating the user's emotion may be combined. That is, the specific processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform specific processing using the user's emotion.

[0304] The present invention is a system that realizes more advanced interaction and effective learning by combining an emotion engine with an AI agent. This system is mainly composed of a server, a terminal, and a user.

[0305] 1. Server Processing

[0306] The server monitors the performance of the AI agent and collects the user's emotion data together with the performance data. The emotion data is obtained using voice tones, expressions, text analysis, etc. For example, in a customer support system, the emotional state is analyzed from the tone of the user's voice and logged in real time.

[0307] 2. Integration of the Emotion Engine

[0308] The server incorporates an emotion engine, and the AI agent analyzes the user's emotion through the emotion engine. This data affects the response of the AI agent, and the response is adjusted to adapt to the user's emotion. For example, when the user is dissatisfied, the AI agent makes a more personal response.

[0309] 3. Feedback and Retraining

[0310] The server sends feedback, including the user's emotional state, to the learning agent. Based on this feedback, the AI ​​agent retrains and adjusts its learning algorithm. This allows the AI ​​agent to strengthen its emotion-based response strategy and improve the accuracy of its responses.

[0311] 4. User Interaction

[0312] Users interact with an AI agent through their device, and their emotional state is recognized by the system. Users can feel that the AI's responses are appropriate and considerate of their emotions. For example, in educational services, the AI ​​can assess a student's emotional state and provide optimal learning support.

[0313] By incorporating an emotion engine in this way, AI agents can achieve interactions that respond to the user's emotions, providing a more personalized experience. This system provides a foundation for AI agents to perform their tasks while flexibly responding to changes in the environment and the user's emotional state.

[0314] The following describes the processing flow.

[0315] Step 1:

[0316] The server initiates interaction with the user, receiving voice input and text data. This data is immediately sent to the emotion engine for analysis.

[0317] Step 2:

[0318] The emotion engine analyzes the tone of voice, vocabulary, and context of the text to identify the user's emotional state. This analysis is returned to the server as a quantified emotion index.

[0319] Step 3:

[0320] The server adjusts the AI ​​agent's response based on sentiment indicators. For example, if a user expresses dissatisfaction, the AI ​​agent is instructed to apologize and offer quick solutions to resolve the problem.

[0321] Step 4:

[0322] On the device, the AI ​​agent provides the user with tailored responses. This involves using natural speech or text to engage in emotionally conscious dialogue.

[0323] Step 5:

[0324] Through interaction with the AI ​​agent, users receive responses tailored to their needs. Emotionally sensitive responses make users feel more supported.

[0325] Step 6:

[0326] The server sends user feedback and sentiment data to the learning agent, prompting retraining. Based on this data, the learning agent performs reinforcement learning to develop more refined sentiment recognition and response strategies.

[0327] Step 7:

[0328] The device releases an updated AI agent based on the retraining content, preparing for future user interactions. This process allows the AI ​​agent to continuously evolve and adapt more closely to the user's emotions.

[0329] (Example 2)

[0330] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0331] Traditional AI systems have struggled to provide responses that take into account the user's emotional state, resulting in shortcomings in the user experience. Furthermore, the lack of means to analyze emotions in real time and optimize responses based on that analysis meant that user feedback could not be effectively incorporated, limiting the improvement of the accuracy of the learning models.

[0332] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0333] In this invention, the server includes means for acquiring user data based on an information processing device for analyzing emotional states, means for identifying the user's emotional state by analyzing the user data, and means for adjusting the AI ​​system's response according to the emotional state. This enables the AI ​​system to provide interactions that respond to the user's emotions, improving the quality of the user experience and simultaneously improving the accuracy of the learning model that reflects the feedback.

[0334] An "information processing device" is a device that collects and analyzes user data to identify emotional states.

[0335] "User data" refers to information such as voice, facial expressions, and text acquired during interactions with users.

[0336] "Emotional state" refers to the results of analysis to identify the user's emotions, and serves as fundamental information for adjusting the AI ​​system's response.

[0337] "AI system response" refers to the content of the conversation with the AI ​​agent, which is adjusted based on the user's emotional state.

[0338] "Feedback information" refers to information that users provide regarding their evaluations and suggestions for improvement in response to the AI ​​system's actions.

[0339] A "learning model" is a mathematical model used to improve the response accuracy of an AI system based on data analysis.

[0340] The system of this invention aims to adjust the response of an AI agent based on the user's emotions. The system mainly consists of a server, a terminal, and a user, and provides the environment for each function to operate.

[0341] The server houses the information processing unit that forms the core of the AI ​​agent, and it collects and analyzes user data. Various data collection methods are used here, such as speech recognition software and facial expression recognition tools. For example, general speech analysis service software is used for speech recognition, processing changes in speech tone and other data as digital data. For facial expression recognition, facial expression analysis software is used to quantify the user's emotional state in real time and send the analysis data to the server. Based on this analysis data, the AI ​​system's response is dynamically optimized.

[0342] The terminal functions as a means of providing a user interface and acts as a border for information exchange between the user and the AI ​​agent. Specifically, a chat application or voice interface runs on the terminal, allowing the user to interact with the AI ​​agent in natural language. When the user provides voice input to the terminal, the voice is analyzed on a server, and an appropriate response based on the user's emotional state is returned to the terminal. This allows the user to feel that the AI's responses are considerate of their emotions.

[0343] The feedback information that users provide to the AI ​​agent is collected by the server and used to retrain the learning model. This feedback is important data for improving the AI's performance, and the generative AI model is used to improve response accuracy. The generative AI model used at this time is made to generate appropriate responses using prompt statements.

[0344] A concrete example of a prompt might be, "In the following educational scenario, how should the AI ​​agent respond if a student is struggling to understand the learning material?" By using this prompt to obtain a response from the generative AI model, the AI ​​agent can provide more empathetic and effective support to the user.

[0345] In this way, the present invention provides a system that flexibly responds to user emotions, achieving improved user experience and enhanced performance of AI agents.

[0346] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0347] Step 1:

[0348] The server begins collecting user data. When a user uses a terminal to input voice or text, the terminal sends this data to the server. The input data is analyzed by speech recognition software and text analysis tools. Specifically, in the case of voice input, it is digitized and converted into text format through natural language processing. The output of this process is the text data to be analyzed.

[0349] Step 2:

[0350] The server identifies emotional data based on the collected text data. The emotion analysis engine within the server receives this text data and analyzes the user's emotional state by comprehensively considering the results of voice tone and facial recognition. This analysis is performed by an emotion analysis algorithm, and the output is an emotion category (e.g., joy, anger, sadness).

[0351] Step 3:

[0352] The server inputs the obtained emotion categories into a generative AI model to generate a response appropriate for the user. By providing the emotion categories along with prompts to the model, the AI ​​generates an appropriate response that takes emotions into consideration. Specifically, an example prompt such as "What is the appropriate response if the user is angry?" is set, and the generative AI model is used to generate a response to the user. As a result, the AI ​​agent receives a customized response to the user as output.

[0353] Step 4:

[0354] The server forwards the generated response to the terminal. The terminal receives this response and displays it to the user through the user interface. Specifically, if it is in text format, it is displayed in a chat window on the screen, and if it is a voice response, the text is converted into speech using speech synthesis technology. The user interacts with the AI ​​agent through this response.

[0355] Step 5:

[0356] The user provides feedback on the AI ​​agent's responses. The user sends feedback input to the server via their device, indicating satisfaction and areas for improvement. The server accumulates this feedback and incorporates it into its learning algorithm. The AI ​​model's performance is retrained using this data to improve future response accuracy. The output is an improved, trained model.

[0357] (Application Example 2)

[0358] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0359] Conventional AI agents struggle to dynamically adjust their responses based on user emotions, and appropriate dialogue is especially crucial in emotionally sensitive environments such as caregiving settings. However, existing systems are unable to adequately analyze users' emotions and provide adaptive responses instantly, making it difficult to provide personalized services that take emotions into consideration.

[0360] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0361] In this invention, the server includes means for autonomously monitoring the performance of other AI agents, means for analyzing the user's emotions, and means for the AI ​​agent to adjust its response based on the emotion analysis. This makes it possible to instantly analyze the user's emotions in settings such as nursing care and provide a response that is in harmony with them, thereby providing a more personalized experience.

[0362] "Autonomously monitoring the performance of other AI agents" means that an AI agent independently observes the activities of other AI agents and continuously evaluates their performance.

[0363] "Identifying necessary improvements" means analyzing data on the AI ​​agent's performance to determine the specific elements and areas that need improvement.

[0364] "Retraining" means updating the AI ​​agent's learning model based on identified areas for improvement and running the learning process again to enhance its performance.

[0365] "Emotional analysis," which analyzes the emotions of users, is a process that collects emotional data from the user's facial expressions, tone of voice, text, etc., and clarifies their emotional state.

[0366] "Adjusting responses" means modifying the output of the AI ​​agent based on the results of sentiment analysis to provide users with more emotionally sensitive responses.

[0367] "Natural language dialogue" refers to the communication process between a user and an AI agent using the language that humans use in everyday life.

[0368] A "reinforcement learning algorithm" is a learning process that improves the efficiency of an AI agent by selecting actions based on rewards from the environment and repeatedly refining those selections.

[0369] "Providing feedback" means sending back to the system the user's reactions and interaction results obtained by the AI ​​agent as useful information for performance improvement.

[0370] This invention aims to enhance individualized care by introducing AI agents into small-group living environments such as caregiving settings. The specific elements for implementing this invention are described below.

[0371] The server autonomously monitors the performance of other AI agents and runs software for emotion analysis. For example, it uses the Google Cloud Speech-to-Text API to analyze tone from speech and the Microsoft Face API for facial expression analysis. The server collects this data and evaluates the user's emotional state in real time.

[0372] The terminals used are smartphones and caregiving robots, which acquire audio and video data through sensor devices such as microphones and cameras. The acquired data is sent to a server, where an AI agent generates an appropriate response based on the analysis results. The generated response is provided to the user in natural language through a dialogue module. The AI ​​agent receives feedback within a certain range periodically and improves its learning algorithm with new data.

[0373] Users interact with the AI ​​agent through everyday conversations, evaluating the emotional adaptability of its responses. For example, if the AI ​​agent determines that the user is feeling down, it may suggest prompts such as, "Did anything fun happen today? Shall I put on some relaxing music?" to facilitate a more personalized experience.

[0374] In this way, the entire system adjusts according to the user's emotional changes and provides more adaptive help, aiming to improve efficiency and quality in care settings.

[0375] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0376] Step 1:

[0377] The device uses a microphone and camera to acquire the user's voice and video data in real time. Voice data includes tone, volume, and speaking speed, while video data captures the user's facial expressions and movements. This data is then transferred directly to the server.

[0378] Step 2:

[0379] The server converts the received audio data into text using the Google Cloud Speech-to-Text API and then performs emotional analysis on the voice. It analyzes variations in volume and speaking speed from the audio to profile the emotional state. As a result, it obtains a numerical evaluation of the user's emotions (e.g., excitement, joy, sadness, anger).

[0380] Step 3:

[0381] The server analyzes the video data and uses facial expression analysis tools such as the Microsoft Face API to determine the user's emotions from their facial expressions. Specifically, it analyzes facial feature points (eyes, eyebrows, mouth angle, etc.) and maps changes in facial expressions in the video to emotion categories (e.g., smile, confusion, surprise).

[0382] Step 4:

[0383] By combining this sentiment analysis data, the server uses an AI model to determine the user's current emotional state. Based on this determination, it generates a response from the AI ​​agent and creates a prompt. For example, if it determines that the user is feeling anxious, it prepares a response that will provide reassurance.

[0384] Step 5:

[0385] The AI ​​agent sends the generated prompt message back to the terminal, which then presents it to the user as voice or text. For example, it can engage in a more human-like conversation, such as, "Did anything fun happen today? Shall I put on some relaxing music?"

[0386] Step 6:

[0387] After receiving a response from the AI ​​agent, the user inputs that feedback as a separate dialogue. The device sends this feedback to the server as new input data. The server uses the received feedback as training data for the AI ​​agent and retrains it to improve future responses.

[0388] This entire process enables the entire system to continuously provide a flexible and adaptive experience that responds to the user's changing emotions.

[0389] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0390] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0391] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0392] [Third Embodiment]

[0393] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0394] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0395] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0396] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0397] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0398] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0399] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0400] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0401] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0402] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0403] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0404] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0405] This invention is an advanced system for educating and training AI agents, and is implemented as follows: The system mainly consists of a server, terminals, and users.

[0406] 1. Server Role

[0407] The server autonomously monitors the performance of AI agents and collects log data. For example, in the case of customer service agents, it records the content of customer interactions, response times, and accurate responses. The server analyzes the collected data to identify areas for improvement to enhance performance.

[0408] 2. Feedback and Retraining

[0409] The server sends identified areas for improvement as feedback to the learning agent. The learning agent uses the feedback to retrain itself using a reinforcement learning algorithm. For example, if the price of a product changes, the agent updates its response based on new information. The reinforcement learning algorithm adapts to environmental changes in real time and performs further optimizations.

[0410] 3. User Interaction

[0411] Users can access the system via their devices to visualize the AI ​​agent's current performance and retraining progress. If necessary, users can also provide additional data to the system and guide the learning process. For example, they can provide new sales materials in response to changes in sales strategy.

[0412] 4. Dialogue using natural language

[0413] On the device, an AI agent can interact with the user in natural language. Based on user queries, the AI ​​provides optimal answers and accumulates feedback. Through this interaction, the system understands the user's intent and uses this data to guide better results.

[0414] These components work together to enable the AI ​​agent to continuously evolve, gaining high flexibility and adaptability to meet the needs of businesses. The entire system provides a robust foundation for the AI ​​agent to respond quickly to environmental changes and perform tasks efficiently.

[0415] The following describes the processing flow.

[0416] Step 1:

[0417] The server collects AI agent operation data in real time and stores it as logs. This includes AI response time, accuracy, and user interaction.

[0418] Step 2:

[0419] The server analyzes the collected data to identify areas where the AI ​​agent's performance needs improvement. This analysis may involve heuristic evaluation or machine learning models.

[0420] Step 3:

[0421] The server sends identified areas for improvement as feedback to the learning agent. This feedback includes details of instances where the response was insufficient and instances where it was successful.

[0422] Step 4:

[0423] The device then initiates retraining of the learning agent based on feedback. Reinforcement learning algorithms set new parameters, and the agent improves itself.

[0424] Step 5:

[0425] Users can use their devices to check the retraining status and improvements made to the AI ​​agent. They can also provide additional data and instructions to the server as needed.

[0426] Step 6:

[0427] The device manages user interactions through natural language processing capabilities, supporting the AI ​​agent in providing optimal responses. This conversational data is also sent back to the server for further analysis.

[0428] Step 7:

[0429] The server monitors the retrained AI agent's new performance and evaluates whether the expected improvements are being seen. If further optimization is deemed necessary, the feedback loop is repeated.

[0430] (Example 1)

[0431] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0432] Traditional artificial intelligence (AI) faces challenges such as its inability to quickly adapt to environmental changes and new data, and the difficulty in continuously improving its performance. Furthermore, it often fails to effectively utilize feedback through natural language interaction with users, and the process of identifying areas for improvement and retraining is frequently done manually, making it inefficient. A system is needed to solve these problems and enhance the flexibility and adaptability of AI.

[0433] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0434] In this invention, the server includes means for autonomously monitoring other artificial intelligences, means for collecting and analyzing performance data of the artificial intelligences, and means for identifying areas for improvement based on the analysis results. This makes it possible to efficiently improve the performance of the artificial intelligences and to quickly adapt to new data and environmental changes.

[0435] "Artificial intelligence" is a program or system that learns autonomously based on data and makes decisions.

[0436] "Monitoring" is the act of continuously checking the status and activity of a system or process and collecting necessary information.

[0437] "Performance data" refers to quantitative or qualitative information used to evaluate the operation and effectiveness of artificial intelligence.

[0438] "Analysis" is the process of analyzing collected data to find meaning and derive specific results or conclusions.

[0439] "Improvements" refer to changes or modifications necessary to enhance the capabilities and performance of artificial intelligence.

[0440] "Retraining" is an additional learning process that uses newly collected information and feedback to further improve the performance of artificial intelligence.

[0441] "Natural language" refers to the language that humans use in their daily lives, and which has a specific grammar and structure.

[0442] "Dialogue" is a form of communication in which information and opinions are exchanged between different individuals or systems.

[0443] "Feedback" refers to evaluations or opinions provided based on specific actions or outcomes, serving as guidance for improvement or adjustment.

[0444] Reinforcement learning is a type of machine learning algorithm in which an agent learns the optimal action by receiving rewards through trial and error.

[0445] "Adaptation" is a response that flexibly responds to changes in the environment and new information, thereby improving behavior and abilities.

[0446] This invention improves the performance of artificial intelligence through a process of collecting, analyzing, and improving various data within an artificial intelligence-based system. The system mainly consists of a server, terminals, and users, and each element works in coordination.

[0447] The server is responsible for autonomously monitoring the performance of the artificial intelligence and collecting log data. Specifically, it uses software used as a log collection tool to collect data such as the content of the AI's dialogue, response time, and response accuracy, and stores it in a database. The server then analyzes the collected data using data analysis software such as Python or R to identify areas for improvement to enhance performance.

[0448] The terminal provides an interface for users to access the system and visualize the current performance and retraining progress of the artificial intelligence. A dedicated dashboard application allows users to check the system status in real time. Furthermore, the artificial intelligence interacts with the user in natural language via the terminal, accumulating the feedback received. SpaCy and Transformers are used as natural language processing frameworks.

[0449] Users can provide additional data to the artificial intelligence (AI) via the system's dashboard, and use that data to guide the AI's retraining. This allows users to provide new sales materials in response to changes in sales strategies. Users can also send queries to the AI ​​using prompts and receive responses. For example, they can expect to receive new information by using a prompt such as, "Please tell me the latest information on new products."

[0450] In this way, through the cooperation of servers, terminals, and users, artificial intelligence can adapt to environmental changes in real time and achieve efficient task processing. The entire system combines a generative AI model with natural language dialogue using prompt sentences to achieve efficient and flexible artificial intelligence operation.

[0451] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0452] Step 1:

[0453] The server collects log data from AI-generated conversations. It receives conversation text data, response times, and response accuracy as input. The server aggregates this data using a log collection tool and stores it in a database. Specifically, periodic data imports are performed into the database. The output obtained at this stage is structured log data that can be used for subsequent analysis.

[0454] Step 2:

[0455] The server analyzes the collected log data. It uses the structured log data obtained in Step 1 as input. The server uses a data analysis program and applies machine learning algorithms to calculate metrics related to the AI's performance and identify areas for improvement. Specifically, data cleaning and feature extraction are performed, and analysis results are generated. The output is a list of specific areas for improvement to enhance the AI's performance.

[0456] Step 3:

[0457] The server generates feedback to the artificial intelligence based on the analysis results and triggers retraining. The list of improvements from Step 2 is used as input. The server uses reinforcement learning to prepare a retraining protocol to update the AI's parameters based on the input data. Specifically, this involves preparing the retraining dataset and updating the model. The output is the updated artificial intelligence model.

[0458] Step 4:

[0459] Users access the system through a terminal to check current performance and retraining progress. The input is the latest AI status information sent from the server. The terminal visualizes this information via a dedicated dashboard and presents it to the user. Specifically, the dashboard displays graphs and metrics that are updated in real time. The output is visualized information for the user to use.

[0460] Step 5:

[0461] The device facilitates interaction between the user and artificial intelligence using natural language. It receives prompt text from the user as input. The device uses a generative AI model to generate the optimal response based on the prompt. Specifically, it analyzes the user's intent through natural language processing and generates a response. Therefore, the output is a natural language response provided to the user.

[0462] (Application Example 1)

[0463] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0464] Current content delivery systems struggle to quickly respond to users' diverse viewing needs and rapidly changing preferences, often resulting in insufficient generation of personalized recommendations. This leads to problems such as the inability to appropriately recommend content that viewers are interested in, and thus the inability to improve the user experience.

[0465] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0466] In this invention, the server includes means for collecting viewing history and user evaluation information, means for generating personalized recommendations based on the collected information, and means for updating the generated recommendations in real time. This makes it possible to leverage user feedback to adjust personalized recommendations in a timely manner and increase user satisfaction.

[0467] "Viewing history" refers to data such as records, duration, and frequency of content viewing by a user.

[0468] "Rating information" refers to information such as feedback and rating scores that users have given to content.

[0469] "Personalized recommendations" refer to a list of content customized based on each user's viewing history and rating information.

[0470] "Real-time updates" means instantly reflecting users' new viewing history and rating information, and changing recommendations on the spot.

[0471] "Natural language dialogue" is a form of communication in which users interact with computer systems using human language.

[0472] "Feedback" refers to opinions and impressions that users provide about their experience using content or a system.

[0473] A "recommendation algorithm" is a computational method for automatically selecting content that is suitable for the user.

[0474] To implement this invention, consider the following system configuration. First, the server functions as the main data processing unit, collecting user viewing history and evaluation information. Specifically, it uses log data obtained from various content platforms and stores it in a database. The server is equipped with a machine learning framework such as TensorFlow, and serves as a foundation for data analysis and reinforcement learning.

[0475] The server uses this data to periodically update its recommendation algorithm, which suggests the most suitable content for each user. To enable real-time updates, a Python script is used to ensure the algorithm always reflects the latest information. Furthermore, user feedback is obtained through the device in a natural language dialogue format, and this information is also used to optimize the recommendation algorithm. The devices used are smartphones and tablets equipped with user-friendly interfaces that accept natural language input.

[0476] For example, if a user provides feedback such as "I enjoyed it" after watching a movie, the server analyzes it and adds other works of similar genres or by similar directors to the recommendation list. Another example of a prompt message is, "I watched some new content. What should I watch next?" In this way, it is possible to increase user engagement and provide a more personalized viewing experience.

[0477] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0478] Step 1:

[0479] The server receives user viewing history and rating information from the content platform. The received data is stored in the database. At this point, the input is viewing history and rating data, and the output is the stored data. The server performs format conversion and integrity checks on the log data and inserts it into the database in a formatted state.

[0480] Step 2:

[0481] The server runs a recommendation algorithm using stored data to generate personalized content recommendations. The input for this step is user data retrieved from a database, and the output is a personalized recommendation list. The server leverages a machine learning framework to extract significant patterns from the data, shaping the most relevant content for each user into a list.

[0482] Step 3:

[0483] The terminal receives natural language feedback from the user and sends it to the server. The input is the user's natural language feedback, and the output is the parsed data sent to the server. The terminal has speech recognition or text input capabilities to format the user's input feedback without ambiguity.

[0484] Step 4:

[0485] The server analyzes the received feedback and readjusts the recommended algorithm. The input to this process is the feedback data, and the output is the updated algorithm parameters. The server uses the feedback to perform reinforcement learning, learning in real time to improve the parameters of the recommended model.

[0486] Step 5:

[0487] The server sends the updated recommendation list back to the terminal, displaying the new recommendations when the user accesses it. The input is the latest recommendation data, and the output is the display list on the user's terminal. The server quickly transfers the data to the terminal via a communication protocol, creating an environment where users can smoothly check the new recommendations.

[0488] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0489] This invention is a system that combines an emotion engine with an AI agent to achieve more advanced interaction and effective learning. This system mainly consists of a server, terminals, and users.

[0490] 1. Server processing

[0491] The server monitors the performance of the AI ​​agent and collects user emotion data along with performance data. Emotion data is obtained using methods such as voice tone, facial expressions, and text analysis. For example, in a customer support system, the emotional state of a user can be analyzed from their voice tone and logged in real time.

[0492] 2. Integration of the Emotional Engine

[0493] The server incorporates an emotion engine, and the AI ​​agent analyzes the user's emotions through this engine. This data influences the AI ​​agent's response, which is then adjusted to suit the user's emotions. For example, if the user is dissatisfied, the AI ​​agent will provide a more empathetic response.

[0494] 3. Feedback and Retraining

[0495] The server sends feedback, including the user's emotional state, to the learning agent. Based on this feedback, the AI ​​agent retrains and adjusts its learning algorithm. This allows the AI ​​agent to strengthen its emotion-based response strategy and improve the accuracy of its responses.

[0496] 4. User Interaction

[0497] Users interact with an AI agent through their device, and their emotional state is recognized by the system. Users can feel that the AI's responses are appropriate and considerate of their emotions. For example, in educational services, the AI ​​can assess a student's emotional state and provide optimal learning support.

[0498] By incorporating an emotion engine in this way, AI agents can achieve interactions that respond to the user's emotions, providing a more personalized experience. This system provides a foundation for AI agents to perform their tasks while flexibly responding to changes in the environment and the user's emotional state.

[0499] The following describes the processing flow.

[0500] Step 1:

[0501] The server initiates interaction with the user, receiving voice input and text data. This data is immediately sent to the emotion engine for analysis.

[0502] Step 2:

[0503] The emotion engine analyzes the tone of voice, vocabulary, and context of the text to identify the user's emotional state. This analysis is returned to the server as a quantified emotion index.

[0504] Step 3:

[0505] The server adjusts the AI ​​agent's response based on sentiment indicators. For example, if a user expresses dissatisfaction, the AI ​​agent is instructed to apologize and offer quick solutions to resolve the problem.

[0506] Step 4:

[0507] On the device, the AI ​​agent provides the user with tailored responses. This involves using natural speech or text to engage in emotionally conscious dialogue.

[0508] Step 5:

[0509] Through interaction with the AI ​​agent, users receive responses tailored to their needs. Emotionally sensitive responses make users feel more supported.

[0510] Step 6:

[0511] The server sends user feedback and sentiment data to the learning agent, prompting retraining. Based on this data, the learning agent performs reinforcement learning to develop more refined sentiment recognition and response strategies.

[0512] Step 7:

[0513] The device releases an updated AI agent based on the retraining content, preparing for future user interactions. This process allows the AI ​​agent to continuously evolve and adapt more closely to the user's emotions.

[0514] (Example 2)

[0515] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0516] Traditional AI systems have struggled to provide responses that take into account the user's emotional state, resulting in shortcomings in the user experience. Furthermore, the lack of means to analyze emotions in real time and optimize responses based on that analysis meant that user feedback could not be effectively incorporated, limiting the improvement of the accuracy of the learning models.

[0517] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0518] In this invention, the server includes means for acquiring user data based on an information processing device for analyzing emotional states, means for identifying the user's emotional state by analyzing the user data, and means for adjusting the AI ​​system's response according to the emotional state. This enables the AI ​​system to provide interactions that respond to the user's emotions, improving the quality of the user experience and simultaneously improving the accuracy of the learning model that reflects the feedback.

[0519] An "information processing device" is a device that collects and analyzes user data to identify emotional states.

[0520] "User data" refers to information such as voice, facial expressions, and text acquired during interactions with users.

[0521] "Emotional state" refers to the results of analysis to identify the user's emotions, and serves as fundamental information for adjusting the AI ​​system's response.

[0522] "AI system response" refers to the content of the conversation with the AI ​​agent, which is adjusted based on the user's emotional state.

[0523] "Feedback information" refers to information that users provide regarding their evaluations and suggestions for improvement in response to the AI ​​system's actions.

[0524] A "learning model" is a mathematical model used to improve the response accuracy of an AI system based on data analysis.

[0525] The system of this invention aims to adjust the response of an AI agent based on the user's emotions. The system mainly consists of a server, a terminal, and a user, and provides the environment for each function to operate.

[0526] The server houses the information processing unit that forms the core of the AI ​​agent, and it collects and analyzes user data. Various data collection methods are used here, such as speech recognition software and facial expression recognition tools. For example, general speech analysis service software is used for speech recognition, processing changes in speech tone and other data as digital data. For facial expression recognition, facial expression analysis software is used to quantify the user's emotional state in real time and send the analysis data to the server. Based on this analysis data, the AI ​​system's response is dynamically optimized.

[0527] The terminal functions as a means of providing a user interface and acts as a border for information exchange between the user and the AI ​​agent. Specifically, a chat application or voice interface runs on the terminal, allowing the user to interact with the AI ​​agent in natural language. When the user provides voice input to the terminal, the voice is analyzed on a server, and an appropriate response based on the user's emotional state is returned to the terminal. This allows the user to feel that the AI's responses are considerate of their emotions.

[0528] The feedback information that users provide to the AI ​​agent is collected by the server and used to retrain the learning model. This feedback is important data for improving the AI's performance, and the generative AI model is used to improve response accuracy. The generative AI model used at this time is made to generate appropriate responses using prompt statements.

[0529] A concrete example of a prompt might be, "In the following educational scenario, how should the AI ​​agent respond if a student is struggling to understand the learning material?" By using this prompt to obtain a response from the generative AI model, the AI ​​agent can provide more empathetic and effective support to the user.

[0530] In this way, the present invention provides a system that flexibly responds to user emotions, achieving improved user experience and enhanced performance of AI agents.

[0531] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0532] Step 1:

[0533] The server begins collecting user data. When a user uses a terminal to input voice or text, the terminal sends this data to the server. The input data is analyzed by speech recognition software and text analysis tools. Specifically, in the case of voice input, it is digitized and converted into text format through natural language processing. The output of this process is the text data to be analyzed.

[0534] Step 2:

[0535] The server identifies emotional data based on the collected text data. The emotion analysis engine within the server receives this text data and analyzes the user's emotional state by comprehensively considering the results of voice tone and facial recognition. This analysis is performed by an emotion analysis algorithm, and the output is an emotion category (e.g., joy, anger, sadness).

[0536] Step 3:

[0537] The server inputs the obtained emotion categories into a generative AI model to generate a response appropriate for the user. By providing the emotion categories along with prompts to the model, the AI ​​generates an appropriate response that takes emotions into consideration. Specifically, an example prompt such as "What is the appropriate response if the user is angry?" is set, and the generative AI model is used to generate a response to the user. As a result, the AI ​​agent receives a customized response to the user as output.

[0538] Step 4:

[0539] The server forwards the generated response to the terminal. The terminal receives this response and displays it to the user through the user interface. Specifically, if it is in text format, it is displayed in a chat window on the screen, and if it is a voice response, the text is converted into speech using speech synthesis technology. The user interacts with the AI ​​agent through this response.

[0540] Step 5:

[0541] The user provides feedback on the AI ​​agent's responses. The user sends feedback input to the server via their device, indicating satisfaction and areas for improvement. The server accumulates this feedback and incorporates it into its learning algorithm. The AI ​​model's performance is retrained using this data to improve future response accuracy. The output is an improved, trained model.

[0542] (Application Example 2)

[0543] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0544] Conventional AI agents struggle to dynamically adjust their responses based on user emotions, and appropriate dialogue is especially crucial in emotionally sensitive environments such as caregiving settings. However, existing systems are unable to adequately analyze users' emotions and provide adaptive responses instantly, making it difficult to provide personalized services that take emotions into consideration.

[0545] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0546] In this invention, the server includes means for autonomously monitoring the performance of other AI agents, means for analyzing the user's emotions, and means for the AI ​​agent to adjust its response based on the emotion analysis. This makes it possible to instantly analyze the user's emotions in settings such as nursing care and provide a response that is in harmony with them, thereby providing a more personalized experience.

[0547] "Autonomously monitoring the performance of other AI agents" means that an AI agent independently observes the activities of other AI agents and continuously evaluates their performance.

[0548] "Identifying necessary improvements" means analyzing data on the AI ​​agent's performance to determine the specific elements and areas that need improvement.

[0549] "Retraining" means updating the AI ​​agent's learning model based on identified areas for improvement and running the learning process again to enhance its performance.

[0550] "Emotional analysis," which analyzes the emotions of users, is a process that collects emotional data from the user's facial expressions, tone of voice, text, etc., and clarifies their emotional state.

[0551] "Adjusting responses" means modifying the output of the AI ​​agent based on the results of sentiment analysis to provide users with more emotionally sensitive responses.

[0552] "Natural language dialogue" refers to the communication process between a user and an AI agent using the language that humans use in everyday life.

[0553] A "reinforcement learning algorithm" is a learning process that improves the efficiency of an AI agent by selecting actions based on rewards from the environment and repeatedly refining those selections.

[0554] "Providing feedback" means sending back to the system the user's reactions and interaction results obtained by the AI ​​agent as useful information for performance improvement.

[0555] This invention aims to enhance individualized care by introducing AI agents into small-group living environments such as caregiving settings. The specific elements for implementing this invention are described below.

[0556] The server autonomously monitors the performance of other AI agents and runs software for emotion analysis. For example, it uses the Google Cloud Speech-to-Text API to analyze tone from speech and the Microsoft Face API for facial expression analysis. The server collects this data and evaluates the user's emotional state in real time.

[0557] The terminals used are smartphones and caregiving robots, which acquire audio and video data through sensor devices such as microphones and cameras. The acquired data is sent to a server, where an AI agent generates an appropriate response based on the analysis results. The generated response is provided to the user in natural language through a dialogue module. The AI ​​agent receives feedback within a certain range periodically and improves its learning algorithm with new data.

[0558] Users interact with the AI ​​agent through everyday conversations, evaluating the emotional adaptability of its responses. For example, if the AI ​​agent determines that the user is feeling down, it may suggest prompts such as, "Did anything fun happen today? Shall I put on some relaxing music?" to facilitate a more personalized experience.

[0559] In this way, the entire system adjusts according to the user's emotional changes and provides more adaptive help, aiming to improve efficiency and quality in care settings.

[0560] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0561] Step 1:

[0562] The device uses a microphone and camera to acquire the user's voice and video data in real time. Voice data includes tone, volume, and speaking speed, while video data captures the user's facial expressions and movements. This data is then transferred directly to the server.

[0563] Step 2:

[0564] The server converts the received audio data into text using the Google Cloud Speech-to-Text API and then performs emotional analysis on the voice. It analyzes variations in volume and speaking speed from the audio to profile the emotional state. As a result, it obtains a numerical evaluation of the user's emotions (e.g., excitement, joy, sadness, anger).

[0565] Step 3:

[0566] The server analyzes the video data and uses facial expression analysis tools such as the Microsoft Face API to determine the user's emotions from their facial expressions. Specifically, it analyzes facial feature points (eyes, eyebrows, mouth angle, etc.) and maps changes in facial expressions in the video to emotion categories (e.g., smile, confusion, surprise).

[0567] Step 4:

[0568] By combining this sentiment analysis data, the server uses an AI model to determine the user's current emotional state. Based on this determination, it generates a response from the AI ​​agent and creates a prompt. For example, if it determines that the user is feeling anxious, it prepares a response that will provide reassurance.

[0569] Step 5:

[0570] The AI ​​agent sends the generated prompt message back to the terminal, which then presents it to the user as voice or text. For example, it can engage in a more human-like conversation, such as, "Did anything fun happen today? Shall I put on some relaxing music?"

[0571] Step 6:

[0572] After receiving a response from the AI ​​agent, the user inputs that feedback as a separate dialogue. The device sends this feedback to the server as new input data. The server uses the received feedback as training data for the AI ​​agent and retrains it to improve future responses.

[0573] This entire process enables the entire system to continuously provide a flexible and adaptive experience that responds to the user's changing emotions.

[0574] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0575] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0576] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0577] [Fourth Embodiment]

[0578] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0579] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[0580] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0581] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0582] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0583] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0584] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0585] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0586] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0587] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0588] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0589] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0590] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0591] This invention is an advanced system for educating and training AI agents, and is implemented as follows: The system mainly consists of a server, terminals, and users.

[0592] 1. Server Role

[0593] The server autonomously monitors the performance of AI agents and collects log data. For example, in the case of customer service agents, it records the content of customer interactions, response times, and accurate responses. The server analyzes the collected data to identify areas for improvement to enhance performance.

[0594] 2. Feedback and Retraining

[0595] The server sends identified areas for improvement as feedback to the learning agent. The learning agent uses the feedback to retrain itself using a reinforcement learning algorithm. For example, if the price of a product changes, the agent updates its response based on new information. The reinforcement learning algorithm adapts to environmental changes in real time and performs further optimizations.

[0596] 3. User Interaction

[0597] Users can access the system via their devices to visualize the AI ​​agent's current performance and retraining progress. If necessary, users can also provide additional data to the system and guide the learning process. For example, they can provide new sales materials in response to changes in sales strategy.

[0598] 4. Dialogue using natural language

[0599] On the device, an AI agent can interact with the user in natural language. Based on user queries, the AI ​​provides optimal answers and accumulates feedback. Through this interaction, the system understands the user's intent and uses this data to guide better results.

[0600] These components work together to enable the AI ​​agent to continuously evolve, gaining high flexibility and adaptability to meet the needs of businesses. The entire system provides a robust foundation for the AI ​​agent to respond quickly to environmental changes and perform tasks efficiently.

[0601] The following describes the processing flow.

[0602] Step 1:

[0603] The server collects AI agent operation data in real time and stores it as logs. This includes AI response time, accuracy, and user interaction.

[0604] Step 2:

[0605] The server analyzes the collected data to identify areas where the AI ​​agent's performance needs improvement. This analysis may involve heuristic evaluation or machine learning models.

[0606] Step 3:

[0607] The server sends identified areas for improvement as feedback to the learning agent. This feedback includes details of instances where the response was insufficient and instances where it was successful.

[0608] Step 4:

[0609] The device then initiates retraining of the learning agent based on feedback. Reinforcement learning algorithms set new parameters, and the agent improves itself.

[0610] Step 5:

[0611] Users can use their devices to check the retraining status and improvements made to the AI ​​agent. They can also provide additional data and instructions to the server as needed.

[0612] Step 6:

[0613] The device manages user interactions through natural language processing capabilities, supporting the AI ​​agent in providing optimal responses. This conversational data is also sent back to the server for further analysis.

[0614] Step 7:

[0615] The server monitors the retrained AI agent's new performance and evaluates whether the expected improvements are being seen. If further optimization is deemed necessary, the feedback loop is repeated.

[0616] (Example 1)

[0617] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0618] Traditional artificial intelligence (AI) faces challenges such as its inability to quickly adapt to environmental changes and new data, and the difficulty in continuously improving its performance. Furthermore, it often fails to effectively utilize feedback through natural language interaction with users, and the process of identifying areas for improvement and retraining is frequently done manually, making it inefficient. A system is needed to solve these problems and enhance the flexibility and adaptability of AI.

[0619] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0620] In this invention, the server includes means for autonomously monitoring other artificial intelligences, means for collecting and analyzing performance data of the artificial intelligences, and means for identifying areas for improvement based on the analysis results. This makes it possible to efficiently improve the performance of the artificial intelligences and to quickly adapt to new data and environmental changes.

[0621] "Artificial intelligence" is a program or system that learns autonomously based on data and makes decisions.

[0622] "Monitoring" is the act of continuously checking the status and activity of a system or process and collecting necessary information.

[0623] "Performance data" refers to quantitative or qualitative information used to evaluate the operation and effectiveness of artificial intelligence.

[0624] "Analysis" is the process of analyzing collected data to find meaning and derive specific results or conclusions.

[0625] "Improvements" refer to changes or modifications necessary to enhance the capabilities and performance of artificial intelligence.

[0626] "Retraining" is an additional learning process that uses newly collected information and feedback to further improve the performance of artificial intelligence.

[0627] "Natural language" refers to the language that humans use in their daily lives, and which has a specific grammar and structure.

[0628] "Dialogue" is a form of communication in which information and opinions are exchanged between different individuals or systems.

[0629] "Feedback" refers to evaluations or opinions provided based on specific actions or outcomes, serving as guidance for improvement or adjustment.

[0630] Reinforcement learning is a type of machine learning algorithm in which an agent learns the optimal action by receiving rewards through trial and error.

[0631] "Adaptation" is a response that flexibly responds to changes in the environment and new information, thereby improving behavior and abilities.

[0632] This invention improves the performance of artificial intelligence through a process of collecting, analyzing, and improving various data within an artificial intelligence-based system. The system mainly consists of a server, terminals, and users, and each element works in coordination.

[0633] The server is responsible for autonomously monitoring the performance of the artificial intelligence and collecting log data. Specifically, it uses software used as a log collection tool to collect data such as the content of the AI's dialogue, response time, and response accuracy, and stores it in a database. The server then analyzes the collected data using data analysis software such as Python or R to identify areas for improvement to enhance performance.

[0634] The terminal provides an interface for users to access the system and visualize the current performance and retraining progress of the artificial intelligence. A dedicated dashboard application allows users to check the system status in real time. Furthermore, the artificial intelligence interacts with the user in natural language via the terminal, accumulating the feedback received. SpaCy and Transformers are used as natural language processing frameworks.

[0635] Users can provide additional data to the artificial intelligence (AI) via the system's dashboard, and use that data to guide the AI's retraining. This allows users to provide new sales materials in response to changes in sales strategies. Users can also send queries to the AI ​​using prompts and receive responses. For example, they can expect to receive new information by using a prompt such as, "Please tell me the latest information on new products."

[0636] In this way, through the cooperation of servers, terminals, and users, artificial intelligence can adapt to environmental changes in real time and achieve efficient task processing. The entire system combines a generative AI model with natural language dialogue using prompt sentences to achieve efficient and flexible artificial intelligence operation.

[0637] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0638] Step 1:

[0639] The server collects log data from AI-generated conversations. It receives conversation text data, response times, and response accuracy as input. The server aggregates this data using a log collection tool and stores it in a database. Specifically, periodic data imports are performed into the database. The output obtained at this stage is structured log data that can be used for subsequent analysis.

[0640] Step 2:

[0641] The server analyzes the collected log data. It uses the structured log data obtained in Step 1 as input. The server uses a data analysis program and applies machine learning algorithms to calculate metrics related to the AI's performance and identify areas for improvement. Specifically, data cleaning and feature extraction are performed, and analysis results are generated. The output is a list of specific areas for improvement to enhance the AI's performance.

[0642] Step 3:

[0643] The server generates feedback to the artificial intelligence based on the analysis results and triggers retraining. The list of improvements from Step 2 is used as input. The server uses reinforcement learning to prepare a retraining protocol to update the AI's parameters based on the input data. Specifically, this involves preparing the retraining dataset and updating the model. The output is the updated artificial intelligence model.

[0644] Step 4:

[0645] Users access the system through a terminal to check current performance and retraining progress. The input is the latest AI status information sent from the server. The terminal visualizes this information via a dedicated dashboard and presents it to the user. Specifically, the dashboard displays graphs and metrics that are updated in real time. The output is visualized information for the user to use.

[0646] Step 5:

[0647] The device facilitates interaction between the user and artificial intelligence using natural language. It receives prompt text from the user as input. The device uses a generative AI model to generate the optimal response based on the prompt. Specifically, it analyzes the user's intent through natural language processing and generates a response. Therefore, the output is a natural language response provided to the user.

[0648] (Application Example 1)

[0649] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0650] Current content delivery systems struggle to quickly respond to users' diverse viewing needs and rapidly changing preferences, often resulting in insufficient generation of personalized recommendations. This leads to problems such as the inability to appropriately recommend content that viewers are interested in, and thus the inability to improve the user experience.

[0651] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0652] In this invention, the server includes means for collecting viewing history and user evaluation information, means for generating personalized recommendations based on the collected information, and means for updating the generated recommendations in real time. This makes it possible to leverage user feedback to adjust personalized recommendations in a timely manner and increase user satisfaction.

[0653] "Viewing history" refers to data such as records, duration, and frequency of content viewing by a user.

[0654] "Rating information" refers to information such as feedback and rating scores that users have given to content.

[0655] "Personalized recommendations" refer to a list of content customized based on each user's viewing history and rating information.

[0656] "Real-time updates" means instantly reflecting users' new viewing history and rating information, and changing recommendations on the spot.

[0657] "Natural language dialogue" is a form of communication in which users interact with computer systems using human language.

[0658] "Feedback" refers to opinions and impressions that users provide about their experience using content or a system.

[0659] A "recommendation algorithm" is a computational method for automatically selecting content that is suitable for the user.

[0660] To implement this invention, consider the following system configuration. First, the server functions as the main data processing unit, collecting user viewing history and evaluation information. Specifically, it uses log data obtained from various content platforms and stores it in a database. The server is equipped with a machine learning framework such as TensorFlow, and serves as a foundation for data analysis and reinforcement learning.

[0661] The server uses this data to periodically update its recommendation algorithm, which suggests the most suitable content for each user. To enable real-time updates, a Python script is used to ensure the algorithm always reflects the latest information. Furthermore, user feedback is obtained through the device in a natural language dialogue format, and this information is also used to optimize the recommendation algorithm. The devices used are smartphones and tablets equipped with user-friendly interfaces that accept natural language input.

[0662] For example, if a user provides feedback such as "I enjoyed it" after watching a movie, the server analyzes it and adds other works of similar genres or by similar directors to the recommendation list. Another example of a prompt message is, "I watched some new content. What should I watch next?" In this way, it is possible to increase user engagement and provide a more personalized viewing experience.

[0663] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0664] Step 1:

[0665] The server receives user viewing history and rating information from the content platform. The received data is stored in the database. At this point, the input is viewing history and rating data, and the output is the stored data. The server performs format conversion and integrity checks on the log data and inserts it into the database in a formatted state.

[0666] Step 2:

[0667] The server runs a recommendation algorithm using stored data to generate personalized content recommendations. The input for this step is user data retrieved from a database, and the output is a personalized recommendation list. The server leverages a machine learning framework to extract significant patterns from the data, shaping the most relevant content for each user into a list.

[0668] Step 3:

[0669] The terminal receives natural language feedback from the user and sends it to the server. The input is the user's natural language feedback, and the output is the parsed data sent to the server. The terminal has speech recognition or text input capabilities to format the user's input feedback without ambiguity.

[0670] Step 4:

[0671] The server analyzes the received feedback and readjusts the recommended algorithm. The input to this process is the feedback data, and the output is the updated algorithm parameters. The server uses the feedback to perform reinforcement learning, learning in real time to improve the parameters of the recommended model.

[0672] Step 5:

[0673] The server sends the updated recommendation list back to the terminal, displaying the new recommendations when the user accesses it. The input is the latest recommendation data, and the output is the display list on the user's terminal. The server quickly transfers the data to the terminal via a communication protocol, creating an environment where users can smoothly check the new recommendations.

[0674] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0675] This invention is a system that combines an emotion engine with an AI agent to achieve more advanced interaction and effective learning. This system mainly consists of a server, terminals, and users.

[0676] 1. Server processing

[0677] The server monitors the performance of the AI ​​agent and collects user emotion data along with performance data. Emotion data is obtained using methods such as voice tone, facial expressions, and text analysis. For example, in a customer support system, the emotional state of a user can be analyzed from their voice tone and logged in real time.

[0678] 2. Integration of the Emotional Engine

[0679] The server incorporates an emotion engine, and the AI ​​agent analyzes the user's emotions through this engine. This data influences the AI ​​agent's response, which is then adjusted to suit the user's emotions. For example, if the user is dissatisfied, the AI ​​agent will provide a more empathetic response.

[0680] 3. Feedback and Retraining

[0681] The server sends feedback, including the user's emotional state, to the learning agent. Based on this feedback, the AI ​​agent retrains and adjusts its learning algorithm. This allows the AI ​​agent to strengthen its emotion-based response strategy and improve the accuracy of its responses.

[0682] 4. User Interaction

[0683] Users interact with an AI agent through their device, and their emotional state is recognized by the system. Users can feel that the AI's responses are appropriate and considerate of their emotions. For example, in educational services, the AI ​​can assess a student's emotional state and provide optimal learning support.

[0684] By incorporating an emotion engine in this way, AI agents can achieve interactions that respond to the user's emotions, providing a more personalized experience. This system provides a foundation for AI agents to perform their tasks while flexibly responding to changes in the environment and the user's emotional state.

[0685] The following describes the processing flow.

[0686] Step 1:

[0687] The server initiates interaction with the user, receiving voice input and text data. This data is immediately sent to the emotion engine for analysis.

[0688] Step 2:

[0689] The emotion engine analyzes the tone of voice, vocabulary, and context of the text to identify the user's emotional state. This analysis is returned to the server as a quantified emotion index.

[0690] Step 3:

[0691] The server adjusts the AI ​​agent's response based on sentiment indicators. For example, if a user expresses dissatisfaction, the AI ​​agent is instructed to apologize and offer quick solutions to resolve the problem.

[0692] Step 4:

[0693] On the device, the AI ​​agent provides the user with tailored responses. This involves using natural speech or text to engage in emotionally conscious dialogue.

[0694] Step 5:

[0695] Through interaction with the AI ​​agent, users receive responses tailored to their needs. Emotionally sensitive responses make users feel more supported.

[0696] Step 6:

[0697] The server sends user feedback and sentiment data to the learning agent, prompting retraining. Based on this data, the learning agent performs reinforcement learning to develop more refined sentiment recognition and response strategies.

[0698] Step 7:

[0699] The device releases an updated AI agent based on the retraining content, preparing for future user interactions. This process allows the AI ​​agent to continuously evolve and adapt more closely to the user's emotions.

[0700] (Example 2)

[0701] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0702] Traditional AI systems have struggled to provide responses that take into account the user's emotional state, resulting in shortcomings in the user experience. Furthermore, the lack of means to analyze emotions in real time and optimize responses based on that analysis meant that user feedback could not be effectively incorporated, limiting the improvement of the accuracy of the learning models.

[0703] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0704] In this invention, the server includes means for acquiring user data based on an information processing device for analyzing emotional states, means for identifying the user's emotional state by analyzing the user data, and means for adjusting the AI ​​system's response according to the emotional state. This enables the AI ​​system to provide interactions that respond to the user's emotions, improving the quality of the user experience and simultaneously improving the accuracy of the learning model that reflects the feedback.

[0705] An "information processing device" is a device that collects and analyzes user data to identify emotional states.

[0706] "User data" refers to information such as voice, facial expressions, and text acquired during interactions with users.

[0707] "Emotional state" refers to the results of analysis to identify the user's emotions, and serves as fundamental information for adjusting the AI ​​system's response.

[0708] "AI system response" refers to the content of the conversation with the AI ​​agent, which is adjusted based on the user's emotional state.

[0709] "Feedback information" refers to information that users provide regarding their evaluations and suggestions for improvement in response to the AI ​​system's actions.

[0710] A "learning model" is a mathematical model used to improve the response accuracy of an AI system based on data analysis.

[0711] The system of this invention aims to adjust the response of an AI agent based on the user's emotions. The system mainly consists of a server, a terminal, and a user, and provides the environment for each function to operate.

[0712] The server houses the information processing unit that forms the core of the AI ​​agent, and it collects and analyzes user data. Various data collection methods are used here, such as speech recognition software and facial expression recognition tools. For example, general speech analysis service software is used for speech recognition, processing changes in speech tone and other data as digital data. For facial expression recognition, facial expression analysis software is used to quantify the user's emotional state in real time and send the analysis data to the server. Based on this analysis data, the AI ​​system's response is dynamically optimized.

[0713] The terminal functions as a means of providing a user interface and acts as a border for information exchange between the user and the AI ​​agent. Specifically, a chat application or voice interface runs on the terminal, allowing the user to interact with the AI ​​agent in natural language. When the user provides voice input to the terminal, the voice is analyzed on a server, and an appropriate response based on the user's emotional state is returned to the terminal. This allows the user to feel that the AI's responses are considerate of their emotions.

[0714] The feedback information that users provide to the AI ​​agent is collected by the server and used to retrain the learning model. This feedback is important data for improving the AI's performance, and the generative AI model is used to improve response accuracy. The generative AI model used at this time is made to generate appropriate responses using prompt statements.

[0715] A concrete example of a prompt might be, "In the following educational scenario, how should the AI ​​agent respond if a student is struggling to understand the learning material?" By using this prompt to obtain a response from the generative AI model, the AI ​​agent can provide more empathetic and effective support to the user.

[0716] In this way, the present invention provides a system that flexibly responds to user emotions, achieving improved user experience and enhanced performance of AI agents.

[0717] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0718] Step 1:

[0719] The server begins collecting user data. When a user uses a terminal to input voice or text, the terminal sends this data to the server. The input data is analyzed by speech recognition software and text analysis tools. Specifically, in the case of voice input, it is digitized and converted into text format through natural language processing. The output of this process is the text data to be analyzed.

[0720] Step 2:

[0721] The server identifies emotional data based on the collected text data. The emotion analysis engine within the server receives this text data and analyzes the user's emotional state by comprehensively considering the results of voice tone and facial recognition. This analysis is performed by an emotion analysis algorithm, and the output is an emotion category (e.g., joy, anger, sadness).

[0722] Step 3:

[0723] The server inputs the obtained emotion categories into a generative AI model to generate a response appropriate for the user. By providing the emotion categories along with prompts to the model, the AI ​​generates an appropriate response that takes emotions into consideration. Specifically, an example prompt such as "What is the appropriate response if the user is angry?" is set, and the generative AI model is used to generate a response to the user. As a result, the AI ​​agent receives a customized response to the user as output.

[0724] Step 4:

[0725] The server forwards the generated response to the terminal. The terminal receives this response and displays it to the user through the user interface. Specifically, if it is in text format, it is displayed in a chat window on the screen, and if it is a voice response, the text is converted into speech using speech synthesis technology. The user interacts with the AI ​​agent through this response.

[0726] Step 5:

[0727] The user provides feedback on the AI ​​agent's responses. The user sends feedback input to the server via their device, indicating satisfaction and areas for improvement. The server accumulates this feedback and incorporates it into its learning algorithm. The AI ​​model's performance is retrained using this data to improve future response accuracy. The output is an improved, trained model.

[0728] (Application Example 2)

[0729] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0730] Conventional AI agents struggle to dynamically adjust their responses based on user emotions, and appropriate dialogue is especially crucial in emotionally sensitive environments such as caregiving settings. However, existing systems are unable to adequately analyze users' emotions and provide adaptive responses instantly, making it difficult to provide personalized services that take emotions into consideration.

[0731] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0732] In this invention, the server includes means for autonomously monitoring the performance of other AI agents, means for analyzing the user's emotions, and means for the AI ​​agent to adjust its response based on the emotion analysis. This makes it possible to instantly analyze the user's emotions in settings such as nursing care and provide a response that is in harmony with them, thereby providing a more personalized experience.

[0733] "Autonomously monitoring the performance of other AI agents" means that an AI agent independently observes the activities of other AI agents and continuously evaluates their performance.

[0734] "Identifying necessary improvements" means analyzing data on the AI ​​agent's performance to determine the specific elements and areas that need improvement.

[0735] "Retraining" means updating the AI ​​agent's learning model based on identified areas for improvement and running the learning process again to enhance its performance.

[0736] "Emotional analysis," which analyzes the emotions of users, is a process that collects emotional data from the user's facial expressions, tone of voice, text, etc., and clarifies their emotional state.

[0737] "Adjusting responses" means modifying the output of the AI ​​agent based on the results of sentiment analysis to provide users with more emotionally sensitive responses.

[0738] "Natural language dialogue" refers to the communication process between a user and an AI agent using the language that humans use in everyday life.

[0739] A "reinforcement learning algorithm" is a learning process that improves the efficiency of an AI agent by selecting actions based on rewards from the environment and repeatedly refining those selections.

[0740] "Providing feedback" means sending back to the system the user's reactions and interaction results obtained by the AI ​​agent as useful information for performance improvement.

[0741] This invention aims to enhance individualized care by introducing AI agents into small-group living environments such as caregiving settings. The specific elements for implementing this invention are described below.

[0742] The server autonomously monitors the performance of other AI agents and runs software for emotion analysis. For example, it uses the Google Cloud Speech-to-Text API to analyze tone from speech and the Microsoft Face API for facial expression analysis. The server collects this data and evaluates the user's emotional state in real time.

[0743] The terminals used are smartphones and caregiving robots, which acquire audio and video data through sensor devices such as microphones and cameras. The acquired data is sent to a server, where an AI agent generates an appropriate response based on the analysis results. The generated response is provided to the user in natural language through a dialogue module. The AI ​​agent receives feedback within a certain range periodically and improves its learning algorithm with new data.

[0744] Users interact with the AI ​​agent through everyday conversations, evaluating the emotional adaptability of its responses. For example, if the AI ​​agent determines that the user is feeling down, it may suggest prompts such as, "Did anything fun happen today? Shall I put on some relaxing music?" to facilitate a more personalized experience.

[0745] In this way, the entire system adjusts according to the user's emotional changes and provides more adaptive help, aiming to improve efficiency and quality in care settings.

[0746] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0747] Step 1:

[0748] The device uses a microphone and camera to acquire the user's voice and video data in real time. Voice data includes tone, volume, and speaking speed, while video data captures the user's facial expressions and movements. This data is then transferred directly to the server.

[0749] Step 2:

[0750] The server converts the received audio data into text using the Google Cloud Speech-to-Text API and then performs emotional analysis on the voice. It analyzes variations in volume and speaking speed from the audio to profile the emotional state. As a result, it obtains a numerical evaluation of the user's emotions (e.g., excitement, joy, sadness, anger).

[0751] Step 3:

[0752] The server analyzes the video data and uses facial expression analysis tools such as the Microsoft Face API to determine the user's emotions from their facial expressions. Specifically, it analyzes facial feature points (eyes, eyebrows, mouth angle, etc.) and maps changes in facial expressions in the video to emotion categories (e.g., smile, confusion, surprise).

[0753] Step 4:

[0754] By combining this sentiment analysis data, the server uses an AI model to determine the user's current emotional state. Based on this determination, it generates a response from the AI ​​agent and creates a prompt. For example, if it determines that the user is feeling anxious, it prepares a response that will provide reassurance.

[0755] Step 5:

[0756] The AI ​​agent sends the generated prompt message back to the terminal, which then presents it to the user as voice or text. For example, it can engage in a more human-like conversation, such as, "Did anything fun happen today? Shall I put on some relaxing music?"

[0757] Step 6:

[0758] After receiving a response from the AI ​​agent, the user inputs that feedback as a separate dialogue. The device sends this feedback to the server as new input data. The server uses the received feedback as training data for the AI ​​agent and retrains it to improve future responses.

[0759] This entire process enables the entire system to continuously provide a flexible and adaptive experience that responds to the user's changing emotions.

[0760] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0761] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0762] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0763] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0764] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0765] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0766] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0767] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0768] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0769] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0770] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0771] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0772] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0773] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[0774] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0775] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[0776] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0777] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0778] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0779] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0780] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0781] The following is further disclosed regarding the embodiments described above.

[0782] (Claim 1)

[0783] A means of autonomously monitoring the performance of other AI agents,

[0784] A means for identifying necessary improvements based on the aforementioned performance,

[0785] Based on the identified areas for improvement, a means to retrain other AI agents,

[0786] A means of providing feedback through natural language dialogue,

[0787] A means for learning from the aforementioned feedback and improving the performance of the learning agent,

[0788] A system that includes this.

[0789] (Claim 2)

[0790] The system according to claim 1, wherein the other AI agent adapts to changes in the environment using a reinforcement learning algorithm.

[0791] (Claim 3)

[0792] The system according to claim 1, which automatically updates the content of the retraining in response to new data or changes in the environment.

[0793] "Example 1"

[0794] (Claim 1)

[0795] A means of autonomously monitoring other artificial intelligences,

[0796] A means for collecting and analyzing performance data of the aforementioned artificial intelligence,

[0797] A means for identifying areas for improvement based on the aforementioned analysis results,

[0798] Based on the aforementioned improvements, a means for retraining artificial intelligence,

[0799] A means of exchanging information and providing feedback using natural language,

[0800] A means to improve the performance of artificial intelligence based on the aforementioned feedback,

[0801] A system that includes this.

[0802] (Claim 2)

[0803] The system according to claim 1, wherein the artificial intelligence adapts to external changes using a reinforcement learning method.

[0804] (Claim 3)

[0805] The system according to claim 1, which automatically updates the content of the retraining in response to new information or external changes.

[0806] "Application Example 1"

[0807] (Claim 1)

[0808] Means for collecting viewing history and user rating information,

[0809] A means for generating personalized recommendations based on collected information,

[0810] A means of updating the generated recommendations in real time,

[0811] A means of obtaining user feedback through natural language dialogue,

[0812] A means for learning from the aforementioned feedback and optimizing the recommended algorithm,

[0813] A system that includes this.

[0814] (Claim 2)

[0815] The system according to claim 1, wherein the recommendations are adapted to changes in user behavior using a reinforcement learning algorithm.

[0816] (Claim 3)

[0817] The system according to claim 1, which automatically updates the generation of recommendations in response to new viewing data and user feedback.

[0818] "Example 2 of combining an emotion engine"

[0819] (Claim 1)

[0820] A means of acquiring user data based on an information processing device for analyzing emotional states,

[0821] A means for analyzing the user data to identify the user's emotional state,

[0822] Means for adjusting the AI ​​system's response in accordance with the aforementioned emotional state,

[0823] Means for providing the adjusted response to the user,

[0824] A means of collecting user feedback information and updating the learning model,

[0825] A system that includes this.

[0826] (Claim 2)

[0827] The system according to claim 1, which uses a learning algorithm to optimize a response strategy based on emotional state.

[0828] (Claim 3)

[0829] The system according to claim 1, which automatically updates the content for retraining a learning model based on newly collected data.

[0830] "Application example 2 when combining with an emotional engine"

[0831] (Claim 1)

[0832] A means of autonomously monitoring the performance of other AI agents,

[0833] A means for identifying necessary improvements based on the aforementioned performance,

[0834] Based on the identified areas for improvement, a means to retrain other AI agents,

[0835] A means of analyzing the emotions of users,

[0836] A means by which an AI agent adjusts its response based on emotion analysis,

[0837] A means of providing feedback through natural language dialogue,

[0838] A means for learning from the aforementioned feedback and improving the performance of the learning agent,

[0839] A system that includes this.

[0840] (Claim 2)

[0841] The system according to claim 1, wherein the other AI agent adapts to changes in the environment and also to changes in the user's emotions using a reinforcement learning algorithm.

[0842] (Claim 3)

[0843] The system according to claim 1, which automatically updates the content of the retraining in response to new data, environmental changes, and the user's emotions. [Explanation of Symbols]

[0844] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of autonomously monitoring the performance of other AI agents, A means for identifying necessary improvements based on the aforementioned performance, Based on the identified areas for improvement, a means to retrain other AI agents, A means of providing feedback through natural language dialogue, A means for learning from the aforementioned feedback and improving the performance of the learning agent, A system that includes this.

2. The system according to claim 1, wherein the other AI agent adapts to changes in the environment using a reinforcement learning algorithm.

3. The system according to claim 1, which automatically updates the content of the retraining in response to new data or changes in the environment.

Images