system

Abstract

The system according to this embodiment aims to efficiently collect feedback linked to the user's login activity and reflect it in the retraining of the generated AI. [Solution] The system according to the embodiment comprises a collection unit, an analysis unit, a reflection unit, and a retraining unit. The collection unit collects feedback linked to the user's login activity. The analysis unit analyzes the data collected by the collection unit. The reflection unit reflects the feedback to the generating AI based on the data analyzed by the analysis unit. The retraining unit causes the generating AI to retrain based on the feedback reflected by the reflection unit.

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] In the prior art, there were problems that human feedback was slow in the learning of generative AI and it was difficult to solve the problem of biased learning data.

[0005] The system according to the embodiment aims to efficiently collect feedback in association with the user's login behavior and reflect it in the relearning of generative AI.

Means for Solving the Problems

[0006] The system according to this embodiment comprises a collection unit, an analysis unit, a reflection unit, and a retraining unit. The collection unit collects feedback linked to the user's login activity. The analysis unit analyzes the data collected by the collection unit. The reflection unit reflects the feedback based on the data analyzed by the analysis unit into the generating AI. The retraining unit causes the generating AI to retrain based on the feedback reflected by the reflection unit. [Effects of the Invention]

[0007] The system according to this embodiment can efficiently collect feedback linked to the user's login activity and reflect it in the retraining of the generated AI. [Brief explanation of the drawing]

[0008] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Modes for carrying out the invention]

[0009] Hereinafter, an example of an embodiment of the system relating to the technology of this disclosure will be described with reference to the attached drawings.

[0010] First, let's explain the terminology used in the following explanation.

[0011] In the following embodiments, the signed processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Furthermore, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), or TPU (Tensor Processing Unit).

[0012] In the following embodiments, signed RAM (Random Access Memory) is a memory that temporarily stores information and is used as work memory by the processor.

[0013] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0014] In the following embodiments, the labeled communication I / F (Interface) is an interface including a communication processor, an antenna, and the like. The communication I / F manages communication between a plurality of computers. Examples of communication standards applicable to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0015] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B". That is, "A and / or B" means that it may be only A, only B, or a combination of A and B. Also, in this specification, when expressing three or more matters connected by "and / or", the same concept as "A and / or B" is applied.

[0016] [First Embodiment] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0017] As shown in FIG. 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.

[0018] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. Also, the database 24 and the communication I / F 26 are connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0019] The smart device 14 comprises a computer 36, a receiving 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 receiving device 38, output device 40, and camera 42 are also connected to the bus 52.

[0020] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, and accepts user input. The touch panel 38A accepts user input via touch by detecting contact with an object (e.g., a pen or finger). The microphone 38B accepts user input via voice 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 unit 12. In the data processing unit 12, the specific processing unit 290 (see Figure 2) acquires the data indicating the user input.

[0021] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user by outputting the data in a form perceptible to the user (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.

[0022] 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.

[0023] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0024] 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.

[0025] Storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0026] In the smart device 14, specific processing is performed by the processor 46. The storage 50 stores a specific processing program 60. The specific processing program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the specific processing program 60 from the storage 50 and executes the read specific processing program 60 on the RAM 48. The specific processing is realized by the processor 46 operating as a control unit 46A according to the specific processing program 60 executed on the RAM 48. The smart device 14 also has a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0027] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device having the data generation model 58. The data processing device 12 may also be a server device or a terminal device owned by a user (e.g., a mobile phone, robot, home appliance, etc.). Next, an example of processing by the data processing system 10 according to the first embodiment will be described.

[0028] (Example of form 1) The system according to an embodiment of the present invention is a mechanism to solve the problem of AI retraining and regenerating large amounts of biased training data as generative AI evolves. This system proposes an image feedback agent, "AiCAPTCHA Agent," that uses CAPTCHA and is linked to the user's daily login activity. CAPTCHA is an image authentication system required to distinguish between humans and bots, along with the "I'm not a robot" screen. First, when a user logs in, image authentication is performed through CAPTCHA. At this time, the user is asked to select the correct items for a specific image. For example, the user is instructed to select all the correct items as a description of a photograph of a woman in a kimono. This information is input into the generative AI. Next, the generative AI analyzes the input information and adopts it as feedback according to the user's recognition rate. For example, if a certain number of responses or more are obtained, the result is adopted as feedback. This mechanism allows for the collection of a large amount of feedback in a short period of time, keeping pace with the AI's learning speed. Furthermore, the generative AI retrains based on the feedback results. This reduces the AI's bias problem and improves accuracy. For example, by providing accurate feedback to AI-generated images of "Japanese families," users can enable the AI ​​to produce more accurate images of Japanese families. This mechanism reduces outsourcing costs, improves the consistency of feedback, mitigates biased feedback, and effectively solves the AI ​​bias problem. In this way, the system effectively solves the bias problem of generated AI.

[0029] The system according to this embodiment comprises a collection unit, an analysis unit, a reflection unit, and a retraining unit. The collection unit collects feedback linked to the user's login activity. For example, the collection unit performs image authentication via CAPTCHA when the user logs in. The analysis unit analyzes the data collected by the collection unit. For example, the analysis unit prompts the user to select the correct item for a specific image. The reflection unit reflects the feedback to the generating AI based on the data analyzed by the analysis unit. For example, the reflection unit adopts the feedback according to the user's recognition ratio. The retraining unit causes the generating AI to retrain based on the feedback reflected by the reflection unit. For example, the retraining unit causes the generating AI to retrain based on the feedback results. As a result, the system can effectively solve the AI ​​bias problem by collecting, analyzing, reflecting, and retraining feedback based on the user's login activity.

[0030] The data collection unit collects feedback linked to user login actions. Specifically, when a user logs in, image authentication is performed via CAPTCHA. CAPTCHA is a test to confirm that the user is a human and not a robot, and generally uses a format where the user selects a specific object from a set of images. For example, a user might follow instructions on the login screen such as "Select all traffic lights" and select an image containing traffic lights from multiple images. At this time, the data collection unit collects information such as the user's selected image data, selection time, and accuracy of the selection. This data is important for understanding the user's recognition ability and behavioral patterns and is used for subsequent analysis and retraining. The data collection unit collects this data in real time and stores it in a secure database. Furthermore, the data collection unit anonymizes and encrypts the data to protect user privacy and ensure data security. This allows the data collection unit to efficiently collect high-quality feedback data based on user login actions.

[0031] The analysis unit analyzes the data collected by the collection unit. Specifically, it asks the user to select the correct item for a given image. The analysis unit uses AI to analyze the collected data in real time and evaluate the accuracy of the user's selections and reaction time. For example, when a user selects a traffic light, the analysis unit analyzes whether the correct image was selected, the time taken for the selection, and the selection pattern. The AI ​​utilizes image recognition and pattern recognition technologies to analyze the user's selection behavior in detail and understand the user's recognition ability and behavioral characteristics. Furthermore, the analysis unit can also detect abnormal behavior and patterns by comparing them with past data and data from other users, and assess security risks. For example, if a particular user selects images in an unusual pattern, it can determine that the user may be attempting unauthorized access and issue a warning. In this way, the analysis unit can quickly and accurately analyze the collected data and understand the user's behavioral characteristics and security risks in real time.

[0032] The feedback unit incorporates feedback into the generating AI based on the data analyzed by the analysis unit. Specifically, it adopts feedback according to the user's recognition rate. Based on the analysis results, the feedback unit provides appropriate feedback to the generating AI and incorporates it as training data for the generating AI. For example, if a user correctly selects a traffic light, that data is provided to the generating AI as positive feedback; conversely, if an incorrect selection is made, it is provided as negative feedback. The feedback unit incorporates this feedback into the generating AI's learning algorithm and adjusts it so that the generating AI has more accurate recognition capabilities. Furthermore, the feedback unit can optimize the learning efficiency of the generating AI by adjusting the content and frequency of the feedback. For example, by providing more detailed feedback to a specific user group and simplified feedback to other user groups, the learning balance of the generating AI can be maintained. In this way, the feedback unit can improve the performance of the generating AI by incorporating appropriate feedback into the generating AI based on the analysis results.

[0033] The retraining unit allows the generating AI to retrain based on the feedback reflected by the reflection unit. Specifically, the generating AI retrains based on the results of the feedback. The retraining unit updates the learning model using the newly provided feedback data, improving recognition ability and accuracy. For example, the generating AI learns from data where the user correctly selected a traffic light, improving recognition accuracy for the next time. It can also learn from data where the user made an incorrect selection, thereby identifying and correcting the cause of misrecognition. The retraining unit monitors the generating AI's learning process and performs retraining at the appropriate time. Furthermore, the retraining unit evaluates the generating AI's learning results and adjusts the learning algorithm and parameters as needed to provide an optimal learning environment. As a result, the retraining unit allows the generating AI to continuously retrain based on the feedback provided by the reflection unit, improving recognition ability and accuracy. This enables the entire system to effectively solve AI bias problems and provide a safer and more reliable system by collecting, analyzing, reflecting, and retraining feedback based on user login behavior.

[0034] The data collection unit can perform image authentication via CAPTCHA when a user logs in. For example, the data collection unit performs image authentication via CAPTCHA when a user logs in. This makes it possible to collect feedback by performing image authentication via CAPTCHA when a user logs in. CAPTCHA includes, but is not limited to, image authentication, text authentication, and voice authentication. Some or all of the above processing in the data collection unit may be performed using AI, for example, or without AI. For example, the data collection unit can input CAPTCHA image authentication data into a generating AI and have the generating AI execute the results of the image authentication.

[0035] The analysis unit can prompt the user to select the correct item for a specific image. For example, the analysis unit may prompt the user to select the correct item for a specific image. This improves the accuracy of the feedback by allowing the user to select the correct item for a specific image. Specific images include, but are not limited to, landscape images, object images, and portrait images. Some or all of the processing described above in the analysis unit may be performed using AI, for example, or without AI. For example, the analysis unit can input the image data selected by the user into a generating AI and have the generating AI perform the analysis of the image data.

[0036] The feedback unit can be used as feedback according to the user's recognition ratio. For example, the feedback unit can be used as feedback according to the user's recognition ratio. By using feedback according to the user's recognition ratio, the reliability of the feedback is improved. The recognition ratio includes, but is not limited to, the correct answer rate, incorrect answer rate, and confidence score. Some or all of the processing in the feedback unit described above may be performed using AI, for example, or without AI. For example, the feedback unit can input the user's recognition ratio data into a generating AI and have the generating AI perform an analysis of the recognition ratio.

[0037] The retraining unit allows the generating AI to retrain based on the feedback results. For example, the retraining unit allows the generating AI to retrain based on the feedback results. This improves the accuracy of the AI. Retraining includes, but is not limited to, the dataset used, the frequency of retraining, and the type of algorithm. Some or all of the above processing in the retraining unit may be performed using AI or not using AI. For example, the retraining unit can input feedback data into the generating AI and have the generating AI perform the retraining.

[0038] The data collection unit can collect feedback linked to the user's daily login activities. For example, the data collection unit collects feedback linked to the user's daily login activities. This improves the efficiency of feedback collection. Daily login activities include, but are not limited to, daily logins or logins during specific time periods. Some or all of the above processing in the data collection unit may be performed using AI, for example, or without AI. For example, the data collection unit can input user login data into a generating AI and have the generating AI perform an analysis of login activities.

[0039] The data collection unit can analyze a user's past login history and select the optimal data collection method. For example, the data collection unit can analyze the times when a user frequently logs in and request feedback during those times. For example, if a user tends to log in on a particular day of the week, the data collection unit can request feedback on that day. For example, the data collection unit can adjust the frequency of feedback according to the user's login frequency. This allows the optimal data collection method to be selected by analyzing a user's past login history. Past login history includes, but is not limited to, login date and time, number of logins, and login success rate. Some or all of the above processing in the data collection unit may be performed using, for example, AI, or not using AI. For example, the data collection unit can input user login history data into a generating AI and have the generating AI perform the analysis of the login history.

[0040] The data collection unit can filter feedback based on the user's current activities and areas of interest. For example, if the user is working, the data collection unit may request work-related images as feedback. If the user is spending time on a hobby, the data collection unit may request images related to that hobby as feedback. If the user is traveling, the data collection unit may request travel-related images as feedback. By filtering based on the user's activities and areas of interest, the data collection unit can collect highly relevant feedback. Current activities include, but are not limited to, active sessions and applications being used. Areas of interest include, but are not limited to, past search history, browsing history, and survey results. Some or all of the processing described above in the data collection unit may be performed using, for example, AI, or not using AI. For example, the data collection unit may input user activity data into a generating AI and have the generating AI perform an analysis of the activity data.

[0041] The data collection unit can prioritize collecting highly relevant feedback by considering the user's geographical location information during feedback collection. For example, if the user is in a specific region, the data collection unit may request images related to that region as feedback. For example, if the user is traveling, the data collection unit may request images related to their travel destination as feedback. For example, if the user is at home, the data collection unit may request images related to their home area as feedback. This allows for the priority collection of highly relevant feedback by considering the user's geographical location information. Geographical location information includes, but is not limited to, GPS data, IP addresses, and location services. Some or all of the processing described above in the data collection unit may be performed using, for example, AI, or without AI. For example, the data collection unit may input the user's geographical location data into a generating AI and have the generating AI perform location information analysis.

[0042] The data collection unit can analyze a user's social media activity and collect relevant feedback when collecting feedback. For example, the data collection unit may request feedback related to images shared by the user on social media. For example, the data collection unit may request feedback related to accounts followed by the user on social media. For example, the data collection unit may request feedback related to images liked by the user on social media. In this way, relevant feedback can be collected by analyzing the user's social media activity. Social media activity includes, but is not limited to, posts, the number of likes, and the number of followers. Some or all of the processing described above in the data collection unit may be performed using, for example, AI, or not using AI. For example, the data collection unit can input the user's social media data into a generating AI and have the generating AI perform the analysis of social media activity.

[0043] The analysis unit can adjust the level of detail of the analysis based on the importance of the feedback during the analysis. For example, the analysis unit performs a detailed analysis for high-importance feedback. For example, the analysis unit performs a concise analysis for low-importance feedback. For example, the analysis unit performs an analysis with an appropriate level of detail for medium-importance feedback. By adjusting the level of detail of the analysis based on the importance of the feedback, efficient analysis becomes possible. Importance includes, but is not limited to, user evaluation, content of the feedback, and impact. Some or all of the above processing in the analysis unit may be performed using, for example, AI, or without AI. For example, the analysis unit can input feedback data into a generating AI and have the generating AI perform an importance analysis.

[0044] The analysis unit can apply different analysis algorithms depending on the category of the feedback during analysis. For example, the analysis unit applies an image analysis algorithm to feedback related to image recognition. For example, the analysis unit applies a natural language processing algorithm to feedback related to text recognition. For example, the analysis unit applies a speech analysis algorithm to feedback related to speech recognition. By applying different analysis algorithms depending on the category of feedback, the accuracy of the analysis is improved. Categories include, but are not limited to, the content, theme, and topic of the feedback. Some or all of the above processing in the analysis unit may be performed using AI, for example, or without AI. For example, the analysis unit can input feedback data into a generating AI and have the generating AI perform analysis for each category.

[0045] The analysis unit can determine the priority of analysis based on the timing of feedback submission during the analysis process. For example, the analysis unit may prioritize the analysis of recently submitted feedback. For example, the analysis unit may postpone the analysis of older feedback. For example, the analysis unit may analyze feedback submitted at a moderate pace. This allows for efficient analysis by prioritizing the analysis based on the timing of feedback submission. The submission timing includes, but is not limited to, the submission date and time, submission frequency, and submission timing. Some or all of the above processing in the analysis unit may be performed using, for example, AI, or not using AI. For example, the analysis unit may input feedback data into a generating AI and have the generating AI perform the analysis of submission timing.

[0046] The analysis unit can adjust the order of analysis based on the relevance of the feedback during the analysis. For example, the analysis unit may prioritize analyzing feedback with high relevance. For example, it may postpone analyzing feedback with low relevance. For example, it may analyze feedback with moderate relevance to a reasonable extent. This allows for efficient analysis by adjusting the order of analysis based on the relevance of the feedback. Relevance includes, but is not limited to, the content, theme, and topic of the feedback. Some or all of the above processing in the analysis unit may be performed using, for example, AI, or without AI. For example, the analysis unit can input feedback data into a generating AI and have the generating AI perform the relevance analysis.

[0047] The reflection unit can adjust the level of detail in the reflection process based on the importance of the feedback. For example, the reflection unit will perform detailed reflection for high-importance feedback. For example, the reflection unit will perform concise reflection for low-importance feedback. For example, the reflection unit will perform reflection with an appropriate level of detail for medium-importance feedback. This allows for efficient reflection by adjusting the level of detail in the reflection process based on the importance of the feedback. Importance includes, but is not limited to, user ratings, feedback content, and impact. Some or all of the above processing in the reflection unit may be performed using, for example, AI, or without AI. For example, the reflection unit can input feedback data into a generating AI and have the generating AI perform an analysis of importance.

[0048] The reflection unit can apply different reflection algorithms depending on the category of the feedback during the reflection process. For example, the reflection unit applies an image reflection algorithm to feedback related to image recognition. For example, the reflection unit applies a natural language processing reflection algorithm to feedback related to text recognition. For example, the reflection unit applies a speech reflection algorithm to feedback related to speech recognition. By applying different reflection algorithms depending on the category of feedback, the accuracy of the reflection is improved. Categories include, but are not limited to, the content, theme, and topic of the feedback. Some or all of the processing described above in the reflection unit may be performed using AI, for example, or without AI. For example, the reflection unit can input feedback data into a generating AI and have the generating AI perform reflection for each category.

[0049] The feedback processing unit can adjust the order of feedback processing based on the submission date. For example, the feedback processing unit may prioritize recently submitted feedback. For example, it may postpone older feedback. For example, it may appropriately process feedback submitted at a moderate time. This allows for efficient feedback processing by adjusting the order of processing based on the submission date. The submission date includes, but is not limited to, the submission date and time, submission frequency, and submission timing. Some or all of the processing described above in the feedback processing unit may be performed using, for example, AI, or not using AI. For example, the feedback processing unit may input feedback data into a generating AI and have the generating AI perform an analysis of the submission date.

[0050] The reflection unit can adjust the order of reflection based on the relevance of the feedback. For example, the reflection unit prioritizes reflecting highly relevant feedback. For example, the reflection unit postpones reflecting less relevant feedback. For example, the reflection unit reflects moderately relevant feedback appropriately. This allows for efficient reflection by adjusting the order of reflection based on the relevance of the feedback. Relevance includes, but is not limited to, the content, theme, and topic of the feedback. Some or all of the above processing in the reflection unit may be performed using, for example, AI, or without AI. For example, the reflection unit can input feedback data into a generating AI and have the generating AI perform relevance analysis.

[0051] The retraining unit can optimize the retraining algorithm by referring to past training data during retraining. For example, the retraining unit can analyze past training data and select the optimal retraining algorithm. For example, the retraining unit can adjust the retraining algorithm based on trends in past training data. For example, the retraining unit can correct errors in past training data and optimize the retraining algorithm. This improves the accuracy of retraining by optimizing the retraining algorithm by referring to past training data. Past training data includes, but is not limited to, past feedback data, training history, and evaluation results. Some or all of the above processes in the retraining unit may be performed using, for example, AI, or not using AI. For example, the retraining unit can input past training data into a generating AI and have the generating AI perform the optimization of the retraining algorithm.

[0052] The retraining unit can apply different retraining methods to each category of feedback during retraining. For example, the retraining unit can apply an image retraining method to feedback related to image recognition. For example, the retraining unit can apply a natural language processing retraining method to feedback related to text recognition. For example, the retraining unit can apply a speech retraining method to feedback related to speech recognition. By applying different retraining methods to each category of feedback, the accuracy of retraining is improved. Categories include, but are not limited to, the content, theme, and topic of the feedback. Some or all of the processing described above in the retraining unit may be performed using, for example, AI, or without AI. For example, the retraining unit can input feedback data into a generating AI and have the generating AI perform category-specific retraining.

[0053] The retraining unit can weight the retraining data based on the timing of feedback submissions during retraining. For example, the retraining unit may weight the retraining data by giving more weight to recently submitted feedback. For example, the retraining unit may lighten the weight of older feedback. For example, the retraining unit may give a moderate weight to feedback that was submitted at a moderate time. This enables efficient retraining by weighting the retraining data based on the timing of feedback submissions. The submission timing includes, but is not limited to, submission date and time, submission frequency, and submission timing. Some or all of the above processing in the retraining unit may be performed using, for example, AI, or not using AI. For example, the retraining unit may input feedback data into a generating AI and have the generating AI perform an analysis of the submission timing.

[0054] The retraining unit can weight the retraining data based on the relevance of the feedback during retraining. For example, the retraining unit may weight the retraining data by giving more weight to feedback with high relevance, or lighter weight to feedback with low relevance, or moderate weight to feedback with moderate relevance. This enables efficient retraining by weighting the retraining data based on the relevance of the feedback. Relevance includes, but is not limited to, the content, theme, and topic of the feedback. Some or all of the above processing in the retraining unit may be performed using AI, for example, or without AI. For example, the retraining unit may input feedback data into a generating AI and have the generating AI perform relevance analysis.

[0055] The system according to the embodiment is not limited to the example described above, and various modifications are possible, for example, as follows.

[0056] The data collection unit can collect not only user login activity but also user browsing history. For example, if a user frequently visits a particular website, it can request images related to that website as feedback. Similarly, if a user searches for a specific keyword, it can request images related to that keyword as feedback. Furthermore, if a user tends to visit a particular website at a specific time, it can request feedback during that time. This allows for the collection of more relevant feedback by basing it on the user's browsing history.

[0057] The feedback unit can adjust the output of the generating AI in real time based on user feedback. For example, if a user provides correct feedback on a particular image, that feedback can be immediately reflected in the generating AI. Conversely, if a user provides incorrect feedback, that feedback can be ignored. Furthermore, the output of the generating AI can be sequentially corrected based on user feedback. This allows for the provision of more accurate output results by adjusting the output of the generating AI in real time based on user feedback.

[0058] The data collection unit can analyze users' social media activity and collect relevant feedback. For example, it can request feedback related to images users have shared on social media. It can also request feedback related to accounts users follow on social media. Furthermore, it can request feedback related to images users have "liked" on social media. In this way, relevant feedback can be collected by analyzing users' social media activity.

[0059] The feedback unit can customize the output of the generating AI based on user feedback. For example, if a user provides correct feedback on a particular image, the output of the generating AI can be customized based on that feedback. Conversely, if a user provides incorrect feedback, that feedback can be ignored. Furthermore, the output of the generating AI can be sequentially corrected based on user feedback. This allows for more accurate output results by customizing the output of the generating AI based on user feedback.

[0060] The data collection unit can prioritize collecting highly relevant feedback by considering the user's geographical location. For example, if a user is in a specific region, it can request images related to that region as feedback. Similarly, if a user is traveling, it can request images related to their travel destination. Furthermore, if a user is at home, it can request images related to their home environment. This allows for the priority collection of highly relevant feedback by considering the user's geographical location.

[0061] The analysis unit can adjust the level of detail of the analysis based on the importance of the feedback during the analysis. For example, it can perform a detailed analysis for high-importance feedback, a concise analysis for low-importance feedback, and an analysis with an appropriate level of detail for moderately important feedback. By adjusting the level of detail of the analysis based on the importance of the feedback, efficient analysis becomes possible.

[0062] The following briefly describes the processing flow for example form 1.

[0063] Step 1: The data collection unit collects feedback linked to the user's login activity. For example, when a user logs in, image verification is performed via CAPTCHA. Step 2: The analysis unit analyzes the data collected by the collection unit. For example, it may prompt the user to select the correct item for a specific image. Step 3: The reflection unit reflects the feedback generated by the analysis unit into the AI. For example, it may adopt feedback based on the user's recognition rate. Step 4: The retraining unit retrains the generating AI based on the feedback reflected by the reflection unit. For example, the generating AI retrains based on the results of the feedback.

[0064] (Example of form 2) The system according to an embodiment of the present invention is a mechanism to solve the problem of AI retraining and regenerating large amounts of biased training data as generative AI evolves. This system proposes an image feedback agent, "AiCAPTCHA Agent," that uses CAPTCHA and is linked to the user's daily login activity. CAPTCHA is an image authentication system required to distinguish between humans and bots, along with the "I'm not a robot" screen. First, when a user logs in, image authentication is performed through CAPTCHA. At this time, the user is asked to select the correct items for a specific image. For example, the user is instructed to select all the correct items as a description of a photograph of a woman in a kimono. This information is input into the generative AI. Next, the generative AI analyzes the input information and adopts it as feedback according to the user's recognition rate. For example, if a certain number of responses or more are obtained, the result is adopted as feedback. This mechanism allows for the collection of a large amount of feedback in a short period of time, keeping pace with the AI's learning speed. Furthermore, the generative AI retrains based on the feedback results. This reduces the AI's bias problem and improves accuracy. For example, by providing accurate feedback to AI-generated images of "Japanese families," users can enable the AI ​​to produce more accurate images of Japanese families. This mechanism reduces outsourcing costs, improves the consistency of feedback, mitigates biased feedback, and effectively solves the AI ​​bias problem. In this way, the system effectively solves the bias problem of generated AI.

[0065] The system according to this embodiment comprises a collection unit, an analysis unit, a reflection unit, and a retraining unit. The collection unit collects feedback linked to the user's login activity. For example, the collection unit performs image authentication via CAPTCHA when the user logs in. The analysis unit analyzes the data collected by the collection unit. For example, the analysis unit prompts the user to select the correct item for a specific image. The reflection unit reflects the feedback to the generating AI based on the data analyzed by the analysis unit. For example, the reflection unit adopts the feedback according to the user's recognition ratio. The retraining unit causes the generating AI to retrain based on the feedback reflected by the reflection unit. For example, the retraining unit causes the generating AI to retrain based on the feedback results. As a result, the system can effectively solve the AI ​​bias problem by collecting, analyzing, reflecting, and retraining feedback based on the user's login activity.

[0066] The data collection unit collects feedback linked to user login actions. Specifically, when a user logs in, image authentication is performed via CAPTCHA. CAPTCHA is a test to confirm that the user is a human and not a robot, and generally uses a format where the user selects a specific object from a set of images. For example, a user might follow instructions on the login screen such as "Select all traffic lights" and select an image containing traffic lights from multiple images. At this time, the data collection unit collects information such as the user's selected image data, selection time, and accuracy of the selection. This data is important for understanding the user's recognition ability and behavioral patterns and is used for subsequent analysis and retraining. The data collection unit collects this data in real time and stores it in a secure database. Furthermore, the data collection unit anonymizes and encrypts the data to protect user privacy and ensure data security. This allows the data collection unit to efficiently collect high-quality feedback data based on user login actions.

[0067] The analysis unit analyzes the data collected by the collection unit. Specifically, it asks the user to select the correct item for a given image. The analysis unit uses AI to analyze the collected data in real time and evaluate the accuracy of the user's selections and reaction time. For example, when a user selects a traffic light, the analysis unit analyzes whether the correct image was selected, the time taken for the selection, and the selection pattern. The AI ​​utilizes image recognition and pattern recognition technologies to analyze the user's selection behavior in detail and understand the user's recognition ability and behavioral characteristics. Furthermore, the analysis unit can also detect abnormal behavior and patterns by comparing them with past data and data from other users, and assess security risks. For example, if a particular user selects images in an unusual pattern, it can determine that the user may be attempting unauthorized access and issue a warning. In this way, the analysis unit can quickly and accurately analyze the collected data and understand the user's behavioral characteristics and security risks in real time.

[0068] The feedback unit incorporates feedback into the generating AI based on the data analyzed by the analysis unit. Specifically, it adopts feedback according to the user's recognition rate. Based on the analysis results, the feedback unit provides appropriate feedback to the generating AI and incorporates it as training data for the generating AI. For example, if a user correctly selects a traffic light, that data is provided to the generating AI as positive feedback; conversely, if an incorrect selection is made, it is provided as negative feedback. The feedback unit incorporates this feedback into the generating AI's learning algorithm and adjusts it so that the generating AI has more accurate recognition capabilities. Furthermore, the feedback unit can optimize the learning efficiency of the generating AI by adjusting the content and frequency of the feedback. For example, by providing more detailed feedback to a specific user group and simplified feedback to other user groups, the learning balance of the generating AI can be maintained. In this way, the feedback unit can improve the performance of the generating AI by incorporating appropriate feedback into the generating AI based on the analysis results.

[0069] The retraining unit allows the generating AI to retrain based on the feedback reflected by the reflection unit. Specifically, the generating AI retrains based on the results of the feedback. The retraining unit updates the learning model using the newly provided feedback data, improving recognition ability and accuracy. For example, the generating AI learns from data where the user correctly selected a traffic light, improving recognition accuracy for the next time. It can also learn from data where the user made an incorrect selection, thereby identifying and correcting the cause of misrecognition. The retraining unit monitors the generating AI's learning process and performs retraining at the appropriate time. Furthermore, the retraining unit evaluates the generating AI's learning results and adjusts the learning algorithm and parameters as needed to provide an optimal learning environment. As a result, the retraining unit allows the generating AI to continuously retrain based on the feedback provided by the reflection unit, improving recognition ability and accuracy. This enables the entire system to effectively solve AI bias problems and provide a safer and more reliable system by collecting, analyzing, reflecting, and retraining feedback based on user login behavior.

[0070] The data collection unit can perform image authentication via CAPTCHA when a user logs in. For example, the data collection unit performs image authentication via CAPTCHA when a user logs in. This makes it possible to collect feedback by performing image authentication via CAPTCHA when a user logs in. CAPTCHA includes, but is not limited to, image authentication, text authentication, and voice authentication. Some or all of the above processing in the data collection unit may be performed using AI, for example, or without AI. For example, the data collection unit can input CAPTCHA image authentication data into a generating AI and have the generating AI execute the results of the image authentication.

[0071] The analysis unit can prompt the user to select the correct item for a specific image. For example, the analysis unit may prompt the user to select the correct item for a specific image. This improves the accuracy of the feedback by allowing the user to select the correct item for a specific image. Specific images include, but are not limited to, landscape images, object images, and portrait images. Some or all of the processing described above in the analysis unit may be performed using AI, for example, or without AI. For example, the analysis unit can input the image data selected by the user into a generating AI and have the generating AI perform the analysis of the image data.

[0072] The feedback unit can be used as feedback according to the user's recognition ratio. For example, the feedback unit can be used as feedback according to the user's recognition ratio. By using feedback according to the user's recognition ratio, the reliability of the feedback is improved. The recognition ratio includes, but is not limited to, the correct answer rate, incorrect answer rate, and confidence score. Some or all of the processing in the feedback unit described above may be performed using AI, for example, or without AI. For example, the feedback unit can input the user's recognition ratio data into a generating AI and have the generating AI perform an analysis of the recognition ratio.

[0073] The retraining unit allows the generating AI to retrain based on the feedback results. For example, the retraining unit allows the generating AI to retrain based on the feedback results. This improves the accuracy of the AI. Retraining includes, but is not limited to, the dataset used, the frequency of retraining, and the type of algorithm. Some or all of the above processing in the retraining unit may be performed using AI or not using AI. For example, the retraining unit can input feedback data into the generating AI and have the generating AI perform the retraining.

[0074] The data collection unit can collect feedback linked to the user's daily login activities. For example, the data collection unit collects feedback linked to the user's daily login activities. This improves the efficiency of feedback collection. Daily login activities include, but are not limited to, daily logins or logins during specific time periods. Some or all of the above processing in the data collection unit may be performed using AI, for example, or without AI. For example, the data collection unit can input user login data into a generating AI and have the generating AI perform an analysis of login activities.

[0075] The data collection unit can estimate the user's emotions and adjust the timing of feedback collection based on the estimated emotions. For example, if the user is stressed, the data collection unit will request feedback when the user is relaxed, rather than immediately after logging in. For example, if the user is focused, the data collection unit will request feedback immediately after logging in. For example, if the user is in a hurry, the data collection unit will request feedback after a certain period of time has elapsed since logging in. By adjusting the timing of feedback collection based on the user's emotions, feedback can be collected at a more appropriate time. Emotion estimation is achieved using an emotion estimation function, such as an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the data collection unit may be performed using AI, or not using AI. For example, the data collection unit can input user emotion data into a generative AI and have the generative AI perform emotion estimation.

[0076] The data collection unit can analyze a user's past login history and select the optimal data collection method. For example, the data collection unit can analyze the times when a user frequently logs in and request feedback during those times. For example, if a user tends to log in on a particular day of the week, the data collection unit can request feedback on that day. For example, the data collection unit can adjust the frequency of feedback according to the user's login frequency. This allows the optimal data collection method to be selected by analyzing a user's past login history. Past login history includes, but is not limited to, login date and time, number of logins, and login success rate. Some or all of the above processing in the data collection unit may be performed using, for example, AI, or not using AI. For example, the data collection unit can input user login history data into a generating AI and have the generating AI perform the analysis of the login history.

[0077] The data collection unit can filter feedback based on the user's current activities and areas of interest. For example, if the user is working, the data collection unit may request work-related images as feedback. If the user is spending time on a hobby, the data collection unit may request images related to that hobby as feedback. If the user is traveling, the data collection unit may request travel-related images as feedback. By filtering based on the user's activities and areas of interest, the data collection unit can collect highly relevant feedback. Current activities include, but are not limited to, active sessions and applications being used. Areas of interest include, but are not limited to, past search history, browsing history, and survey results. Some or all of the processing described above in the data collection unit may be performed using, for example, AI, or not using AI. For example, the data collection unit may input user activity data into a generating AI and have the generating AI perform an analysis of the activity data.

[0078] The data collection unit can estimate the user's emotions and determine the priority of feedback to collect based on the estimated emotions. For example, if the user is relaxed, the data collection unit will prioritize collecting high-importance feedback. If the user is stressed, the data collection unit will prioritize collecting simple feedback. If the user is focused, the data collection unit will prioritize collecting detailed feedback. This allows for the priority collection of important feedback by prioritizing feedback based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, such as an emotion engine or generative AI. Generative AI may be, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the data collection unit may be performed using AI or not. For example, the data collection unit can input user emotion data into a generative AI and have the generative AI perform emotion estimation.

[0079] The data collection unit can prioritize collecting highly relevant feedback by considering the user's geographical location information during feedback collection. For example, if the user is in a specific region, the data collection unit may request images related to that region as feedback. For example, if the user is traveling, the data collection unit may request images related to their travel destination as feedback. For example, if the user is at home, the data collection unit may request images related to their home area as feedback. This allows for the priority collection of highly relevant feedback by considering the user's geographical location information. Geographical location information includes, but is not limited to, GPS data, IP addresses, and location services. Some or all of the processing described above in the data collection unit may be performed using, for example, AI, or without AI. For example, the data collection unit may input the user's geographical location data into a generating AI and have the generating AI perform location information analysis.

[0080] The data collection unit can analyze a user's social media activity and collect relevant feedback when collecting feedback. For example, the data collection unit may request feedback related to images shared by the user on social media. For example, the data collection unit may request feedback related to accounts followed by the user on social media. For example, the data collection unit may request feedback related to images liked by the user on social media. In this way, relevant feedback can be collected by analyzing the user's social media activity. Social media activity includes, but is not limited to, posts, the number of likes, and the number of followers. Some or all of the processing described above in the data collection unit may be performed using, for example, AI, or not using AI. For example, the data collection unit can input the user's social media data into a generating AI and have the generating AI perform the analysis of social media activity.

[0081] The analysis unit can estimate the user's emotions and adjust the presentation of the analysis based on the estimated emotions. For example, if the user is relaxed, the analysis unit provides detailed analysis results. For example, if the user is stressed, the analysis unit provides concise analysis results. For example, if the user is focused, the analysis unit provides visually easy-to-understand analysis results. By adjusting the presentation of the analysis based on the user's emotions, more appropriate analysis results can be provided. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the analysis unit may be performed using AI, for example, or without AI. For example, the analysis unit can input user emotion data into a generative AI and have the generative AI perform emotion estimation.

[0082] The analysis unit can adjust the level of detail of the analysis based on the importance of the feedback during the analysis. For example, the analysis unit performs a detailed analysis for high-importance feedback. For example, the analysis unit performs a concise analysis for low-importance feedback. For example, the analysis unit performs an analysis with an appropriate level of detail for medium-importance feedback. By adjusting the level of detail of the analysis based on the importance of the feedback, efficient analysis becomes possible. Importance includes, but is not limited to, user evaluation, content of the feedback, and impact. Some or all of the above processing in the analysis unit may be performed using, for example, AI, or without AI. For example, the analysis unit can input feedback data into a generating AI and have the generating AI perform an importance analysis.

[0083] The analysis unit can apply different analysis algorithms depending on the category of the feedback during analysis. For example, the analysis unit applies an image analysis algorithm to feedback related to image recognition. For example, the analysis unit applies a natural language processing algorithm to feedback related to text recognition. For example, the analysis unit applies a speech analysis algorithm to feedback related to speech recognition. By applying different analysis algorithms depending on the category of feedback, the accuracy of the analysis is improved. Categories include, but are not limited to, the content, theme, and topic of the feedback. Some or all of the above processing in the analysis unit may be performed using AI, for example, or without AI. For example, the analysis unit can input feedback data into a generating AI and have the generating AI perform analysis for each category.

[0084] The analysis unit can estimate the user's emotions and adjust the length of the analysis based on the estimated emotions. For example, if the user is in a hurry, the analysis unit will provide a short analysis result. For example, if the user is relaxed, the analysis unit will provide a detailed analysis result. For example, if the user is focused, the analysis unit will provide an analysis result of appropriate length. By adjusting the length of the analysis based on the user's emotions, more appropriate analysis results can be provided. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. Some or all of the above processing in the analysis unit may be performed using AI, for example, or without AI. For example, the analysis unit can input user emotion data into a generative AI and have the generative AI perform emotion estimation.

[0085] The analysis unit can determine the priority of analysis based on the timing of feedback submission during the analysis process. For example, the analysis unit may prioritize the analysis of recently submitted feedback. For example, the analysis unit may postpone the analysis of older feedback. For example, the analysis unit may analyze feedback submitted at a moderate pace. This allows for efficient analysis by prioritizing the analysis based on the timing of feedback submission. The submission timing includes, but is not limited to, the submission date and time, submission frequency, and submission timing. Some or all of the above processing in the analysis unit may be performed using, for example, AI, or not using AI. For example, the analysis unit may input feedback data into a generating AI and have the generating AI perform the analysis of submission timing.

[0086] The analysis unit can adjust the order of analysis based on the relevance of the feedback during the analysis. For example, the analysis unit may prioritize analyzing feedback with high relevance. For example, it may postpone analyzing feedback with low relevance. For example, it may analyze feedback with moderate relevance to a reasonable extent. This allows for efficient analysis by adjusting the order of analysis based on the relevance of the feedback. Relevance includes, but is not limited to, the content, theme, and topic of the feedback. Some or all of the above processing in the analysis unit may be performed using, for example, AI, or without AI. For example, the analysis unit can input feedback data into a generating AI and have the generating AI perform the relevance analysis.

[0087] The reflection unit can estimate the user's emotions and adjust the reflection method based on the estimated user emotions. For example, if the user is relaxed, the reflection unit will reflect detailed feedback. For example, if the user is stressed, the reflection unit will reflect concise feedback. For example, if the user is focused, the reflection unit will reflect visually easy-to-understand feedback. This allows for more appropriate feedback to be reflected by adjusting the reflection method based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the reflection unit may be performed using AI, for example, or without AI. For example, the reflection unit can input user emotion data into a generative AI and have the generative AI perform emotion estimation.

[0088] The reflection unit can adjust the level of detail in the reflection process based on the importance of the feedback. For example, the reflection unit will perform detailed reflection for high-importance feedback. For example, the reflection unit will perform concise reflection for low-importance feedback. For example, the reflection unit will perform reflection with an appropriate level of detail for medium-importance feedback. This allows for efficient reflection by adjusting the level of detail in the reflection process based on the importance of the feedback. Importance includes, but is not limited to, user ratings, feedback content, and impact. Some or all of the above processing in the reflection unit may be performed using, for example, AI, or without AI. For example, the reflection unit can input feedback data into a generating AI and have the generating AI perform an analysis of importance.

[0089] The reflection unit can apply different reflection algorithms depending on the category of the feedback during the reflection process. For example, the reflection unit applies an image reflection algorithm to feedback related to image recognition. For example, the reflection unit applies a natural language processing reflection algorithm to feedback related to text recognition. For example, the reflection unit applies a speech reflection algorithm to feedback related to speech recognition. By applying different reflection algorithms depending on the category of feedback, the accuracy of the reflection is improved. Categories include, but are not limited to, the content, theme, and topic of the feedback. Some or all of the processing described above in the reflection unit may be performed using AI, for example, or without AI. For example, the reflection unit can input feedback data into a generating AI and have the generating AI perform reflection for each category.

[0090] The feedback unit can estimate the user's emotions and determine the priority of feedback based on the estimated emotions. For example, if the user is relaxed, the feedback unit will prioritize reflecting high-importance feedback. For example, if the user is stressed, the feedback unit will prioritize reflecting simple feedback. For example, if the user is focused, the feedback unit will prioritize reflecting detailed feedback. In this way, important feedback can be prioritized by determining the priority of feedback based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the feedback unit may be performed using AI, for example, or not using AI. For example, the feedback unit can input user emotion data into a generative AI and have the generative AI perform emotion estimation.

[0091] The feedback processing unit can adjust the order of feedback processing based on the submission date. For example, the feedback processing unit may prioritize recently submitted feedback. For example, it may postpone older feedback. For example, it may appropriately process feedback submitted at a moderate time. This allows for efficient feedback processing by adjusting the order of processing based on the submission date. The submission date includes, but is not limited to, the submission date and time, submission frequency, and submission timing. Some or all of the processing described above in the feedback processing unit may be performed using, for example, AI, or not using AI. For example, the feedback processing unit may input feedback data into a generating AI and have the generating AI perform an analysis of the submission date.

[0092] The reflection unit can adjust the order of reflection based on the relevance of the feedback. For example, the reflection unit prioritizes reflecting highly relevant feedback. For example, the reflection unit postpones reflecting less relevant feedback. For example, the reflection unit reflects moderately relevant feedback appropriately. This allows for efficient reflection by adjusting the order of reflection based on the relevance of the feedback. Relevance includes, but is not limited to, the content, theme, and topic of the feedback. Some or all of the above processing in the reflection unit may be performed using, for example, AI, or without AI. For example, the reflection unit can input feedback data into a generating AI and have the generating AI perform relevance analysis.

[0093] The retraining unit can estimate the user's emotions and select retraining data based on the estimated user emotions. For example, if the user is relaxed, the retraining unit will select detailed retraining data. For example, if the user is stressed, the retraining unit will select concise retraining data. For example, if the user is focused, the retraining unit will select visually easy-to-understand retraining data. This allows for more appropriate retraining by selecting retraining data based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the retraining unit may be performed using AI, for example, or without AI. For example, the retraining unit can input user emotion data into a generative AI and have the generative AI perform emotion estimation.

[0094] The retraining unit can optimize the retraining algorithm by referring to past training data during retraining. For example, the retraining unit can analyze past training data and select the optimal retraining algorithm. For example, the retraining unit can adjust the retraining algorithm based on trends in past training data. For example, the retraining unit can correct errors in past training data and optimize the retraining algorithm. This improves the accuracy of retraining by optimizing the retraining algorithm by referring to past training data. Past training data includes, but is not limited to, past feedback data, training history, and evaluation results. Some or all of the above processes in the retraining unit may be performed using, for example, AI, or not using AI. For example, the retraining unit can input past training data into a generating AI and have the generating AI perform the optimization of the retraining algorithm.

[0095] The retraining unit can apply different retraining methods to each category of feedback during retraining. For example, the retraining unit can apply an image retraining method to feedback related to image recognition. For example, the retraining unit can apply a natural language processing retraining method to feedback related to text recognition. For example, the retraining unit can apply a speech retraining method to feedback related to speech recognition. By applying different retraining methods to each category of feedback, the accuracy of retraining is improved. Categories include, but are not limited to, the content, theme, and topic of the feedback. Some or all of the processing described above in the retraining unit may be performed using, for example, AI, or without AI. For example, the retraining unit can input feedback data into a generating AI and have the generating AI perform category-specific retraining.

[0096] The relearning unit can estimate the user's emotions and adjust the frequency of relearning based on the estimated emotions. For example, the relearning unit increases the frequency of relearning when the user is relaxed. For example, the relearning unit decreases the frequency of relearning when the user is stressed. For example, the relearning unit performs relearning at an appropriate frequency when the user is focused. By adjusting the frequency of relearning based on the user's emotions, more appropriate relearning becomes possible. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. Some or all of the above processing in the relearning unit may be performed using AI, for example, or without AI. For example, the relearning unit can input user emotion data into a generative AI and have the generative AI perform emotion estimation.

[0097] The retraining unit can weight the retraining data based on the timing of feedback submissions during retraining. For example, the retraining unit may weight the retraining data by giving more weight to recently submitted feedback. For example, the retraining unit may lighten the weight of older feedback. For example, the retraining unit may give a moderate weight to feedback that was submitted at a moderate time. This enables efficient retraining by weighting the retraining data based on the timing of feedback submissions. The submission timing includes, but is not limited to, submission date and time, submission frequency, and submission timing. Some or all of the above processing in the retraining unit may be performed using, for example, AI, or not using AI. For example, the retraining unit may input feedback data into a generating AI and have the generating AI perform an analysis of the submission timing.

[0098] The retraining unit can weight the retraining data based on the relevance of the feedback during retraining. For example, the retraining unit may weight the retraining data by giving more weight to feedback with high relevance, or lighter weight to feedback with low relevance, or moderate weight to feedback with moderate relevance. This enables efficient retraining by weighting the retraining data based on the relevance of the feedback. Relevance includes, but is not limited to, the content, theme, and topic of the feedback. Some or all of the above processing in the retraining unit may be performed using AI, for example, or without AI. For example, the retraining unit may input feedback data into a generating AI and have the generating AI perform relevance analysis.

[0099] The system according to the embodiment is not limited to the example described above, and various modifications are possible, for example, as follows.

[0100] The data collection unit can collect not only user login activity but also user browsing history. For example, if a user frequently visits a particular website, it can request images related to that website as feedback. Similarly, if a user searches for a specific keyword, it can request images related to that keyword as feedback. Furthermore, if a user tends to visit a particular website at a specific time, it can request feedback during that time. This allows for the collection of more relevant feedback by basing it on the user's browsing history.

[0101] The analysis unit can estimate the user's emotions and adjust the accuracy of the analysis based on those emotions. For example, if the user is relaxed, a detailed analysis can be performed. If the user is stressed, a concise analysis can be performed. Furthermore, if the user is focused, a visually easy-to-understand analysis can be performed. By adjusting the accuracy of the analysis based on the user's emotions, more appropriate analysis results can be provided.

[0102] The feedback unit can adjust the output of the generating AI in real time based on user feedback. For example, if a user provides correct feedback on a particular image, that feedback can be immediately reflected in the generating AI. Conversely, if a user provides incorrect feedback, that feedback can be ignored. Furthermore, the output of the generating AI can be sequentially corrected based on user feedback. This allows for the provision of more accurate output results by adjusting the output of the generating AI in real time based on user feedback.

[0103] The relearning unit can estimate the user's emotions and adjust the timing of relearning based on those emotions. For example, if the user is relaxed, the frequency of relearning can be increased. Conversely, if the user is stressed, the frequency of relearning can be decreased. Furthermore, if the user is focused, relearning can be performed at an appropriate frequency. By adjusting the timing of relearning based on the user's emotions, more appropriate relearning becomes possible.

[0104] The data collection unit can analyze users' social media activity and collect relevant feedback. For example, it can request feedback related to images users have shared on social media. It can also request feedback related to accounts users follow on social media. Furthermore, it can request feedback related to images users have "liked" on social media. In this way, relevant feedback can be collected by analyzing users' social media activity.

[0105] The analysis unit can estimate the user's emotions and adjust the presentation of the analysis based on those emotions. For example, if the user is relaxed, it can provide detailed analysis results. If the user is stressed, it can provide concise analysis results. Furthermore, if the user is focused, it can provide visually easy-to-understand analysis results. In this way, by adjusting the presentation of the analysis based on the user's emotions, more appropriate analysis results can be provided.

[0106] The feedback unit can customize the output of the generating AI based on user feedback. For example, if a user provides correct feedback on a particular image, the output of the generating AI can be customized based on that feedback. Conversely, if a user provides incorrect feedback, that feedback can be ignored. Furthermore, the output of the generating AI can be sequentially corrected based on user feedback. This allows for more accurate output results by customizing the output of the generating AI based on user feedback.

[0107] The relearning unit can estimate the user's emotions and select relearning data based on those emotions. For example, if the user is relaxed, detailed relearning data can be selected. If the user is stressed, concise relearning data can be selected. Furthermore, if the user is focused, visually easy-to-understand relearning data can be selected. This allows for more appropriate relearning by selecting relearning data based on the user's emotions.

[0108] The data collection unit can prioritize collecting highly relevant feedback by considering the user's geographical location. For example, if a user is in a specific region, it can request images related to that region as feedback. Similarly, if a user is traveling, it can request images related to their travel destination. Furthermore, if a user is at home, it can request images related to their home environment. This allows for the priority collection of highly relevant feedback by considering the user's geographical location.

[0109] The analysis unit can adjust the level of detail of the analysis based on the importance of the feedback during the analysis. For example, it can perform a detailed analysis for high-importance feedback, a concise analysis for low-importance feedback, and an analysis with an appropriate level of detail for moderately important feedback. By adjusting the level of detail of the analysis based on the importance of the feedback, efficient analysis becomes possible.

[0110] The following briefly describes the processing flow for example form 2.

[0111] Step 1: The data collection unit collects feedback linked to the user's login activity. For example, when a user logs in, image verification is performed via CAPTCHA. Step 2: The analysis unit analyzes the data collected by the collection unit. For example, it may prompt the user to select the correct item for a specific image. Step 3: The reflection unit reflects the feedback generated by the analysis unit into the AI. For example, it may adopt feedback based on the user's recognition rate. Step 4: The retraining unit retrains the generating AI based on the feedback reflected by the reflection unit. For example, the generating AI retrains based on the results of the feedback.

[0112] 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.

[0113] Data generation model 58 is a form of so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> Examples of generative AI include text generation AI, image generation AI, and multimodal generation AI. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats from audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVMs), k-means clustering, convolutional neural networks (CNNs), recurrent neural networks (RNNs), generative adversarial networks (GANs), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each of the above parts is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example.Furthermore, processing performed by AI, including generative AI, may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by AI, including generative AI.

[0114] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart device 14 or an external device, and the smart device 14 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0115] Each of the multiple elements described above, including the collection unit, analysis unit, reflection unit, and retraining unit, is implemented in at least one of the smart device 14 and the data processing device 12. For example, the collection unit is implemented by the control unit 46A of the smart device 14 and collects feedback linked to the user's login activity. The analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the collected data. The reflection unit is implemented by the specific processing unit 290 of the data processing device 12 and reflects the feedback to the generating AI based on the analyzed data. The retraining unit is implemented by the specific processing unit 290 of the data processing device 12 and the generating AI retrains based on the reflected feedback. The correspondence between each unit and the device or control unit is not limited to the example described above and can be changed in various ways.

[0116] [Second Embodiment] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0117] 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.

[0118] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0119] 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.

[0120] The microphone 238 receives voice signals from the user and accepts instructions from the user. The microphone 238 captures the voice signals from the user, 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.

[0121] 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, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0122] 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.

[0123] 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 by the processor 28. The storage 32 stores the specific processing program 56.

[0124] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0125] Storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0126] In the smart glasses 214, specific processing is performed by the processor 46. The storage 50 stores a specific processing program 60. The processor 46 reads the specific processing program 60 from the storage 50 and executes the read specific processing program 60 on the RAM 48. The specific processing is realized by the processor 46 acting as a control unit 46A according to the specific processing program 60 executed on the RAM 48. The smart glasses 214 also have a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0127] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0128] 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.

[0129] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0130] The data processing system 210 according to the second embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 210 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but it may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart glasses 214 or an external device, and the smart glasses 214 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0131] Each of the multiple elements described above, including the collection unit, analysis unit, reflection unit, and retraining unit, is implemented in at least one of the smart glasses 214 and the data processing device 12. For example, the collection unit is implemented by the control unit 46A of the smart glasses 214 and collects feedback linked to the user's login activity. The analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the collected data. The reflection unit is implemented by the specific processing unit 290 of the data processing device 12 and reflects the feedback to the generating AI based on the analyzed data. The retraining unit is implemented by the specific processing unit 290 of the data processing device 12 and the generating AI retrains based on the reflected feedback. The correspondence between each unit and the device or control unit is not limited to the example described above and can be changed in various ways.

[0132] [Third Embodiment] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0133] 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.

[0134] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0135] 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.

[0136] The microphone 238 receives voice signals from the user and accepts instructions from the user. The microphone 238 captures the voice signals from the user, 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.

[0137] 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, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0138] 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.

[0139] 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.

[0140] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0141] Storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0142] In the headset terminal 314, specific processing is performed by the processor 46. The storage 50 stores a specific program 60. The processor 46 reads the specific program 60 from the storage 50 and executes the read specific program 60 on the RAM 48. The specific processing is realized by the processor 46 acting as a control unit 46A according to the specific program 60 executed on the RAM 48. The headset terminal 314 also has a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0143] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0144] 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.

[0145] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0146] The data processing system 310 according to the third embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 310 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the headset terminal 314, but may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the headset terminal 314. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the headset terminal 314 or an external device, and the headset terminal 314 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0147] Each of the multiple elements described above, including the collection unit, analysis unit, reflection unit, and retraining unit, is implemented in at least one of the headset terminal 314 and the data processing device 12. For example, the collection unit is implemented by the control unit 46A of the headset terminal 314 and collects feedback linked to the user's login activity. The analysis unit is implemented by the specific processing unit 290 of the data processing device 12 and analyzes the collected data. The reflection unit is implemented by the specific processing unit 290 of the data processing device 12 and reflects the feedback to the generating AI based on the analyzed data. The retraining unit is implemented by the specific processing unit 290 of the data processing device 12 and the generating AI retrains based on the reflected feedback. The correspondence between each unit and the device or control unit is not limited to the example described above and can be changed in various ways.

[0148] [Fourth Embodiment] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0149] 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.

[0150] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0151] 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.

[0152] The microphone 238 receives voice signals from the user and accepts instructions from the user. The microphone 238 captures the voice signals from the user, 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.

[0153] 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 image sensor or CCD image sensor, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0154] 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.

[0155] 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. The robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0156] 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.

[0157] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0158] Storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0159] In robot 414, specific processing is performed by processor 46. A specific program 60 is stored in storage 50. Processor 46 reads the specific program 60 from storage 50 and executes it on RAM 48. The specific processing is achieved by processor 46 acting as a control unit 46A according to the specific program 60 executed on RAM 48. Robot 414 also has data generation model 58 and emotion identification model 59, similar to those of the robot, and can perform processing similar to that of the specific processing unit 290 using these models.

[0160] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0161] 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.

[0162] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0163] The data processing system 410 according to the fourth embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 410 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but it may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the robot 414 or an external device, and the robot 414 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0164] Each of the multiple elements described above, including the collection unit, analysis unit, reflection unit, and relearning unit, is implemented in at least one of the robot 414 and the data processing unit 12. For example, the collection unit is implemented by the control unit 46A of the robot 414 and collects feedback linked to the user's login activity. The analysis unit is implemented by the specific processing unit 290 of the data processing unit 12 and analyzes the collected data. The reflection unit is implemented by the specific processing unit 290 of the data processing unit 12 and reflects the feedback to the generating AI based on the analyzed data. The relearning unit is implemented by the specific processing unit 290 of the data processing unit 12 and the generating AI retrains based on the reflected feedback. The correspondence between each unit and the device or control unit is not limited to the example described above and can be changed in various ways.

[0165] 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.

[0166] Figure 9 shows the 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.

[0167] 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.

[0168] 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.

[0169] 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, and motorcycles, 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 based, for example, 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.

[0170] 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."

[0171] 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.

[0172] 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 method for the specific process may be used, which includes computer 22 and multiple other computers.

[0173] 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.

[0174] 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.

[0175] 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.

[0176] 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.

[0177] 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.

[0178] 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.

[0179] 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.

[0180] Furthermore, although the above-described examples were divided into four embodiments, some or all of these embodiments may be combined. Also, the smart device 14, smart glasses 214, headset terminal 314, and robot 414 are just examples, and they may be combined, or other devices may be used. Also, although the above-described examples were divided into two embodiments, Embodiment 1 and Embodiment 2, these may be combined.

[0181] 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 other things 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.

[0182] 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.

[0183] (Note 1) A collection unit that collects feedback linked to user login actions, An analysis unit analyzes the data collected by the aforementioned collection unit, A reflection unit that reflects feedback to the generating AI based on the data analyzed by the aforementioned analysis unit, The system comprises a retraining unit that performs retraining on the generating AI based on the feedback reflected by the reflection unit. A system characterized by the following features. (Note 2) The aforementioned collection unit is When a user logs in, image verification is performed via CAPTCHA. The system described in Appendix 1, characterized by the features described herein. (Note 3) The aforementioned analysis unit, The user is prompted to select the correct item for a specific image. The system described in Appendix 1, characterized by the features described herein. (Note 4) The aforementioned reflection unit is, The feedback will be adopted based on the user recognition rate. The system described in Appendix 1, characterized by the features described herein. (Note 5) The aforementioned relearning unit, The generative AI retrains based on the feedback results. The system described in Appendix 1, characterized by the features described herein. (Note 6) The aforementioned collection unit is We collect feedback linked to users' daily login activities. The system described in Appendix 1, characterized by the features described herein. (Note 7) The aforementioned collection unit is It estimates the user's emotions and adjusts the timing of feedback collection based on the estimated user emotions. The system described in Appendix 1, characterized by the features described herein. (Note 8) The aforementioned collection unit is Analyze the user's past login history and select the optimal data collection method. The system described in Appendix 1, characterized by the features described herein. (Note 9) The aforementioned collection unit is When collecting feedback, filter it based on the user's current activities and areas of interest. The system described in Appendix 1, characterized by the features described herein. (Note 10) The aforementioned collection unit is It estimates the user's emotions and determines the priority of feedback to collect based on those estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 11) The aforementioned collection unit is When collecting feedback, the system prioritizes collecting highly relevant feedback by considering the user's geographical location. The system described in Appendix 1, characterized by the features described herein. (Note 12) The aforementioned collection unit is When collecting feedback, analyze users' social media activity and gather relevant feedback. The system described in Appendix 1, characterized by the features described herein. (Note 13) The aforementioned analysis unit, The system estimates the user's emotions and adjusts the representation of the analysis based on the estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 14) The aforementioned analysis unit, During analysis, adjust the level of detail based on the importance of the feedback. The system described in Appendix 1, characterized by the features described herein. (Note 15) The aforementioned analysis unit, During analysis, different analysis algorithms are applied depending on the feedback category. The system described in Appendix 1, characterized by the features described herein. (Note 16) The aforementioned analysis unit, It estimates the user's emotions and adjusts the length of the analysis based on the estimated user emotions. The system described in Appendix 1, characterized by the features described herein. (Note 17) The aforementioned analysis unit, During the analysis, the analysis priority is determined based on when feedback is submitted. The system described in Appendix 1, characterized by the features described herein. (Note 18) The aforementioned analysis unit, During analysis, the order of analysis is adjusted based on the relevance of the feedback. The system described in Appendix 1, characterized by the features described herein. (Note 19) The aforementioned reflection unit is, We estimate the user's emotions and adjust how those emotions are reflected based on that estimation. The system described in Appendix 1, characterized by the features described herein. (Note 20) The aforementioned reflection unit is, When implementing changes, adjust the level of detail based on the importance of the feedback. The system described in Appendix 1, characterized by the features described herein. (Note 21) The aforementioned reflection unit is, When the feedback is reflected, a different reflection algorithm is applied depending on the feedback category. The system described in Appendix 1, characterized by the features described herein. (Note 22) The aforementioned reflection unit is, The system estimates the user's emotions and determines the priority of what to reflect based on those estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 23) The aforementioned reflection unit is, When implementing changes, the order in which changes are implemented will be adjusted based on when the feedback was submitted. The system described in Appendix 1, characterized by the features described herein. (Note 24) The aforementioned reflection unit is, When implementing changes, the order of implementation will be adjusted based on the relevance of the feedback. The system described in Appendix 1, characterized by the features described herein. (Note 25) The aforementioned relearning unit, The system estimates the user's emotions and selects retraining data based on the estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 26) The aforementioned relearning unit, During retraining, the retraining algorithm is optimized by referring to past training data. The system described in Appendix 1, characterized by the features described herein. (Note 27) The aforementioned relearning unit, During retraining, different retraining methods are applied for each feedback category. The system described in Appendix 1, characterized by the features described herein. (Note 28) The aforementioned relearning unit, It estimates the user's emotions and adjusts the frequency of retraining based on the estimated user emotions. The system described in Appendix 1, characterized by the features described herein. (Note 29) The aforementioned relearning unit, During retraining, the retraining data is weighted based on when feedback was submitted. The system described in Appendix 1, characterized by the features described herein. (Note 30) The aforementioned relearning unit, During retraining, the retraining data is weighted based on the relevance of the feedback. The system described in Appendix 1, characterized by the features described herein. [Explanation of symbols]

[0184] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots

Claims

1. A collection unit that collects feedback linked to user login actions, An analysis unit analyzes the data collected by the aforementioned collection unit, A reflection unit that reflects feedback to the generating AI based on the data analyzed by the analysis unit, The system includes a relearning unit that performs relearning on the generating AI based on the feedback reflected by the reflection unit. A system characterized by the following features.

2. The aforementioned collection unit is When a user logs in, image verification is performed via CAPTCHA. The system according to feature 1.

3. The aforementioned analysis unit, The user is prompted to select the correct item for a specific image. The system according to feature 1.

4. The aforementioned reflection unit is, The feedback will be adopted based on the user recognition rate. The system according to feature 1.

5. The aforementioned relearning unit, The generative AI retrains based on the feedback results. The system according to feature 1.

6. The aforementioned collection unit is We collect feedback linked to users' daily login activities. The system according to feature 1.

7. The aforementioned collection unit is It estimates the user's emotions and adjusts the timing of feedback collection based on the estimated user emotions. The system according to feature 1.

8. The aforementioned collection unit is Analyze the user's past login history and select the optimal data collection method. The system according to feature 1.

Images