System

Abstract

A system is provided.SOLUTION: A system, comprising: means for collecting and processing historical audio and video of an in-house review meeting; means for training a generative AI model based on the processed audio and video; means for simulating review of material submitted by a user using the generative AI model; and means for providing the simulation results as feedback to the user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] In many companies, it is difficult to coordinate schedules with senior executives when holding internal review meetings, making it difficult to receive reviews frequently. Furthermore, some employees become nervous when interacting with senior executives and are unable to receive effective feedback. Furthermore, while there is a need to improve the quality of reviews and practice them repeatedly, there is a limit to the number of actual review meetings that can be held. It is necessary to solve these problems and provide an environment where employees can receive effective reviews at any time. [Means for solving the problem]

[0005] The present invention provides a system that includes a means for collecting and processing audio and video data from past internal review meetings, a means for training a generative AI model based on the processed audio and video data, a means for simulating a review of materials submitted by a user using the generative AI model, and a means for providing the results of the simulation as feedback to the user. This allows users to receive simulated reviews with a quality similar to that of a real review without having to coordinate schedules with senior executives. Furthermore, the generative AI model reproduces tone and emotional patterns, allowing employees to relax and receive effective feedback. Furthermore, the system allows employees to conduct multiple reviews with a virtual director, helping them prepare for a real director review.

[0006] An "internal review meeting" is a meeting held within a company that includes evaluation and feedback of documents such as proposals and project plans.

[0007] "Past data" refers to audio and video data from internal review meetings that have already been held.

[0008] "Audio data" refers to data that records the content of conversations or words in the form of sound waves.

[0009] "Video data" refers to data that visually records the state of a conference and the actions of participants.

[0010] "Processing means" refers to a method or device for analyzing, transforming, tagging, and storing audio and video data.

[0011] A "generative AI model" is an artificial intelligence model that is trained using machine learning technology to reproduce the tone and emotional patterns of a specific person in a specific position (e.g., a general manager).

[0012] "Training means" refers to the method or process by which a generative AI model is trained using collected data.

[0013] "Means for simulating" refers to a method or apparatus for virtually reviewing, evaluating, and providing feedback on user-submitted materials using a generative AI model.

[0014] "Feedback" means the information, advice, and evaluation provided as a result of a Review.

[0015] The "System" is a comprehensive mechanism that includes a set of means for processing audio and video data from internal review meetings, reviewing user materials using generative AI models, and providing feedback. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0017] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0018] First, the terms used in the following description will be explained.

[0019] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0020] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

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

[0022] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0023] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0024] [First embodiment]

[0025] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0026] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0027] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0028] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0029] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0030] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0031] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0032] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0034] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0035] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0036] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0037] MODE FOR CARRYING OUT THE INVENTION

[0038] The present invention provides a system that allows users to virtually perform review simulations using a generative AI model by using past data from internal review meetings. The following describes an embodiment of this system.

[0039] Data Collection Phase

[0040] The server collects past data (audio and video data) from internal review meetings. This data can be acquired using audio and video recording devices such as microphones and cameras. It is also possible to export data from existing conference systems. The collected audio data is stored in the server's storage in sound wave format, and the video data is stored in video format.

[0041] Data analysis and processing

[0042] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it also identifies the speaker and tags the speech content temporally. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is integrated with the text data and saved as a single unified data set.

[0043] Training a generative AI model

[0044] The server uses the unified dataset to train a generative AI model, which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0045] Simulation Phase

[0046] The user sends a request to the server from their device to conduct an internal review simulation. The user uploads the document they wish to simulate (for example, a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual head of the department reviews it using the generative AI model.

[0047] Conducting a review

[0048] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the situation in an interactive format. The generative AI model reproduces tone and emotional patterns, providing an atmosphere similar to that of a conversation with a real director.

[0049] Providing Feedback

[0050] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0051] Specific examples

[0052] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI Director. The Generative AI Director reviews the proposal in detail and provides comments and questions to the user in real time. Through dialogue with the virtual Director, the user understands which parts of their materials need improvement. Finally, the server generates a feedback report and provides it to the user, allowing them to prepare for the next actual review.

[0053] The above is a specific description of the embodiment of the present invention, which allows the user to perform simulations multiple times and improve the quality of the review.

[0054] The processing flow will be explained below.

[0055] Program processing steps

[0056] Step 1: Data collection

[0057] The server collects historical audio and video data from internal review meetings, either from recorders and cameras or using files exported from existing conferencing systems.

[0058] Step 2: Initial saving of data

[0059] The server temporarily stores the collected audio and video data in storage, where it is used in subsequent processing steps.

[0060] Step 3: Process the audio data

[0061] The server uses a speech recognition engine to convert the voice data into text data, thereby obtaining the spoken content in text format.

[0062] Step 4: Video data processing

[0063] The server uses a video analysis engine to extract facial expressions and movement characteristics from the video data, and through this process, emotion and movement patterns are analyzed.

[0064] Step 5: Integrate the data

[0065] The server combines the text data obtained from the audio data with the features obtained from the video data, creating a single dataset that includes who said what, and what emotions and facial expressions they showed.

[0066] Step 6: Training the generative AI model

[0067] The server uses the combined dataset to train a generative AI model, using machine learning techniques to create a model capable of reproducing the general manager's tone, pronunciation, and emotional patterns.

[0068] Step 7: Saving the generated AI model

[0069] The server stores the trained generative AI model in storage, which is then used in subsequent simulation phases.

[0070] Step 8: Request a simulation

[0071] A user accesses the simulation tool from his / her own terminal and sends a review simulation request to the server.

[0072] Step 9: Upload your materials

[0073] The user uploads the document (e.g., proposal) they wish to review and simulate. The uploaded document is transferred to the server.

[0074] Step 10: Initializing the generative AI model

[0075] The server analyzes the uploaded materials and prepares the simulation. The head of the AI ​​generation department loads the information necessary for reviewing the user's materials.

[0076] Step 11: Conduct a review simulation

[0077] The server launches the Generative AI Headquarters and begins reviewing the content of the materials. The Generative AI Headquarters analyzes the user's materials in detail and provides appropriate comments and questions. The user receives this feedback in real time on their device.

[0078] Step 12: Save the review results

[0079] The server saves the review results of the generated AI chief as a log, which records all review content.

[0080] Step 13: Generate a feedback report

[0081] The server analyzes the reviews and generates a comprehensive feedback report for the user, including suggestions for improvement.

[0082] Step 14: Provide feedback

[0083] The server will provide the generated feedback report to the user, who can then review the feedback on their own device and make any necessary improvements.

[0084] The above are the specific processing steps of this system.

[0085] Example 1

[0086] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0087] Current internal review meetings present difficulties in efficiently reviewing materials, particularly when evaluating new proposals and project plans, as they require time and effort. Furthermore, the quality of the review is heavily dependent on subjective factors, making it difficult to ensure consistency and objectivity. This leaves employees preparing for the review with few opportunities to simulate the actual opinions and comments of the general manager in advance. To address these issues, there is a need to develop a system that efficiently utilizes generative AI models using data from past review meetings, allowing users to effectively perform review simulations.

[0088] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0089] In this invention, the server includes means for collecting and processing audio and video data, means for training a generative AI model based on the processed audio and video data, means for using the generative AI model to review materials submitted by users, means for providing interactive interaction with the users during the review, and means for analyzing the results of the simulation and generating and providing a comprehensive feedback report, thereby enabling users to virtually simulate a review and understand and prepare the review content in advance.

[0090] "Audio data" is a recording format of audio acquired using an audio input device such as a microphone.

[0091] "Video data" refers to the recording format of video captured using a video capture device such as a camera.

[0092] A "generative AI model" is an artificial intelligence model trained using machine learning techniques to reproduce specific speech patterns and emotions.

[0093] "User" refers to a person who uses this system to conduct a document review simulation.

[0094] "Materials" refers to documents such as proposals and project plans uploaded by users and subject to review.

[0095] "Review" refers to the process of feedback and comments provided by the generative AI model.

[0096] "Dialogue format" refers to a format in which the user and the generative AI model interact through dialogue.

[0097] "Feedback" refers to the evaluation and advice that the generative AI model provides to the user on their material.

[0098] A "feedback report" is a document containing a comprehensive evaluation and advice generated based on the simulation results.

[0099] "Interaction" refers to the interaction or communication that takes place between a user and a system.

[0100] "Analysis" refers to the process of processing collected audio and video data and extracting useful information.

[0101] The present invention provides a system that allows users to virtually perform review simulations using a generative AI model based on past data from internal review meetings. The following describes an embodiment of this system.

[0102] Data Collection Phase

[0103] The server collects audio and video data from past internal review meetings. This data is collected using audio and video recording devices such as microphones and cameras, or an existing conference system. The collected audio data is saved in sound wave format, and the video data is saved in video format in the server's storage. This allows the contents of review meetings to be recorded in detail and used for later analysis.

[0104] Data analysis and processing

[0105] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it identifies the speaker and temporally tags the speech content. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. In particular, it is possible to identify the speaker's emotions and changes in facial expression. The analyzed voice data and video data are integrated and saved as a single unified data set.

[0106] Training a generative AI model

[0107] The server uses this unified dataset to train a generative AI model using machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0108] Simulation Phase

[0109] The user sends a request to the server from their device to conduct a review simulation. The user uploads the document they wish to simulate (for example, a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual head of the headquarters reviews it using the generative AI model.

[0110] Conducting reviews and providing feedback

[0111] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to interactively simulate the situation. The generative AI model reproduces tone and emotional patterns to create an atmosphere similar to a conversation with a real director. After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report. This report includes points for improvement, good points, and specific advice. The user can use this as a reference to prepare for the actual review meeting.

[0112] Specific examples

[0113] An employee has prepared a proposal for a new project and wants to have it reviewed. The user accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI Director. The Generative AI Director reviews the proposal's contents in detail and provides comments and questions to the user in real time. Through dialogue with the virtual Director, the user understands which parts of their materials need improvement. Finally, the server generates a feedback report and provides it to the user.

[0114] An example of a prompt is as follows:

[0115] "Please evaluate the following proposal. The title is 'Proposal for a Next-Generation Marketing Strategy.' The contents of the proposal are as follows."

[0116] The above is a specific description of the embodiment of the present invention. Based on this embodiment, the user can perform simulations multiple times to improve the quality of the review.

[0117] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0118] Step 1: Data collection

[0119] The server collects audio and video data from internal review meetings. Specifically, it records the audio and video of the meetings using a microphone and camera. Data exported from existing conference systems can also be used. The input is audio and video data, and the output is audio files in sound wave format and video files in video format. These are stored in the server's storage.

[0120] Step 2: Analyzing the audio data

[0121] The server analyzes the collected voice data using a speech recognition engine. This engine receives an audio file as input and converts the speech into text data. Specifically, it identifies the speaker and adds time tags to the speech. The output is text data that includes speaker information and time tags.

[0122] Step 3: Analyzing the video data

[0123] The server analyzes the collected video data using a video analysis engine. This engine receives video files as input and extracts facial and movement features. Specifically, it identifies changes in the speaker's emotions and adds time tags to the analysis results of facial expressions and movements. The output is the analyzed video data, which includes emotion and movement tags.

[0124] Step 4: Generate a unified dataset

[0125] The server integrates the text data of the audio data with the results of video analysis to generate a unified dataset. The input is the text data and the analyzed video data, and the output is a unified dataset. This dataset includes information such as audio content, speaker identification, time tags, facial expression changes, and motion analysis.

[0126] Step 5: Training the generative AI model

[0127] The server trains the generative AI model using the unified dataset. The input is the unified dataset, and the output is the trained generative AI model. Specifically, it uses a machine learning algorithm to learn the tone, pronunciation, and emotional patterns of the director and generates a trained model.

[0128] Step 6: Processing the User Request

[0129] A user sends a review simulation request from their own device to the server. The input is the user's desired document (proposal or project plan), and the output is the document uploaded to the server.

[0130] Step 7: Analyze and review the materials

[0131] The server analyzes the materials received from the user using a generative AI model and conducts a review. The input is the uploaded materials, and the output is comments and questions from the head of the generative AI department. The generative AI model analyzes the contents of the materials in detail and generates feedback in the form of a statement from the head of the department.

[0132] Step 8: Provide feedback

[0133] The server provides the generated feedback to the user. Specifically, the head of the Generating AI displays comments and questions on the user's device in real time. The input is the generated feedback, and the output is the feedback displayed on the user's device. The user receives this in an interactive format.

[0134] Step 9: Generate a comprehensive report

[0135] After the simulation is complete, the server analyzes the generated feedback and generates a comprehensive report. The input is the feedback data, and the output is a comprehensive feedback report. This report includes points for improvement, good points, and specific advice. Users can use this as a reference to prepare for the actual review meeting.

[0136] (Application example 1)

[0137] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0138] When reviewing operation plans or maintenance reports for industrial equipment, it is difficult to efficiently simulate the opinions of personnel with specialized knowledge and experience. Furthermore, conventional review methods require physical meetings to apply for reviews and receive feedback, which places significant constraints on time and location. This makes it difficult to conduct reviews quickly and accurately.

[0139] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0140] In this invention, the server includes means for collecting and processing audio and video data that are past data from internal review meetings, means for training a generative AI model based on the processed audio and video data, means for simulating a review of materials submitted by a user using the generative AI model, means for providing the simulation results to the user as feedback, means for reviewing operation plans and maintenance reports for industrial equipment using the generative AI model, and means for requesting reviews and uploading materials from a smartphone, thereby enabling fast and accurate reviews without relying on physical meetings.

[0141] A "generative AI model" is an artificial intelligence that is trained based on past data from internal review meetings and can simulate specific speech content and emotional patterns.

[0142] "Audio data" refers to data that has been recorded and saved in digital format, including speech and sound during a meeting.

[0143] "Video data" refers to data that has been recorded and saved in digital format, including the state of a meeting, the facial expressions and movements of speakers, etc.

[0144] "Industrial equipment" refers to machinery and equipment used in factories to process and assemble products, such as robots and production line equipment.

[0145] An "operation plan" is a document that defines the operating procedures and timing of industrial equipment and production lines.

[0146] A "maintenance report" is a report that records the maintenance and repair status of industrial equipment.

[0147] "Feedback" refers to providing evaluation, suggestions for improvement, and advice regarding materials submitted by users.

[0148] A "smartphone" is a portable information terminal equipped with Internet connectivity and application execution capabilities.

[0149] To implement this invention, it is desirable to build the following system, which trains a generative AI model using past data from internal review meetings and performs a virtual review of documents submitted by users.

[0150] System configuration

[0151] The system consists of the following elements:

[0152] 1. Server:

[0153] The server collects, processes, and stores audio and video data, including a speech recognition engine and a video analysis engine. Specifically, it uses the speech_recognition library to convert audio data into text data and the moviepy library to analyze video.

[0154] 2. Generative AI Model:

[0155] The generative AI model reviews user-submitted materials based on the trained data. The generative AI model implements the GPT-3 model using the transformers library.

[0156] 3. User Device:

[0157] Users can access the system from their smartphone or computer and upload operation plans and maintenance reports.

[0158] 4. Building a network:

[0159] A network is required to exchange data between the user's device and the server. To ensure a secure connection, the HTTPS protocol is used.

[0160] Data collection and processing

[0161] The server collects audio and video data from past internal review meetings. The audio data is converted to text using the speech_recognition library. The video data is analyzed using the moviepy library to analyze the speaker's facial expressions and behavior.

[0162] Training a generative AI model

[0163] Based on the audio and video data, a generative AI model was trained using the transformers library to build a GPT-3 model that reproduced the Director's tone and emotional patterns.

[0164] System operation example

[0165] Users access the system from their smartphones and upload operation plans and maintenance reports for industrial equipment. The server receives the documents and performs a virtual review using a generative AI model. Feedback is displayed on the user's device in real time.

[0166] Specific examples

[0167] For example, suppose a user wants to review the following motion plan:

[0168] Please review the following plans:

[0169] The plan states:

[0170] 1. Basic robot movements

[0171] 2. Maintenance Procedures

[0172] 3. Safety Measures

[0173] Please point out the positive and negative aspects of this plan.

[0174] The server inputs this prompt into a generative AI model to generate specific feedback, allowing the user to understand how to improve their motion plan.

[0175] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0176] Step 1:

[0177] The server collects audio and video data from past internal review meetings. Audio data is recorded via a microphone, and video data is recorded via a camera. These data are stored in the server's storage. The input is from the microphone and camera feed, and the output is from audio data files and video data files.

[0178] Step 2:

[0179] The server processes the collected voice data using a voice recognition engine and converts it into text data. This process uses the speech_recognition library. The server inputs the voice file and outputs text data. Specifically, the voice data is passed to the recognition engine, which analyzes the speech and converts it into text format along with timestamps and speaker information.

[0180] Step 3:

[0181] The server processes the video data using a video analysis engine to analyze the speaker's facial expressions and behavioral features. This process uses the moviepy library. It inputs the video file and outputs the analysis results. Specifically, it extracts facial expressions and behaviors from each video frame, identifies the speaker's emotional state and gestures, and digitizes them.

[0182] Step 4:

[0183] The server integrates the audio and video data to generate a unified dataset. It also concatenates the audio-text data and video analysis data and stores them as a single dataset. The input is the audio-text and video analysis results, and the output is a unified dataset.

[0184] Step 5:

[0185] The server trains a generative AI model using a unified dataset. It uses the Transformers library and the GPT-3 model for training. It inputs the unified dataset and outputs a trained generative AI model. Specifically, the dataset is passed to the model to learn specific patterns and features.

[0186] Step 6:

[0187] Users access the system from their own devices and upload the materials they wish to review. The materials are sent as input to the server. Users can then select and upload materials using their smartphones or PCs.

[0188] Step 7:

[0189] The server analyzes the materials uploaded by the user and reviews them using a generative AI model. The server analyzes the document file as input and passes it to the generative AI model as a prompt. Specifically, it analyzes the content of the document, forms an appropriate prompt, and gives it to the generative AI model.

[0190] Step 8:

[0191] The generative AI model reviews materials based on prompts and generates feedback. The prompts are input and the review results are output. Specifically, the model interprets the prompts and generates a textual evaluation of the materials and suggestions for improvement.

[0192] Step 9:

[0193] The server provides the generated feedback to the user in real time by sending the feedback in text format to the user's device. Specifically, the output of the generative AI model is displayed on the user interface so that the user can check the feedback.

[0194] Step 10:

[0195] After the simulation is complete, the server generates a comprehensive report and provides it to the user. The server then aggregates the feedback results and creates a final report. Specifically, it compiles all the feedback, generates a report with specific improvements and advice, and sends it to the user.

[0196] Through the above steps, users can receive a quick and accurate review without relying on a physical meeting.

[0197] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0198] MODE FOR CARRYING OUT THE INVENTION

[0199] The present invention relates to a system that creates a generative AI model using past data from internal review meetings and uses the model to virtually simulate a review. Furthermore, by adding an emotion engine that recognizes the user's emotions, the system aims to provide feedback based on the user's emotional state. The following describes in detail an embodiment of this system.

[0200] Data Collection Phase

[0201] The server collects past audio and video data from internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in the form of sound waves, and the video data is stored in the server's storage in the form of video.

[0202] Data analysis and processing

[0203] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it also identifies the speaker and tags the speech content temporally. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is integrated with the text data and saved as a single unified data set.

[0204] Training a generative AI model

[0205] The server uses the unified dataset to train a generative AI model, which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0206] Simulation Phase

[0207] The user sends a request to the server from their device to conduct an internal review simulation. The user uploads the document they wish to simulate (e.g., a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual general manager uses the generative AI model to conduct the review. Furthermore, the server uses an emotion engine to analyze the user's video data and identify their emotional state in real time.

[0208] Conducting a review

[0209] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the situation interactively. The generative AI model reproduces tone and emotional patterns, providing an atmosphere similar to that of a conversation with a real director. Furthermore, the emotion engine recognizes the user's emotional state and adjusts the content of the generative AI director's feedback. For example, if the user is nervous, the generative AI director can provide feedback in a softer tone.

[0210] Providing Feedback

[0211] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0212] Specific examples

[0213] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI General Manager. The Generative AI General Manager reviews the proposal in detail and provides comments and questions to the user in real time. Through dialogue with the virtual General Manager, the user understands which parts of their document need improvement. In addition, an emotion engine grasps the user's emotional state in real time, and if the user is nervous, the Generative AI General Manager will respond flexibly. Finally, the server generates a feedback report and provides it to the user, allowing them to prepare for the next actual review.

[0214] The above is a detailed description of an embodiment of the present invention. Based on this embodiment, users can perform multiple simulations to improve the quality of their reviews. The emotion engine provides feedback tailored to the user's individual emotional state, which can lead to more effective learning and improvement.

[0215] The processing flow will be explained below.

[0216] Program processing steps

[0217] Step 1: Data collection

[0218] The server collects past audio and video data from internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in the form of sound waves, and the video data is stored in the server's storage in the form of video.

[0219] Step 2: Initial saving of data

[0220] The server temporarily stores the collected audio and video data in storage, where it is used in subsequent processing steps.

[0221] Step 3: Process the audio data

[0222] The server uses a speech recognition engine to convert the voice data into text data, which allows the spoken content to be obtained in text format. It also identifies the speaker and tags the timing of the speech.

[0223] Step 4: Video data processing

[0224] The server uses a video analysis engine to extract facial expressions and movement characteristics from the video data, allowing the speaker's emotions and facial expressions to be analyzed.

[0225] Step 5: Integrate the data

[0226] The server combines the text data obtained from the audio data with the features obtained from the video data, creating a single dataset that includes who said what, and what emotions and facial expressions they showed.

[0227] Step 6: Training the generative AI model

[0228] The server uses the combined dataset to train a generative AI model, using machine learning techniques to create a model capable of reproducing the general manager's tone, pronunciation, and emotional patterns.

[0229] Step 7: Saving the generated AI model

[0230] The server stores the trained generative AI model in storage, which is then used in subsequent simulation phases.

[0231] Step 8: Request a simulation

[0232] A user accesses the simulation tool from his / her own terminal and sends a review simulation request to the server.

[0233] Step 9: Upload your materials

[0234] The user uploads the document (e.g., proposal) they wish to review and simulate. The uploaded document is transferred to the server.

[0235] Step 10: Initializing the generative AI model

[0236] The server analyzes the uploaded materials and prepares the simulation. The head of the AI ​​generation department loads the information necessary for reviewing the user's materials.

[0237] Step 11: Activating the Emotion Engine

[0238] The server activates the emotion engine and acquires video data from the user's device in real time. The emotion engine analyzes this data and identifies the user's emotional state.

[0239] Step 12: Conduct a review simulation

[0240] The server launches the Generative AI Director, who begins reviewing the materials based on their content. The Generative AI Director analyzes the user's materials in detail and provides appropriate comments and questions in real time. The user receives this feedback in real time on their own device. The emotion engine monitors the user's emotional state, and if it detects tension or stress, it adjusts the tone of the Generative AI Director's feedback.

[0241] Step 13: Save the review results

[0242] The server stores the review results of the generated AI chief as a log, which records all review content and emotional state.

[0243] Step 14: Generate a feedback report

[0244] The server analyzes the review results and emotional state and generates a comprehensive feedback report for the user, including suggestions for improvement and recommendations.

[0245] Step 15: Provide feedback

[0246] The server will provide the generated feedback report to the user, who can then review the feedback on their own device and make any necessary improvements.

[0247] The above are the specific processing steps of this system.

[0248] Example 2

[0249] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0250] Conventional internal review meeting systems do not fully utilize past review results, limiting opportunities for users to practice for actual review meetings or receive effective feedback. Furthermore, since feedback does not take into account the user's emotional state, it is difficult to realize true improvements. To solve these issues, a system that can utilize past data and provide feedback that matches the user's emotional state is needed.

[0251] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0252] In this invention, the server includes a means for collecting and processing audio and video data, which are past data from internal review meetings; a means for training a generative AI model using machine learning technology based on the processed audio and video data; and a means including an emotion engine for recognizing the user's emotional state and adjusting the feedback content. This enables simulations based on data from past review meetings, allowing users to receive more effective feedback when preparing for an actual review. Furthermore, the emotion engine enables appropriate responses that take into account the user's tension and anxiety, allowing users to perform review simulations more effectively.

[0253] "Audio data" refers to data in which audio from meetings, reviews, etc. is recorded in digital format.

[0254] "Video data" refers to data that has been recorded in digital format, such as footage of meetings or reviews.

[0255] "Machine learning technology" is a technology that allows a computer to extract patterns from given data and learn automatically.

[0256] A "generative AI model" is an artificial intelligence model trained using machine learning technology to reproduce the tone and emotional patterns of a specific executive.

[0257] An "emotion engine" is a system that recognizes a user's emotional state and adjusts the feedback content based on that state.

[0258] "Review simulation" is a process in which documents submitted by users are virtually reviewed using a generative AI model.

[0259] "Feedback" refers to evaluations and advice provided to users based on the results of the review simulation.

[0260] "Materials" are documents such as proposals and project plans that users submit for review purposes.

[0261] "Server" is the central computing system that performs data collection, analysis, model training and simulation.

[0262] MODE FOR CARRYING OUT THE INVENTION

[0263] The present invention relates to a system that creates a generative AI model using past data from internal review meetings and uses the model to virtually simulate a review. Furthermore, by adding an emotion engine that recognizes the user's emotions, the system aims to provide feedback based on the user's emotional state. The following describes in detail an embodiment of this system.

[0264] Data Collection Phase

[0265] The server collects past audio and video data from internal review meetings. This data is obtained from recorders and cameras, or files exported from existing conference systems are used. For example, audio data is collected in .wav file format, and video data is collected in .mp4 file format. The collected audio and video data is stored in the server's storage.

[0266] Data analysis and processing

[0267] The server processes the collected voice data using a speech recognition engine such as the Google Speech-to-Text API and converts it into text data. This process also identifies the speaker and tags the speech over time. The server then processes the video data using a video analysis engine such as the Microsoft Azure Face API to analyze facial and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is then integrated with the text data and saved as a single unified dataset.

[0268] Training a generative AI model

[0269] The server uses the unified dataset to train the generative AI model. This training uses machine learning techniques such as TensorFlow. The dataset is used to train the model to reproduce the tone, pronunciation, and emotional patterns of a specific executive. Once trained, the generative AI model can faithfully simulate the executive's speech and facial expressions.

[0270] Simulation Phase

[0271] A user sends a request for an internal review simulation to the server from their device. The user uploads the document they wish to simulate (e.g., a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and a virtual executive reviews it using the generative AI model. The server then uses an emotion engine to analyze the user's video data and identify their emotional state in real time.

[0272] Conducting a review

[0273] The generated AI executive reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the interaction in an interactive format. The generative AI model reproduces the executive's tone and emotional patterns, providing an atmosphere similar to that of a conversation with an actual executive. Furthermore, the emotion engine recognizes the user's emotional state and adjusts the content of the generated AI executive's feedback. For example, if the user is nervous, the generated AI executive can provide feedback in a softer tone.

[0274] Providing Feedback

[0275] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0276] Specific examples

[0277] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches a generating AI executive. The generating AI executive reviews the proposal's contents in detail and provides comments and questions to the user in real time. Through dialogue with the virtual executive, the user understands which parts of their document need improvement. In addition, an emotion engine grasps the user's emotional state in real time, and if the user is nervous, the generating AI executive will respond flexibly. Finally, the server generates a feedback report and provides it to the user.

[0278] As a concrete example, let's input the following prompt to the generative AI model:

[0279] "Present your proposal and get feedback from virtual executives."

[0280] The above is a detailed description of an embodiment of the present invention. Based on this embodiment, users can perform multiple simulations to improve the quality of their reviews. The emotion engine provides feedback tailored to the user's individual emotional state, which can lead to more effective learning and improvement.

[0281] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0282] Step 1: Data collection

[0283] The server collects past audio and video data from internal review meetings. Input includes data captured from internal recorders and cameras, or data files exported from existing conference systems. Data processing involves saving the audio data (e.g., .wav files) and video data (e.g., .mp4 files) in a specified directory. The output is the audio and video data saved in storage.

[0284] Step 2: Analyzing the audio data

[0285] The voice data collected by the server is processed by a voice recognition engine (e.g., Google Speech-to-Text API) and converted into text data. The input includes voice data. Data calculations include the process of performing voice recognition and converting the data into text data with time tags and speaker information attached. Specific operations include making an API call and saving the results as a text file. The output is text data with time tags attached.

[0286] Step 3: Analyzing the video data

[0287] The video data collected by the server is processed using a video analysis engine (e.g., Microsoft Azure Face API) to extract facial and movement characteristics. The input includes video data. Data calculations include the process of analyzing the video and extracting facial expression data (e.g., smiling, sadness, anger, etc.) and movement data (e.g., hand movements, eye focus, etc.). Specific operations include making an API call and saving the results in JSON format. Analyzed facial expression and movement data are obtained as output.

[0288] Step 4: Integrate the data

[0289] The server integrates the analyzed audio and video data and stores it as a single unified dataset. The input includes time-tagged text data and analyzed facial and movement data. The data processing involves merging the audio and video data based on the time tags to create a single unified dataset. The specific operation is to store it as a unified dataset in a database. The unified dataset is obtained as the output.

[0290] Step 5: Training the generative AI model

[0291] The server uses a unified dataset to train the generative AI model. The input includes the unified dataset. Data processing involves using a machine learning framework (e.g., TensorFlow) to learn the tone, pronunciation, and emotional patterns of a specific executive. Specifically, the dataset is fed into a neural network model, training is performed, and the optimal model parameters are saved. The output is a trained generative AI model.

[0292] Step 6: Receiving a simulation request

[0293] The user sends a request for an internal review simulation to the server from their own device. The input includes the user's login information and the desired materials for the simulation. Specific operations include uploading the materials through the user interface and sending the request to the server. The output shows the status of the request being sent to the server.

[0294] Step 7: Analyze the data and perform simulations

[0295] The server analyzes the materials received from the user, and the virtual executive reviews them using a generative AI model. Inputs include the user's materials and the generative AI model. Data calculations include analyzing the content of the materials using a text analysis engine (e.g., SpaCy or NLTK) and extracting important points and keywords. This data is passed to the generative AI model, and the virtual executive begins the review. Specifically, the model's output is fed back to the user's device in real time. Real-time comments and questions are provided to the user as output.

[0296] Step 8: Run the Emotion Engine

[0297] The server uses an emotion engine to analyze the user's video data and identify their emotional state in real time. The input includes the user's video data. The data calculation process involves analyzing the data using an emotion recognition algorithm and extracting the user's emotional state (e.g., tension, surprise, relief). Specific operations include calling the emotion engine's API and adjusting the feedback content based on the analysis results. The output is feedback corresponding to the user's emotional state.

[0298] Step 9: Provide feedback

[0299] The generated AI manager evaluates the user's materials and provides appropriate comments and feedback in real time. Specifically, it responds flexibly to the user's tension or anxiety and adjusts the content of the feedback. The output is the comments and advice provided as feedback that is displayed to the user.

[0300] Step 10: Generate a feedback report

[0301] The server analyzes the feedback generated after the simulation has finished and provides a comprehensive report to the user. The input includes a record of all interactions during the simulation. Data processing involves analyzing the log data and summarizing areas for improvement, positives, and specific action items. Specific operations include applying a report generation algorithm to create a report in a format that the user can download (e.g., PDF). The output is a comprehensive feedback report provided to the user.

[0302] The above is a detailed description of the specific processing steps of the system. By following this procedure, users can effectively conduct a review simulation and prepare for an actual review meeting.

[0303] (Application example 2)

[0304] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0305] In the design and development of conventional autonomous vehicles, there is a problem that the review simulations performed by engineers and designers often deviate significantly from the feedback given in actual review meetings. In addition, there is a lack of a means to provide feedback that takes into account the user's emotional state, which hinders more effective improvements.

[0306] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes: means for collecting and processing audio data and video data, which are past data of internal review meetings; means for training a generative AI model based on the processed audio data and video data; means for simulating a review of materials submitted by a user using the generative AI model; means for adjusting feedback content, including an emotion engine that analyzes the user's emotional state; and means for providing the simulation results to the user as feedback. This enables a virtual review simulation using the generative AI model, and can provide feedback based on the user's emotional state, enabling effective improvements in an environment that more closely resembles an actual meeting.

[0307] "Audio data" refers to digital data that records the sounds generated during a meeting, conversation, or the like.

[0308] "Video data" is digital data that records visual information generated during meetings, conversations, etc.

[0309] A "generative AI model" is an artificial intelligence model that is trained using machine learning techniques based on past data and can simulate judgments and actions for specific tasks.

[0310] "Review simulation" is the process of virtually evaluating and examining the materials submitted by the user and providing the results as feedback.

[0311] "Feedback" refers to evaluations and advice obtained as a result of the simulation and is provided to the user.

[0312] An "emotion engine" is a technology that analyzes a user's emotional state from facial expressions, tone of voice, etc.

[0313] "Server" means a computer system for collecting, analyzing data, and training generative AI models.

[0314] "Training" is the process by which a generative AI model learns using past data to improve its performance on a specific task.

[0315] A "user" is a person who uses the system to undergo a review simulation.

[0316] The following describes an embodiment of the present invention.

[0317] System Configuration

[0318] The system of this invention includes a server, a user terminal, an emotion engine, and a generative AI model. The server collects and analyzes data, trains the generative AI model, and sends the generated feedback to the user terminal. The user terminal provides an interface for the user to review the simulation. The emotion engine analyzes the user's video data and detects their emotional state in real time.

[0319] Program processing

[0320] The server collects audio and video data from past internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in sound wave format, and the video data is stored in video format.

[0321] The server then processes the collected voice data using a speech recognition engine (e.g., Google Speech-to-Text API) and converts it into text data. During this process, it also identifies the speaker and tags the speech over time. The video data is then processed using a video analysis engine (e.g., OpenCV) to analyze facial and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is then integrated with the text data and saved as a single unified dataset.

[0322] Using the unified dataset, the server trains a generative AI model (such as GPT-3 or BERT), which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0323] To conduct an internal review simulation, a user sends a simulation request to the server from their own device. The user uploads the materials they wish to simulate (e.g., design documents or prototypes). These materials are transferred to the server and become the subject of review by the generative AI model. The server analyzes the materials, and the virtual head of the department conducts the review using the generative AI model.

[0324] Furthermore, the server uses an emotion engine to analyze the user's video data and identify their emotional state in real time. The Generative AI Headquarters can then adjust the feedback content based on the user's emotional state, enabling more flexible responses. For example, if the user is nervous, the server will provide feedback in a softer tone.

[0325] Specific examples

[0326] An engineer has prepared a design document for a new autonomous vehicle and wants to have it reviewed. The engineer accesses the "Automotive Development Simulation Trainer" from his or her PC and logs into the system. Next, he or she uploads the design document file to the server. The server receives the file and launches the Generative AI Headquarters. The Generative AI Headquarters reviews the contents of the design document in detail and provides comments and questions to the engineer in real time. Through dialogue with the virtual headquarters, the engineer understands which parts of his or her document need improvement. In addition, the emotion engine grasps the engineer's emotional state in real time, and if the engineer seems nervous, the Generative AI Headquarters will respond flexibly. Finally, the server generates a feedback report and provides it to the engineer, allowing him or her to prepare for the next actual review.

[0327] Prompt Sentence Examples

[0328] "Please upload the design document below. We will start the simulation and flexibly adjust the feedback based on the user's emotional state."

[0329] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0330] Step 1:

[0331] The server collects audio and video data from past internal review meetings. The audio and video data is acquired from recorders and cameras, or files exported from existing conference systems are used. This data is saved in the server's storage. The input is audio and video data, and the output is the saved raw data.

[0332] Step 2:

[0333] The server processes the collected voice data using a voice recognition engine (e.g., Google Speech-to-Text API) and converts it into text data. This process also identifies the speaker and tags the speech with time. The input is voice data, and the output is text data with speaker identification and time tags.

[0334] Step 3:

[0335] The server processes the video data using a video analysis engine (e.g., OpenCV) to analyze facial expressions and movement features. This makes it possible to identify the speaker's emotions and changes in facial expressions. The input is video data, and the output is data containing emotional states and facial expression features.

[0336] Step 4:

[0337] The server integrates the analyzed audio and video data and stores them as a unified dataset. The input is text data and facial expression data, and the output is the integrated dataset.

[0338] Step 5:

[0339] The server uses the unified dataset to train a generative AI model (e.g., GPT-3 or BERT). This training uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief executive. The input is the unified dataset, and the output is the trained generative AI model.

[0340] Step 6:

[0341] A user sends a request to the server from their own device to run an internal review simulation. The user uploads materials such as design documents and prototypes. The input is the user request and the uploaded materials, and the output is the transfer of the materials to the server.

[0342] Step 7:

[0343] The server analyzes the uploaded materials, and the virtual director reviews them using a generative AI model. The input is the uploaded materials, and the output is the review comments and feedback of the generative AI director.

[0344] Step 8:

[0345] The server uses an emotion engine to analyze the user's video data and identify the user's emotional state in real time. The input is the user's video data, and the output is the user's emotional state.

[0346] Step 9:

[0347] The Generative AI Director adjusts the feedback content based on the user's emotional state, enabling flexible responses. For example, if the user is nervous, the feedback will be given in a softer tone. The input is the user's emotional state and the Generative AI Director's review comments, and the output is the adjusted feedback.

[0348] Step 10:

[0349] The server generates a feedback report and provides it to the user. The report includes areas for improvement, good points, and specific advice. The input is the tailored feedback, and the output is the feedback report.

[0350] Through these steps, the system enables users to make effective improvements in an environment similar to an actual review meeting.

[0351] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0352] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0353] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0354] [Second embodiment]

[0355] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0356] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0357] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0358] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0359] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0360] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0361] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0362] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0363] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0364] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0365] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0366] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0367] MODE FOR CARRYING OUT THE INVENTION

[0368] The present invention provides a system that allows users to virtually perform review simulations using a generative AI model by using past data from internal review meetings. The following describes an embodiment of this system.

[0369] Data Collection Phase

[0370] The server collects past data (audio and video data) from internal review meetings. This data can be acquired using audio and video recording devices such as microphones and cameras. It is also possible to export data from existing conference systems. The collected audio data is stored in the server's storage in sound wave format, and the video data is stored in video format.

[0371] Data analysis and processing

[0372] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it also identifies the speaker and tags the speech content temporally. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is integrated with the text data and saved as a single unified data set.

[0373] Training a generative AI model

[0374] The server uses the unified dataset to train a generative AI model, which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0375] Simulation Phase

[0376] The user sends a request to the server from their device to conduct an internal review simulation. The user uploads the document they wish to simulate (for example, a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual head of the department reviews it using the generative AI model.

[0377] Conducting a review

[0378] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the situation in an interactive format. The generative AI model reproduces tone and emotional patterns, providing an atmosphere similar to that of a conversation with a real director.

[0379] Providing Feedback

[0380] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0381] Specific examples

[0382] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI Director. The Generative AI Director reviews the proposal in detail and provides comments and questions to the user in real time. Through dialogue with the virtual Director, the user understands which parts of their materials need improvement. Finally, the server generates a feedback report and provides it to the user, allowing them to prepare for the next actual review.

[0383] The above is a specific description of the embodiment of the present invention, which allows the user to perform simulations multiple times and improve the quality of the review.

[0384] The processing flow will be explained below.

[0385] Program processing steps

[0386] Step 1: Data collection

[0387] The server collects historical audio and video data from internal review meetings, either from recorders and cameras or using files exported from existing conferencing systems.

[0388] Step 2: Initial saving of data

[0389] The server temporarily stores the collected audio and video data in storage, where it is used in subsequent processing steps.

[0390] Step 3: Process the audio data

[0391] The server uses a speech recognition engine to convert the voice data into text data, thereby obtaining the spoken content in text format.

[0392] Step 4: Video data processing

[0393] The server uses a video analysis engine to extract facial expressions and movement characteristics from the video data, and through this process, emotion and movement patterns are analyzed.

[0394] Step 5: Integrate the data

[0395] The server combines the text data obtained from the audio data with the features obtained from the video data, creating a single dataset that includes who said what, and what emotions and facial expressions they showed.

[0396] Step 6: Training the generative AI model

[0397] The server uses the combined dataset to train a generative AI model, using machine learning techniques to create a model capable of reproducing the general manager's tone, pronunciation, and emotional patterns.

[0398] Step 7: Saving the generated AI model

[0399] The server stores the trained generative AI model in storage, which is then used in subsequent simulation phases.

[0400] Step 8: Request a simulation

[0401] A user accesses the simulation tool from his / her own terminal and sends a review simulation request to the server.

[0402] Step 9: Upload your materials

[0403] The user uploads the document (e.g., proposal) they wish to review and simulate. The uploaded document is transferred to the server.

[0404] Step 10: Initializing the generative AI model

[0405] The server analyzes the uploaded materials and prepares the simulation. The head of the AI ​​generation department loads the information necessary for reviewing the user's materials.

[0406] Step 11: Conduct a review simulation

[0407] The server launches the Generative AI Headquarters and begins reviewing the content of the materials. The Generative AI Headquarters analyzes the user's materials in detail and provides appropriate comments and questions. The user receives this feedback in real time on their device.

[0408] Step 12: Save the review results

[0409] The server saves the review results of the generated AI chief as a log, which records all review content.

[0410] Step 13: Generate a feedback report

[0411] The server analyzes the reviews and generates a comprehensive feedback report for the user, including suggestions for improvement.

[0412] Step 14: Provide feedback

[0413] The server will provide the generated feedback report to the user, who can then review the feedback on their own device and make any necessary improvements.

[0414] The above are the specific processing steps of this system.

[0415] Example 1

[0416] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0417] Current internal review meetings present difficulties in efficiently reviewing materials, particularly when evaluating new proposals and project plans, as they require time and effort. Furthermore, the quality of the review is heavily dependent on subjective factors, making it difficult to ensure consistency and objectivity. This leaves employees preparing for the review with few opportunities to simulate the actual opinions and comments of the general manager in advance. To address these issues, there is a need to develop a system that efficiently utilizes generative AI models using data from past review meetings, allowing users to effectively perform review simulations.

[0418] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0419] In this invention, the server includes means for collecting and processing audio and video data, means for training a generative AI model based on the processed audio and video data, means for using the generative AI model to review materials submitted by users, means for providing interactive interaction with the users during the review, and means for analyzing the results of the simulation and generating and providing a comprehensive feedback report, thereby enabling users to virtually simulate a review and understand and prepare the review content in advance.

[0420] "Audio data" is a recording format of audio acquired using an audio input device such as a microphone.

[0421] "Video data" refers to the recording format of video captured using a video capture device such as a camera.

[0422] A "generative AI model" is an artificial intelligence model trained using machine learning techniques to reproduce specific speech patterns and emotions.

[0423] "User" refers to a person who uses this system to conduct a document review simulation.

[0424] "Materials" refers to documents such as proposals and project plans uploaded by users and subject to review.

[0425] "Review" refers to the process of feedback and comments provided by the generative AI model.

[0426] "Dialogue format" refers to a format in which the user and the generative AI model interact through dialogue.

[0427] "Feedback" refers to the evaluation and advice that the generative AI model provides to the user on their material.

[0428] A "feedback report" is a document containing a comprehensive evaluation and advice generated based on the simulation results.

[0429] "Interaction" refers to the interaction or communication that takes place between a user and a system.

[0430] "Analysis" refers to the process of processing collected audio and video data and extracting useful information.

[0431] The present invention provides a system that allows users to virtually perform review simulations using a generative AI model based on past data from internal review meetings. The following describes an embodiment of this system.

[0432] Data Collection Phase

[0433] The server collects audio and video data from past internal review meetings. This data is collected using audio and video recording devices such as microphones and cameras, or an existing conference system. The collected audio data is saved in sound wave format, and the video data is saved in video format in the server's storage. This allows the contents of review meetings to be recorded in detail and used for later analysis.

[0434] Data analysis and processing

[0435] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it identifies the speaker and temporally tags the speech content. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. In particular, it is possible to identify the speaker's emotions and changes in facial expression. The analyzed voice data and video data are integrated and saved as a single unified data set.

[0436] Training a generative AI model

[0437] The server uses this unified dataset to train a generative AI model using machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0438] Simulation Phase

[0439] The user sends a request to the server from their device to conduct a review simulation. The user uploads the document they wish to simulate (for example, a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual head of the headquarters reviews it using the generative AI model.

[0440] Conducting reviews and providing feedback

[0441] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to interactively simulate the situation. The generative AI model reproduces tone and emotional patterns to create an atmosphere similar to a conversation with a real director. After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report. This report includes points for improvement, good points, and specific advice. The user can use this as a reference to prepare for the actual review meeting.

[0442] Specific examples

[0443] An employee has prepared a proposal for a new project and wants to have it reviewed. The user accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI Director. The Generative AI Director reviews the proposal's contents in detail and provides comments and questions to the user in real time. Through dialogue with the virtual Director, the user understands which parts of their materials need improvement. Finally, the server generates a feedback report and provides it to the user.

[0444] An example of a prompt is as follows:

[0445] "Please evaluate the following proposal. The title is 'Proposal for a Next-Generation Marketing Strategy.' The contents of the proposal are as follows."

[0446] The above is a specific description of the embodiment of the present invention. Based on this embodiment, the user can perform simulations multiple times to improve the quality of the review.

[0447] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0448] Step 1: Data collection

[0449] The server collects audio and video data from internal review meetings. Specifically, it records the audio and video of the meetings using a microphone and camera. Data exported from existing conference systems can also be used. The input is audio and video data, and the output is audio files in sound wave format and video files in video format. These are stored in the server's storage.

[0450] Step 2: Analyzing the audio data

[0451] The server analyzes the collected voice data using a speech recognition engine. This engine receives an audio file as input and converts the speech into text data. Specifically, it identifies the speaker and adds time tags to the speech. The output is text data that includes speaker information and time tags.

[0452] Step 3: Analyzing the video data

[0453] The server analyzes the collected video data using a video analysis engine. This engine receives video files as input and extracts facial and movement features. Specifically, it identifies changes in the speaker's emotions and adds time tags to the analysis results of facial expressions and movements. The output is the analyzed video data, which includes emotion and movement tags.

[0454] Step 4: Generate a unified dataset

[0455] The server integrates the text data of the audio data with the results of video analysis to generate a unified dataset. The input is the text data and the analyzed video data, and the output is a unified dataset. This dataset includes information such as audio content, speaker identification, time tags, facial expression changes, and motion analysis.

[0456] Step 5: Training the generative AI model

[0457] The server trains the generative AI model using the unified dataset. The input is the unified dataset, and the output is the trained generative AI model. Specifically, it uses a machine learning algorithm to learn the tone, pronunciation, and emotional patterns of the director and generates a trained model.

[0458] Step 6: Processing the User Request

[0459] A user sends a review simulation request from their own device to the server. The input is the user's desired document (proposal or project plan), and the output is the document uploaded to the server.

[0460] Step 7: Analyze and review the materials

[0461] The server analyzes the materials received from the user using a generative AI model and conducts a review. The input is the uploaded materials, and the output is comments and questions from the head of the generative AI department. The generative AI model analyzes the contents of the materials in detail and generates feedback in the form of a statement from the head of the department.

[0462] Step 8: Provide feedback

[0463] The server provides the generated feedback to the user. Specifically, the head of the Generating AI displays comments and questions on the user's device in real time. The input is the generated feedback, and the output is the feedback displayed on the user's device. The user receives this in an interactive format.

[0464] Step 9: Generate a comprehensive report

[0465] After the simulation is complete, the server analyzes the generated feedback and generates a comprehensive report. The input is the feedback data, and the output is a comprehensive feedback report. This report includes points for improvement, good points, and specific advice. Users can use this as a reference to prepare for the actual review meeting.

[0466] (Application example 1)

[0467] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0468] When reviewing operation plans or maintenance reports for industrial equipment, it is difficult to efficiently simulate the opinions of personnel with specialized knowledge and experience. Furthermore, conventional review methods require physical meetings to apply for reviews and receive feedback, which places significant constraints on time and location. This makes it difficult to conduct reviews quickly and accurately.

[0469] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0470] In this invention, the server includes means for collecting and processing audio and video data that are past data from internal review meetings, means for training a generative AI model based on the processed audio and video data, means for simulating a review of materials submitted by a user using the generative AI model, means for providing the simulation results to the user as feedback, means for reviewing operation plans and maintenance reports for industrial equipment using the generative AI model, and means for requesting reviews and uploading materials from a smartphone, thereby enabling fast and accurate reviews without relying on physical meetings.

[0471] A "generative AI model" is an artificial intelligence that is trained based on past data from internal review meetings and can simulate specific speech content and emotional patterns.

[0472] "Audio data" refers to data that has been recorded and saved in digital format, including speech and sound during a meeting.

[0473] "Video data" refers to data that has been recorded and saved in digital format, including the state of a meeting, the facial expressions and movements of speakers, etc.

[0474] "Industrial equipment" refers to machinery and equipment used in factories to process and assemble products, such as robots and production line equipment.

[0475] An "operation plan" is a document that defines the operating procedures and timing of industrial equipment and production lines.

[0476] A "maintenance report" is a report that records the maintenance and repair status of industrial equipment.

[0477] "Feedback" refers to providing evaluation, suggestions for improvement, and advice regarding materials submitted by users.

[0478] A "smartphone" is a portable information terminal equipped with Internet connectivity and application execution capabilities.

[0479] To implement this invention, it is desirable to build the following system, which trains a generative AI model using past data from internal review meetings and performs a virtual review of documents submitted by users.

[0480] System configuration

[0481] The system consists of the following elements:

[0482] 1. Server:

[0483] The server collects, processes, and stores audio and video data, including a speech recognition engine and a video analysis engine. Specifically, it uses the speech_recognition library to convert audio data into text data and the moviepy library to analyze video.

[0484] 2. Generative AI Model:

[0485] The generative AI model reviews user-submitted materials based on the trained data. The generative AI model implements the GPT-3 model using the transformers library.

[0486] 3. User Device:

[0487] Users can access the system from their smartphone or computer and upload operation plans and maintenance reports.

[0488] 4. Building a network:

[0489] A network is required to exchange data between the user's device and the server. To ensure a secure connection, the HTTPS protocol is used.

[0490] Data collection and processing

[0491] The server collects audio and video data from past internal review meetings. The audio data is converted to text using the speech_recognition library. The video data is analyzed using the moviepy library to analyze the speaker's facial expressions and behavior.

[0492] Training a generative AI model

[0493] Based on the audio and video data, a generative AI model was trained using the transformers library to build a GPT-3 model that reproduced the Director's tone and emotional patterns.

[0494] System operation example

[0495] Users access the system from their smartphones and upload operation plans and maintenance reports for industrial equipment. The server receives the documents and performs a virtual review using a generative AI model. Feedback is displayed on the user's device in real time.

[0496] Specific examples

[0497] For example, suppose a user wants to review the following motion plan:

[0498] Please review the following plans:

[0499] The plan states:

[0500] 1. Basic robot movements

[0501] 2. Maintenance Procedures

[0502] 3. Safety Measures

[0503] Please point out the positive and negative aspects of this plan.

[0504] The server inputs this prompt into a generative AI model to generate specific feedback, allowing the user to understand how to improve their motion plan.

[0505] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0506] Step 1:

[0507] The server collects audio and video data from past internal review meetings. Audio data is recorded via a microphone, and video data is recorded via a camera. These data are stored in the server's storage. The input is from the microphone and camera feed, and the output is from audio data files and video data files.

[0508] Step 2:

[0509] The server processes the collected voice data using a voice recognition engine and converts it into text data. This process uses the speech_recognition library. The server inputs the voice file and outputs text data. Specifically, the voice data is passed to the recognition engine, which analyzes the speech and converts it into text format along with timestamps and speaker information.

[0510] Step 3:

[0511] The server processes the video data using a video analysis engine to analyze the speaker's facial expressions and behavioral features. This process uses the moviepy library. It inputs the video file and outputs the analysis results. Specifically, it extracts facial expressions and behaviors from each video frame, identifies the speaker's emotional state and gestures, and digitizes them.

[0512] Step 4:

[0513] The server integrates the audio and video data to generate a unified dataset. It also concatenates the audio-text data and video analysis data and stores them as a single dataset. The input is the audio-text and video analysis results, and the output is a unified dataset.

[0514] Step 5:

[0515] The server trains a generative AI model using a unified dataset. It uses the Transformers library and the GPT-3 model for training. It inputs the unified dataset and outputs a trained generative AI model. Specifically, the dataset is passed to the model to learn specific patterns and features.

[0516] Step 6:

[0517] Users access the system from their own devices and upload the materials they wish to review. The materials are sent as input to the server. Users can then select and upload materials using their smartphones or PCs.

[0518] Step 7:

[0519] The server analyzes the materials uploaded by the user and reviews them using a generative AI model. The server analyzes the document file as input and passes it to the generative AI model as a prompt. Specifically, it analyzes the content of the document, forms an appropriate prompt, and gives it to the generative AI model.

[0520] Step 8:

[0521] The generative AI model reviews materials based on prompts and generates feedback. The prompts are input and the review results are output. Specifically, the model interprets the prompts and generates a textual evaluation of the materials and suggestions for improvement.

[0522] Step 9:

[0523] The server provides the generated feedback to the user in real time by sending the feedback in text format to the user's device. Specifically, the output of the generative AI model is displayed on the user interface so that the user can check the feedback.

[0524] Step 10:

[0525] After the simulation is complete, the server generates a comprehensive report and provides it to the user. The server then aggregates the feedback results and creates a final report. Specifically, it compiles all the feedback, generates a report with specific improvements and advice, and sends it to the user.

[0526] Through the above steps, users can receive a quick and accurate review without relying on a physical meeting.

[0527] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0528] MODE FOR CARRYING OUT THE INVENTION

[0529] The present invention relates to a system that creates a generative AI model using past data from internal review meetings and uses the model to virtually simulate a review. Furthermore, by adding an emotion engine that recognizes the user's emotions, the system aims to provide feedback based on the user's emotional state. The following describes in detail an embodiment of this system.

[0530] Data Collection Phase

[0531] The server collects past audio and video data from internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in the form of sound waves, and the video data is stored in the server's storage in the form of video.

[0532] Data analysis and processing

[0533] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it also identifies the speaker and tags the speech content temporally. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is integrated with the text data and saved as a single unified data set.

[0534] Training a generative AI model

[0535] The server uses the unified dataset to train a generative AI model, which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0536] Simulation Phase

[0537] The user sends a request to the server from their device to conduct an internal review simulation. The user uploads the document they wish to simulate (e.g., a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual general manager uses the generative AI model to conduct the review. Furthermore, the server uses an emotion engine to analyze the user's video data and identify their emotional state in real time.

[0538] Conducting a review

[0539] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the situation interactively. The generative AI model reproduces tone and emotional patterns, providing an atmosphere similar to that of a conversation with a real director. Furthermore, the emotion engine recognizes the user's emotional state and adjusts the content of the generative AI director's feedback. For example, if the user is nervous, the generative AI director can provide feedback in a softer tone.

[0540] Providing Feedback

[0541] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0542] Specific examples

[0543] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI General Manager. The Generative AI General Manager reviews the proposal in detail and provides comments and questions to the user in real time. Through dialogue with the virtual General Manager, the user understands which parts of their document need improvement. In addition, an emotion engine grasps the user's emotional state in real time, and if the user is nervous, the Generative AI General Manager will respond flexibly. Finally, the server generates a feedback report and provides it to the user, allowing them to prepare for the next actual review.

[0544] The above is a detailed description of an embodiment of the present invention. Based on this embodiment, users can perform multiple simulations to improve the quality of their reviews. The emotion engine provides feedback tailored to the user's individual emotional state, which can lead to more effective learning and improvement.

[0545] The processing flow will be explained below.

[0546] Program processing steps

[0547] Step 1: Data collection

[0548] The server collects past audio and video data from internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in the form of sound waves, and the video data is stored in the server's storage in the form of video.

[0549] Step 2: Initial saving of data

[0550] The server temporarily stores the collected audio and video data in storage, where it is used in subsequent processing steps.

[0551] Step 3: Process the audio data

[0552] The server uses a speech recognition engine to convert the voice data into text data, which allows the spoken content to be obtained in text format. It also identifies the speaker and tags the timing of the speech.

[0553] Step 4: Video data processing

[0554] The server uses a video analysis engine to extract facial expressions and movement characteristics from the video data, allowing the speaker's emotions and facial expressions to be analyzed.

[0555] Step 5: Integrate the data

[0556] The server combines the text data obtained from the audio data with the features obtained from the video data, creating a single dataset that includes who said what, and what emotions and facial expressions they showed.

[0557] Step 6: Training the generative AI model

[0558] The server uses the combined dataset to train a generative AI model, using machine learning techniques to create a model capable of reproducing the general manager's tone, pronunciation, and emotional patterns.

[0559] Step 7: Saving the generated AI model

[0560] The server stores the trained generative AI model in storage, which is then used in subsequent simulation phases.

[0561] Step 8: Request a simulation

[0562] A user accesses the simulation tool from his / her own terminal and sends a review simulation request to the server.

[0563] Step 9: Upload your materials

[0564] The user uploads the document (e.g., proposal) they wish to review and simulate. The uploaded document is transferred to the server.

[0565] Step 10: Initializing the generative AI model

[0566] The server analyzes the uploaded materials and prepares the simulation. The head of the AI ​​generation department loads the information necessary for reviewing the user's materials.

[0567] Step 11: Activating the Emotion Engine

[0568] The server activates the emotion engine and acquires video data from the user's device in real time. The emotion engine analyzes this data and identifies the user's emotional state.

[0569] Step 12: Conduct a review simulation

[0570] The server launches the Generative AI Director, who begins reviewing the materials based on their content. The Generative AI Director analyzes the user's materials in detail and provides appropriate comments and questions in real time. The user receives this feedback in real time on their own device. The emotion engine monitors the user's emotional state, and if it detects tension or stress, it adjusts the tone of the Generative AI Director's feedback.

[0571] Step 13: Save the review results

[0572] The server stores the review results of the generated AI chief as a log, which records all review content and emotional state.

[0573] Step 14: Generate a feedback report

[0574] The server analyzes the review results and emotional state and generates a comprehensive feedback report for the user, including suggestions for improvement and recommendations.

[0575] Step 15: Provide feedback

[0576] The server will provide the generated feedback report to the user, who can then review the feedback on their own device and make any necessary improvements.

[0577] The above are the specific processing steps of this system.

[0578] Example 2

[0579] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0580] Conventional internal review meeting systems do not fully utilize past review results, limiting opportunities for users to practice for actual review meetings or receive effective feedback. Furthermore, since feedback does not take into account the user's emotional state, it is difficult to realize true improvements. To solve these issues, a system that can utilize past data and provide feedback that matches the user's emotional state is needed.

[0581] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0582] In this invention, the server includes a means for collecting and processing audio and video data, which are past data from internal review meetings; a means for training a generative AI model using machine learning technology based on the processed audio and video data; and a means including an emotion engine for recognizing the user's emotional state and adjusting the feedback content. This enables simulations based on data from past review meetings, allowing users to receive more effective feedback when preparing for an actual review. Furthermore, the emotion engine enables appropriate responses that take into account the user's tension and anxiety, allowing users to perform review simulations more effectively.

[0583] "Audio data" refers to data in which audio from meetings, reviews, etc. is recorded in digital format.

[0584] "Video data" refers to data that has been recorded in digital format, such as footage of meetings or reviews.

[0585] "Machine learning technology" is a technology that allows a computer to extract patterns from given data and learn automatically.

[0586] A "generative AI model" is an artificial intelligence model trained using machine learning technology to reproduce the tone and emotional patterns of a specific executive.

[0587] An "emotion engine" is a system that recognizes a user's emotional state and adjusts the feedback content based on that state.

[0588] "Review simulation" is a process in which documents submitted by users are virtually reviewed using a generative AI model.

[0589] "Feedback" refers to evaluations and advice provided to users based on the results of the review simulation.

[0590] "Materials" are documents such as proposals and project plans that users submit for review purposes.

[0591] "Server" is the central computing system that performs data collection, analysis, model training and simulation.

[0592] MODE FOR CARRYING OUT THE INVENTION

[0593] The present invention relates to a system that creates a generative AI model using past data from internal review meetings and uses the model to virtually simulate a review. Furthermore, by adding an emotion engine that recognizes the user's emotions, the system aims to provide feedback based on the user's emotional state. The following describes in detail an embodiment of this system.

[0594] Data Collection Phase

[0595] The server collects past audio and video data from internal review meetings. This data is obtained from recorders and cameras, or files exported from existing conference systems are used. For example, audio data is collected in .wav file format, and video data is collected in .mp4 file format. The collected audio and video data is stored in the server's storage.

[0596] Data analysis and processing

[0597] The server processes the collected voice data using a speech recognition engine such as the Google Speech-to-Text API and converts it into text data. This process also identifies the speaker and tags the speech over time. The server then processes the video data using a video analysis engine such as the Microsoft Azure Face API to analyze facial and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is then integrated with the text data and saved as a single unified dataset.

[0598] Training a generative AI model

[0599] The server uses the unified dataset to train the generative AI model. This training uses machine learning techniques such as TensorFlow. The dataset is used to train the model to reproduce the tone, pronunciation, and emotional patterns of a specific executive. Once trained, the generative AI model can faithfully simulate the executive's speech and facial expressions.

[0600] Simulation Phase

[0601] A user sends a request for an internal review simulation to the server from their device. The user uploads the document they wish to simulate (e.g., a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and a virtual executive reviews it using the generative AI model. The server then uses an emotion engine to analyze the user's video data and identify their emotional state in real time.

[0602] Conducting a review

[0603] The generated AI executive reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the interaction in an interactive format. The generative AI model reproduces the executive's tone and emotional patterns, providing an atmosphere similar to that of a conversation with an actual executive. Furthermore, the emotion engine recognizes the user's emotional state and adjusts the content of the generated AI executive's feedback. For example, if the user is nervous, the generated AI executive can provide feedback in a softer tone.

[0604] Providing Feedback

[0605] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0606] Specific examples

[0607] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches a generating AI executive. The generating AI executive reviews the proposal's contents in detail and provides comments and questions to the user in real time. Through dialogue with the virtual executive, the user understands which parts of their document need improvement. In addition, an emotion engine grasps the user's emotional state in real time, and if the user is nervous, the generating AI executive will respond flexibly. Finally, the server generates a feedback report and provides it to the user.

[0608] As a concrete example, let's input the following prompt to the generative AI model:

[0609] "Present your proposal and get feedback from virtual executives."

[0610] The above is a detailed description of an embodiment of the present invention. Based on this embodiment, users can perform multiple simulations to improve the quality of their reviews. The emotion engine provides feedback tailored to the user's individual emotional state, which can lead to more effective learning and improvement.

[0611] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0612] Step 1: Data collection

[0613] The server collects past audio and video data from internal review meetings. Input includes data captured from internal recorders and cameras, or data files exported from existing conference systems. Data processing involves saving the audio data (e.g., .wav files) and video data (e.g., .mp4 files) in a specified directory. The output is the audio and video data saved in storage.

[0614] Step 2: Analyzing the audio data

[0615] The voice data collected by the server is processed by a voice recognition engine (e.g., Google Speech-to-Text API) and converted into text data. The input includes voice data. Data calculations include the process of performing voice recognition and converting the data into text data with time tags and speaker information attached. Specific operations include making an API call and saving the results as a text file. The output is text data with time tags attached.

[0616] Step 3: Analyzing the video data

[0617] The video data collected by the server is processed using a video analysis engine (e.g., Microsoft Azure Face API) to extract facial and movement characteristics. The input includes video data. Data calculations include the process of analyzing the video and extracting facial expression data (e.g., smiling, sadness, anger, etc.) and movement data (e.g., hand movements, eye focus, etc.). Specific operations include making an API call and saving the results in JSON format. Analyzed facial expression and movement data are obtained as output.

[0618] Step 4: Integrate the data

[0619] The server integrates the analyzed audio and video data and stores it as a single unified dataset. The input includes time-tagged text data and analyzed facial and movement data. The data processing involves merging the audio and video data based on the time tags to create a single unified dataset. The specific operation is to store it as a unified dataset in a database. The unified dataset is obtained as the output.

[0620] Step 5: Training the generative AI model

[0621] The server uses a unified dataset to train the generative AI model. The input includes the unified dataset. Data processing involves using a machine learning framework (e.g., TensorFlow) to learn the tone, pronunciation, and emotional patterns of a specific executive. Specifically, the dataset is fed into a neural network model, training is performed, and the optimal model parameters are saved. The output is a trained generative AI model.

[0622] Step 6: Receiving a simulation request

[0623] The user sends a request for an internal review simulation to the server from their own device. The input includes the user's login information and the desired materials for the simulation. Specific operations include uploading the materials through the user interface and sending the request to the server. The output shows the status of the request being sent to the server.

[0624] Step 7: Analyze the data and perform simulations

[0625] The server analyzes the materials received from the user, and the virtual executive reviews them using a generative AI model. Inputs include the user's materials and the generative AI model. Data calculations include analyzing the content of the materials using a text analysis engine (e.g., SpaCy or NLTK) and extracting important points and keywords. This data is passed to the generative AI model, and the virtual executive begins the review. Specifically, the model's output is fed back to the user's device in real time. Real-time comments and questions are provided to the user as output.

[0626] Step 8: Run the Emotion Engine

[0627] The server uses an emotion engine to analyze the user's video data and identify their emotional state in real time. The input includes the user's video data. The data calculation process involves analyzing the data using an emotion recognition algorithm and extracting the user's emotional state (e.g., tension, surprise, relief). Specific operations include calling the emotion engine's API and adjusting the feedback content based on the analysis results. The output is feedback corresponding to the user's emotional state.

[0628] Step 9: Provide feedback

[0629] The generated AI manager evaluates the user's materials and provides appropriate comments and feedback in real time. Specifically, it responds flexibly to the user's tension or anxiety and adjusts the content of the feedback. The output is the comments and advice provided as feedback that is displayed to the user.

[0630] Step 10: Generate a feedback report

[0631] The server analyzes the feedback generated after the simulation has finished and provides a comprehensive report to the user. The input includes a record of all interactions during the simulation. Data processing involves analyzing the log data and summarizing areas for improvement, positives, and specific action items. Specific operations include applying a report generation algorithm to create a report in a format that the user can download (e.g., PDF). The output is a comprehensive feedback report provided to the user.

[0632] The above is a detailed description of the specific processing steps of the system. By following this procedure, users can effectively conduct a review simulation and prepare for an actual review meeting.

[0633] (Application example 2)

[0634] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0635] In the design and development of conventional autonomous vehicles, there is a problem that the review simulations performed by engineers and designers often deviate significantly from the feedback given in actual review meetings. In addition, there is a lack of a means to provide feedback that takes into account the user's emotional state, which hinders more effective improvements.

[0636] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes: means for collecting and processing audio data and video data, which are past data of internal review meetings; means for training a generative AI model based on the processed audio data and video data; means for simulating a review of materials submitted by a user using the generative AI model; means for adjusting feedback content, including an emotion engine that analyzes the user's emotional state; and means for providing the simulation results to the user as feedback. This enables a virtual review simulation using the generative AI model, and can provide feedback based on the user's emotional state, enabling effective improvements in an environment that more closely resembles an actual meeting.

[0637] "Audio data" refers to digital data that records the sounds generated during a meeting, conversation, or the like.

[0638] "Video data" is digital data that records visual information generated during meetings, conversations, etc.

[0639] A "generative AI model" is an artificial intelligence model that is trained using machine learning techniques based on past data and can simulate judgments and actions for specific tasks.

[0640] "Review simulation" is the process of virtually evaluating and examining the materials submitted by the user and providing the results as feedback.

[0641] "Feedback" refers to evaluations and advice obtained as a result of the simulation and is provided to the user.

[0642] An "emotion engine" is a technology that analyzes a user's emotional state from facial expressions, tone of voice, etc.

[0643] "Server" means a computer system for collecting, analyzing data, and training generative AI models.

[0644] "Training" is the process by which a generative AI model learns using past data to improve its performance on a specific task.

[0645] A "user" is a person who uses the system to undergo a review simulation.

[0646] The following describes an embodiment of the present invention.

[0647] System Configuration

[0648] The system of this invention includes a server, a user terminal, an emotion engine, and a generative AI model. The server collects and analyzes data, trains the generative AI model, and sends the generated feedback to the user terminal. The user terminal provides an interface for the user to review the simulation. The emotion engine analyzes the user's video data and detects their emotional state in real time.

[0649] Program processing

[0650] The server collects audio and video data from past internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in sound wave format, and the video data is stored in video format.

[0651] The server then processes the collected voice data using a speech recognition engine (e.g., Google Speech-to-Text API) and converts it into text data. During this process, it also identifies the speaker and tags the speech over time. The video data is then processed using a video analysis engine (e.g., OpenCV) to analyze facial and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is then integrated with the text data and saved as a single unified dataset.

[0652] Using the unified dataset, the server trains a generative AI model (such as GPT-3 or BERT), which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0653] To conduct an internal review simulation, a user sends a simulation request to the server from their own device. The user uploads the materials they wish to simulate (e.g., design documents or prototypes). These materials are transferred to the server and become the subject of review by the generative AI model. The server analyzes the materials, and the virtual head of the department conducts the review using the generative AI model.

[0654] Furthermore, the server uses an emotion engine to analyze the user's video data and identify their emotional state in real time. The Generative AI Headquarters can then adjust the feedback content based on the user's emotional state, enabling more flexible responses. For example, if the user is nervous, the server will provide feedback in a softer tone.

[0655] Specific examples

[0656] An engineer has prepared a design document for a new autonomous vehicle and wants to have it reviewed. The engineer accesses the "Automotive Development Simulation Trainer" from his or her PC and logs into the system. Next, he or she uploads the design document file to the server. The server receives the file and launches the Generative AI Headquarters. The Generative AI Headquarters reviews the contents of the design document in detail and provides comments and questions to the engineer in real time. Through dialogue with the virtual headquarters, the engineer understands which parts of his or her document need improvement. In addition, the emotion engine grasps the engineer's emotional state in real time, and if the engineer seems nervous, the Generative AI Headquarters will respond flexibly. Finally, the server generates a feedback report and provides it to the engineer, allowing him or her to prepare for the next actual review.

[0657] Prompt Sentence Examples

[0658] "Please upload the design document below. We will start the simulation and flexibly adjust the feedback based on the user's emotional state."

[0659] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0660] Step 1:

[0661] The server collects audio and video data from past internal review meetings. The audio and video data is acquired from recorders and cameras, or files exported from existing conference systems are used. This data is saved in the server's storage. The input is audio and video data, and the output is the saved raw data.

[0662] Step 2:

[0663] The server processes the collected voice data using a voice recognition engine (e.g., Google Speech-to-Text API) and converts it into text data. This process also identifies the speaker and tags the speech with time. The input is voice data, and the output is text data with speaker identification and time tags.

[0664] Step 3:

[0665] The server processes the video data using a video analysis engine (e.g., OpenCV) to analyze facial expressions and movement features. This makes it possible to identify the speaker's emotions and changes in facial expressions. The input is video data, and the output is data containing emotional states and facial expression features.

[0666] Step 4:

[0667] The server integrates the analyzed audio and video data and stores them as a unified dataset. The input is text data and facial expression data, and the output is the integrated dataset.

[0668] Step 5:

[0669] The server uses the unified dataset to train a generative AI model (e.g., GPT-3 or BERT). This training uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief executive. The input is the unified dataset, and the output is the trained generative AI model.

[0670] Step 6:

[0671] A user sends a request to the server from their own device to run an internal review simulation. The user uploads materials such as design documents and prototypes. The input is the user request and the uploaded materials, and the output is the transfer of the materials to the server.

[0672] Step 7:

[0673] The server analyzes the uploaded materials, and the virtual director reviews them using a generative AI model. The input is the uploaded materials, and the output is the review comments and feedback of the generative AI director.

[0674] Step 8:

[0675] The server uses an emotion engine to analyze the user's video data and identify the user's emotional state in real time. The input is the user's video data, and the output is the user's emotional state.

[0676] Step 9:

[0677] The Generative AI Director adjusts the feedback content based on the user's emotional state, enabling flexible responses. For example, if the user is nervous, the feedback will be given in a softer tone. The input is the user's emotional state and the Generative AI Director's review comments, and the output is the adjusted feedback.

[0678] Step 10:

[0679] The server generates a feedback report and provides it to the user. The report includes areas for improvement, good points, and specific advice. The input is the tailored feedback, and the output is the feedback report.

[0680] Through these steps, the system enables users to make effective improvements in an environment similar to an actual review meeting.

[0681] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0682] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0683] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0684] [Third embodiment]

[0685] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0686] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0687] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0688] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0689] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0690] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0691] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0692] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0693] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0694] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0695] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0696] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0697] MODE FOR CARRYING OUT THE INVENTION

[0698] The present invention provides a system that allows users to virtually perform review simulations using a generative AI model by using past data from internal review meetings. The following describes an embodiment of this system.

[0699] Data Collection Phase

[0700] The server collects past data (audio and video data) from internal review meetings. This data can be acquired using audio and video recording devices such as microphones and cameras. It is also possible to export data from existing conference systems. The collected audio data is stored in the server's storage in sound wave format, and the video data is stored in video format.

[0701] Data analysis and processing

[0702] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it also identifies the speaker and tags the speech content temporally. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is integrated with the text data and saved as a single unified data set.

[0703] Training a generative AI model

[0704] The server uses the unified dataset to train a generative AI model, which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0705] Simulation Phase

[0706] The user sends a request to the server from their device to conduct an internal review simulation. The user uploads the document they wish to simulate (for example, a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual head of the department reviews it using the generative AI model.

[0707] Conducting a review

[0708] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the situation in an interactive format. The generative AI model reproduces tone and emotional patterns, providing an atmosphere similar to that of a conversation with a real director.

[0709] Providing Feedback

[0710] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0711] Specific examples

[0712] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI Director. The Generative AI Director reviews the proposal in detail and provides comments and questions to the user in real time. Through dialogue with the virtual Director, the user understands which parts of their materials need improvement. Finally, the server generates a feedback report and provides it to the user, allowing them to prepare for the next actual review.

[0713] The above is a specific description of the embodiment of the present invention, which allows the user to perform simulations multiple times and improve the quality of the review.

[0714] The processing flow will be explained below.

[0715] Program processing steps

[0716] Step 1: Data collection

[0717] The server collects historical audio and video data from internal review meetings, either from recorders and cameras or using files exported from existing conferencing systems.

[0718] Step 2: Initial saving of data

[0719] The server temporarily stores the collected audio and video data in storage, where it is used in subsequent processing steps.

[0720] Step 3: Process the audio data

[0721] The server uses a speech recognition engine to convert the voice data into text data, thereby obtaining the spoken content in text format.

[0722] Step 4: Video data processing

[0723] The server uses a video analysis engine to extract facial expressions and movement characteristics from the video data, and through this process, emotion and movement patterns are analyzed.

[0724] Step 5: Integrate the data

[0725] The server combines the text data obtained from the audio data with the features obtained from the video data, creating a single dataset that includes who said what, and what emotions and facial expressions they showed.

[0726] Step 6: Training the generative AI model

[0727] The server uses the combined dataset to train a generative AI model, using machine learning techniques to create a model capable of reproducing the general manager's tone, pronunciation, and emotional patterns.

[0728] Step 7: Saving the generated AI model

[0729] The server stores the trained generative AI model in storage, which is then used in subsequent simulation phases.

[0730] Step 8: Request a simulation

[0731] A user accesses the simulation tool from his / her own terminal and sends a review simulation request to the server.

[0732] Step 9: Upload your materials

[0733] The user uploads the document (e.g., proposal) they wish to review and simulate. The uploaded document is transferred to the server.

[0734] Step 10: Initializing the generative AI model

[0735] The server analyzes the uploaded materials and prepares the simulation. The head of the AI ​​generation department loads the information necessary for reviewing the user's materials.

[0736] Step 11: Conduct a review simulation

[0737] The server launches the Generative AI Headquarters and begins reviewing the content of the materials. The Generative AI Headquarters analyzes the user's materials in detail and provides appropriate comments and questions. The user receives this feedback in real time on their device.

[0738] Step 12: Save the review results

[0739] The server saves the review results of the generated AI chief as a log, which records all review content.

[0740] Step 13: Generate a feedback report

[0741] The server analyzes the reviews and generates a comprehensive feedback report for the user, including suggestions for improvement.

[0742] Step 14: Provide feedback

[0743] The server will provide the generated feedback report to the user, who can then review the feedback on their own device and make any necessary improvements.

[0744] The above are the specific processing steps of this system.

[0745] Example 1

[0746] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0747] Current internal review meetings present difficulties in efficiently reviewing materials, particularly when evaluating new proposals and project plans, as they require time and effort. Furthermore, the quality of the review is heavily dependent on subjective factors, making it difficult to ensure consistency and objectivity. This leaves employees preparing for the review with few opportunities to simulate the actual opinions and comments of the general manager in advance. To address these issues, there is a need to develop a system that efficiently utilizes generative AI models using data from past review meetings, allowing users to effectively perform review simulations.

[0748] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0749] In this invention, the server includes means for collecting and processing audio and video data, means for training a generative AI model based on the processed audio and video data, means for using the generative AI model to review materials submitted by users, means for providing interactive interaction with the users during the review, and means for analyzing the results of the simulation and generating and providing a comprehensive feedback report, thereby enabling users to virtually simulate a review and understand and prepare the review content in advance.

[0750] "Audio data" is a recording format of audio acquired using an audio input device such as a microphone.

[0751] "Video data" refers to the recording format of video captured using a video capture device such as a camera.

[0752] A "generative AI model" is an artificial intelligence model trained using machine learning techniques to reproduce specific speech patterns and emotions.

[0753] "User" refers to a person who uses this system to conduct a document review simulation.

[0754] "Materials" refers to documents such as proposals and project plans uploaded by users and subject to review.

[0755] "Review" refers to the process of feedback and comments provided by the generative AI model.

[0756] "Dialogue format" refers to a format in which the user and the generative AI model interact through dialogue.

[0757] "Feedback" refers to the evaluation and advice that the generative AI model provides to the user on their material.

[0758] A "feedback report" is a document containing a comprehensive evaluation and advice generated based on the simulation results.

[0759] "Interaction" refers to the interaction or communication that takes place between a user and a system.

[0760] "Analysis" refers to the process of processing collected audio and video data and extracting useful information.

[0761] The present invention provides a system that allows users to virtually perform review simulations using a generative AI model based on past data from internal review meetings. The following describes an embodiment of this system.

[0762] Data Collection Phase

[0763] The server collects audio and video data from past internal review meetings. This data is collected using audio and video recording devices such as microphones and cameras, or an existing conference system. The collected audio data is saved in sound wave format, and the video data is saved in video format in the server's storage. This allows the contents of review meetings to be recorded in detail and used for later analysis.

[0764] Data analysis and processing

[0765] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it identifies the speaker and temporally tags the speech content. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. In particular, it is possible to identify the speaker's emotions and changes in facial expression. The analyzed voice data and video data are integrated and saved as a single unified data set.

[0766] Training a generative AI model

[0767] The server uses this unified dataset to train a generative AI model using machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0768] Simulation Phase

[0769] The user sends a request to the server from their device to conduct a review simulation. The user uploads the document they wish to simulate (for example, a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual head of the headquarters reviews it using the generative AI model.

[0770] Conducting reviews and providing feedback

[0771] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to interactively simulate the situation. The generative AI model reproduces tone and emotional patterns to create an atmosphere similar to a conversation with a real director. After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report. This report includes points for improvement, good points, and specific advice. The user can use this as a reference to prepare for the actual review meeting.

[0772] Specific examples

[0773] An employee has prepared a proposal for a new project and wants to have it reviewed. The user accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI Director. The Generative AI Director reviews the proposal's contents in detail and provides comments and questions to the user in real time. Through dialogue with the virtual Director, the user understands which parts of their materials need improvement. Finally, the server generates a feedback report and provides it to the user.

[0774] An example of a prompt is as follows:

[0775] "Please evaluate the following proposal. The title is 'Proposal for a Next-Generation Marketing Strategy.' The contents of the proposal are as follows."

[0776] The above is a specific description of the embodiment of the present invention. Based on this embodiment, the user can perform simulations multiple times to improve the quality of the review.

[0777] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0778] Step 1: Data collection

[0779] The server collects audio and video data from internal review meetings. Specifically, it records the audio and video of the meetings using a microphone and camera. Data exported from existing conference systems can also be used. The input is audio and video data, and the output is audio files in sound wave format and video files in video format. These are stored in the server's storage.

[0780] Step 2: Analyzing the audio data

[0781] The server analyzes the collected voice data using a speech recognition engine. This engine receives an audio file as input and converts the speech into text data. Specifically, it identifies the speaker and adds time tags to the speech. The output is text data that includes speaker information and time tags.

[0782] Step 3: Analyzing the video data

[0783] The server analyzes the collected video data using a video analysis engine. This engine receives video files as input and extracts facial and movement features. Specifically, it identifies changes in the speaker's emotions and adds time tags to the analysis results of facial expressions and movements. The output is the analyzed video data, which includes emotion and movement tags.

[0784] Step 4: Generate a unified dataset

[0785] The server integrates the text data of the audio data with the results of video analysis to generate a unified dataset. The input is the text data and the analyzed video data, and the output is a unified dataset. This dataset includes information such as audio content, speaker identification, time tags, facial expression changes, and motion analysis.

[0786] Step 5: Training the generative AI model

[0787] The server trains the generative AI model using the unified dataset. The input is the unified dataset, and the output is the trained generative AI model. Specifically, it uses a machine learning algorithm to learn the tone, pronunciation, and emotional patterns of the director and generates a trained model.

[0788] Step 6: Processing the User Request

[0789] A user sends a review simulation request from their own device to the server. The input is the user's desired document (proposal or project plan), and the output is the document uploaded to the server.

[0790] Step 7: Analyze and review the materials

[0791] The server analyzes the materials received from the user using a generative AI model and conducts a review. The input is the uploaded materials, and the output is comments and questions from the head of the generative AI department. The generative AI model analyzes the contents of the materials in detail and generates feedback in the form of a statement from the head of the department.

[0792] Step 8: Provide feedback

[0793] The server provides the generated feedback to the user. Specifically, the head of the Generating AI displays comments and questions on the user's device in real time. The input is the generated feedback, and the output is the feedback displayed on the user's device. The user receives this in an interactive format.

[0794] Step 9: Generate a comprehensive report

[0795] After the simulation is complete, the server analyzes the generated feedback and generates a comprehensive report. The input is the feedback data, and the output is a comprehensive feedback report. This report includes points for improvement, good points, and specific advice. Users can use this as a reference to prepare for the actual review meeting.

[0796] (Application example 1)

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

[0798] When reviewing operation plans or maintenance reports for industrial equipment, it is difficult to efficiently simulate the opinions of personnel with specialized knowledge and experience. Furthermore, conventional review methods require physical meetings to apply for reviews and receive feedback, which places significant constraints on time and location. This makes it difficult to conduct reviews quickly and accurately.

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

[0800] In this invention, the server includes means for collecting and processing audio and video data that are past data from internal review meetings, means for training a generative AI model based on the processed audio and video data, means for simulating a review of materials submitted by a user using the generative AI model, means for providing the simulation results to the user as feedback, means for reviewing operation plans and maintenance reports for industrial equipment using the generative AI model, and means for requesting reviews and uploading materials from a smartphone, thereby enabling fast and accurate reviews without relying on physical meetings.

[0801] A "generative AI model" is an artificial intelligence that is trained based on past data from internal review meetings and can simulate specific speech content and emotional patterns.

[0802] "Audio data" refers to data that has been recorded and saved in digital format, including speech and sound during a meeting.

[0803] "Video data" refers to data that has been recorded and saved in digital format, including the state of a meeting, the facial expressions and movements of speakers, etc.

[0804] "Industrial equipment" refers to machinery and equipment used in factories to process and assemble products, such as robots and production line equipment.

[0805] An "operation plan" is a document that defines the operating procedures and timing of industrial equipment and production lines.

[0806] A "maintenance report" is a report that records the maintenance and repair status of industrial equipment.

[0807] "Feedback" refers to providing evaluation, suggestions for improvement, and advice regarding materials submitted by users.

[0808] A "smartphone" is a portable information terminal equipped with Internet connectivity and application execution capabilities.

[0809] To implement this invention, it is desirable to build the following system, which trains a generative AI model using past data from internal review meetings and performs a virtual review of documents submitted by users.

[0810] System configuration

[0811] The system consists of the following elements:

[0812] 1. Server:

[0813] The server collects, processes, and stores audio and video data, including a speech recognition engine and a video analysis engine. Specifically, it uses the speech_recognition library to convert audio data into text data and the moviepy library to analyze video.

[0814] 2. Generative AI Model:

[0815] The generative AI model reviews user-submitted materials based on the trained data. The generative AI model implements the GPT-3 model using the transformers library.

[0816] 3. User Device:

[0817] Users can access the system from their smartphone or computer and upload operation plans and maintenance reports.

[0818] 4. Building a network:

[0819] A network is required to exchange data between the user's device and the server. To ensure a secure connection, the HTTPS protocol is used.

[0820] Data collection and processing

[0821] The server collects audio and video data from past internal review meetings. The audio data is converted to text using the speech_recognition library. The video data is analyzed using the moviepy library to analyze the speaker's facial expressions and behavior.

[0822] Training a generative AI model

[0823] Based on the audio and video data, a generative AI model was trained using the transformers library to build a GPT-3 model that reproduced the Director's tone and emotional patterns.

[0824] System operation example

[0825] Users access the system from their smartphones and upload operation plans and maintenance reports for industrial equipment. The server receives the documents and performs a virtual review using a generative AI model. Feedback is displayed on the user's device in real time.

[0826] Specific examples

[0827] For example, suppose a user wants to review the following motion plan:

[0828] Please review the following plans:

[0829] The plan states:

[0830] 1. Basic robot movements

[0831] 2. Maintenance Procedures

[0832] 3. Safety Measures

[0833] Please point out the positive and negative aspects of this plan.

[0834] The server inputs this prompt into a generative AI model to generate specific feedback, allowing the user to understand how to improve their motion plan.

[0835] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0836] Step 1:

[0837] The server collects audio and video data from past internal review meetings. Audio data is recorded via a microphone, and video data is recorded via a camera. These data are stored in the server's storage. The input is from the microphone and camera feed, and the output is from audio data files and video data files.

[0838] Step 2:

[0839] The server processes the collected voice data using a voice recognition engine and converts it into text data. This process uses the speech_recognition library. The server inputs the voice file and outputs text data. Specifically, the voice data is passed to the recognition engine, which analyzes the speech and converts it into text format along with timestamps and speaker information.

[0840] Step 3:

[0841] The server processes the video data using a video analysis engine to analyze the speaker's facial expressions and behavioral features. This process uses the moviepy library. It inputs the video file and outputs the analysis results. Specifically, it extracts facial expressions and behaviors from each video frame, identifies the speaker's emotional state and gestures, and digitizes them.

[0842] Step 4:

[0843] The server integrates the audio and video data to generate a unified dataset. It also concatenates the audio-text data and video analysis data and stores them as a single dataset. The input is the audio-text and video analysis results, and the output is a unified dataset.

[0844] Step 5:

[0845] The server trains a generative AI model using a unified dataset. It uses the Transformers library and the GPT-3 model for training. It inputs the unified dataset and outputs a trained generative AI model. Specifically, the dataset is passed to the model to learn specific patterns and features.

[0846] Step 6:

[0847] Users access the system from their own devices and upload the materials they wish to review. The materials are sent as input to the server. Users can then select and upload materials using their smartphones or PCs.

[0848] Step 7:

[0849] The server analyzes the materials uploaded by the user and reviews them using a generative AI model. The server analyzes the document file as input and passes it to the generative AI model as a prompt. Specifically, it analyzes the content of the document, forms an appropriate prompt, and gives it to the generative AI model.

[0850] Step 8:

[0851] The generative AI model reviews materials based on prompts and generates feedback. The prompts are input and the review results are output. Specifically, the model interprets the prompts and generates a textual evaluation of the materials and suggestions for improvement.

[0852] Step 9:

[0853] The server provides the generated feedback to the user in real time by sending the feedback in text format to the user's device. Specifically, the output of the generative AI model is displayed on the user interface so that the user can check the feedback.

[0854] Step 10:

[0855] After the simulation is complete, the server generates a comprehensive report and provides it to the user. The server then aggregates the feedback results and creates a final report. Specifically, it compiles all the feedback, generates a report with specific improvements and advice, and sends it to the user.

[0856] Through the above steps, users can receive a quick and accurate review without relying on a physical meeting.

[0857] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0858] MODE FOR CARRYING OUT THE INVENTION

[0859] The present invention relates to a system that creates a generative AI model using past data from internal review meetings and uses the model to virtually simulate a review. Furthermore, by adding an emotion engine that recognizes the user's emotions, the system aims to provide feedback based on the user's emotional state. The following describes in detail an embodiment of this system.

[0860] Data Collection Phase

[0861] The server collects past audio and video data from internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in the form of sound waves, and the video data is stored in the server's storage in the form of video.

[0862] Data analysis and processing

[0863] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it also identifies the speaker and tags the speech content temporally. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is integrated with the text data and saved as a single unified data set.

[0864] Training a generative AI model

[0865] The server uses the unified dataset to train a generative AI model, which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0866] Simulation Phase

[0867] The user sends a request to the server from their device to conduct an internal review simulation. The user uploads the document they wish to simulate (e.g., a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual general manager uses the generative AI model to conduct the review. Furthermore, the server uses an emotion engine to analyze the user's video data and identify their emotional state in real time.

[0868] Conducting a review

[0869] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the situation interactively. The generative AI model reproduces tone and emotional patterns, providing an atmosphere similar to that of a conversation with a real director. Furthermore, the emotion engine recognizes the user's emotional state and adjusts the content of the generative AI director's feedback. For example, if the user is nervous, the generative AI director can provide feedback in a softer tone.

[0870] Providing Feedback

[0871] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0872] Specific examples

[0873] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI General Manager. The Generative AI General Manager reviews the proposal in detail and provides comments and questions to the user in real time. Through dialogue with the virtual General Manager, the user understands which parts of their document need improvement. In addition, an emotion engine grasps the user's emotional state in real time, and if the user is nervous, the Generative AI General Manager will respond flexibly. Finally, the server generates a feedback report and provides it to the user, allowing them to prepare for the next actual review.

[0874] The above is a detailed description of an embodiment of the present invention. Based on this embodiment, users can perform multiple simulations to improve the quality of their reviews. The emotion engine provides feedback tailored to the user's individual emotional state, which can lead to more effective learning and improvement.

[0875] The processing flow will be explained below.

[0876] Program processing steps

[0877] Step 1: Data collection

[0878] The server collects past audio and video data from internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in the form of sound waves, and the video data is stored in the server's storage in the form of video.

[0879] Step 2: Initial saving of data

[0880] The server temporarily stores the collected audio and video data in storage, where it is used in subsequent processing steps.

[0881] Step 3: Process the audio data

[0882] The server uses a speech recognition engine to convert the voice data into text data, which allows the spoken content to be obtained in text format. It also identifies the speaker and tags the timing of the speech.

[0883] Step 4: Video data processing

[0884] The server uses a video analysis engine to extract facial expressions and movement characteristics from the video data, allowing the speaker's emotions and facial expressions to be analyzed.

[0885] Step 5: Integrate the data

[0886] The server combines the text data obtained from the audio data with the features obtained from the video data, creating a single dataset that includes who said what, and what emotions and facial expressions they showed.

[0887] Step 6: Training the generative AI model

[0888] The server uses the combined dataset to train a generative AI model, using machine learning techniques to create a model capable of reproducing the general manager's tone, pronunciation, and emotional patterns.

[0889] Step 7: Saving the generated AI model

[0890] The server stores the trained generative AI model in storage, which is then used in subsequent simulation phases.

[0891] Step 8: Request a simulation

[0892] A user accesses the simulation tool from his / her own terminal and sends a review simulation request to the server.

[0893] Step 9: Upload your materials

[0894] The user uploads the document (e.g., proposal) they wish to review and simulate. The uploaded document is transferred to the server.

[0895] Step 10: Initializing the generative AI model

[0896] The server analyzes the uploaded materials and prepares the simulation. The head of the AI ​​generation department loads the information necessary for reviewing the user's materials.

[0897] Step 11: Activating the Emotion Engine

[0898] The server activates the emotion engine and acquires video data from the user's device in real time. The emotion engine analyzes this data and identifies the user's emotional state.

[0899] Step 12: Conduct a review simulation

[0900] The server launches the Generative AI Director, who begins reviewing the materials based on their content. The Generative AI Director analyzes the user's materials in detail and provides appropriate comments and questions in real time. The user receives this feedback in real time on their own device. The emotion engine monitors the user's emotional state, and if it detects tension or stress, it adjusts the tone of the Generative AI Director's feedback.

[0901] Step 13: Save the review results

[0902] The server stores the review results of the generated AI chief as a log, which records all review content and emotional state.

[0903] Step 14: Generate a feedback report

[0904] The server analyzes the review results and emotional state and generates a comprehensive feedback report for the user, including suggestions for improvement and recommendations.

[0905] Step 15: Provide feedback

[0906] The server will provide the generated feedback report to the user, who can then review the feedback on their own device and make any necessary improvements.

[0907] The above are the specific processing steps of this system.

[0908] Example 2

[0909] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0910] Conventional internal review meeting systems do not fully utilize past review results, limiting opportunities for users to practice for actual review meetings or receive effective feedback. Furthermore, since feedback does not take into account the user's emotional state, it is difficult to realize true improvements. To solve these issues, a system that can utilize past data and provide feedback that matches the user's emotional state is needed.

[0911] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0912] In this invention, the server includes a means for collecting and processing audio and video data, which are past data from internal review meetings; a means for training a generative AI model using machine learning technology based on the processed audio and video data; and a means including an emotion engine for recognizing the user's emotional state and adjusting the feedback content. This enables simulations based on data from past review meetings, allowing users to receive more effective feedback when preparing for an actual review. Furthermore, the emotion engine enables appropriate responses that take into account the user's tension and anxiety, allowing users to perform review simulations more effectively.

[0913] "Audio data" refers to data in which audio from meetings, reviews, etc. is recorded in digital format.

[0914] "Video data" refers to data that has been recorded in digital format, such as footage of meetings or reviews.

[0915] "Machine learning technology" is a technology that allows a computer to extract patterns from given data and learn automatically.

[0916] A "generative AI model" is an artificial intelligence model trained using machine learning technology to reproduce the tone and emotional patterns of a specific executive.

[0917] An "emotion engine" is a system that recognizes a user's emotional state and adjusts the feedback content based on that state.

[0918] "Review simulation" is a process in which documents submitted by users are virtually reviewed using a generative AI model.

[0919] "Feedback" refers to evaluations and advice provided to users based on the results of the review simulation.

[0920] "Materials" are documents such as proposals and project plans that users submit for review purposes.

[0921] "Server" is the central computing system that performs data collection, analysis, model training and simulation.

[0922] MODE FOR CARRYING OUT THE INVENTION

[0923] The present invention relates to a system that creates a generative AI model using past data from internal review meetings and uses the model to virtually simulate a review. Furthermore, by adding an emotion engine that recognizes the user's emotions, the system aims to provide feedback based on the user's emotional state. The following describes in detail an embodiment of this system.

[0924] Data Collection Phase

[0925] The server collects past audio and video data from internal review meetings. This data is obtained from recorders and cameras, or files exported from existing conference systems are used. For example, audio data is collected in .wav file format, and video data is collected in .mp4 file format. The collected audio and video data is stored in the server's storage.

[0926] Data analysis and processing

[0927] The server processes the collected voice data using a speech recognition engine such as the Google Speech-to-Text API and converts it into text data. This process also identifies the speaker and tags the speech over time. The server then processes the video data using a video analysis engine such as the Microsoft Azure Face API to analyze facial and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is then integrated with the text data and saved as a single unified dataset.

[0928] Training a generative AI model

[0929] The server uses the unified dataset to train the generative AI model. This training uses machine learning techniques such as TensorFlow. The dataset is used to train the model to reproduce the tone, pronunciation, and emotional patterns of a specific executive. Once trained, the generative AI model can faithfully simulate the executive's speech and facial expressions.

[0930] Simulation Phase

[0931] A user sends a request for an internal review simulation to the server from their device. The user uploads the document they wish to simulate (e.g., a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and a virtual executive reviews it using the generative AI model. The server then uses an emotion engine to analyze the user's video data and identify their emotional state in real time.

[0932] Conducting a review

[0933] The generated AI executive reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the interaction in an interactive format. The generative AI model reproduces the executive's tone and emotional patterns, providing an atmosphere similar to that of a conversation with an actual executive. Furthermore, the emotion engine recognizes the user's emotional state and adjusts the content of the generated AI executive's feedback. For example, if the user is nervous, the generated AI executive can provide feedback in a softer tone.

[0934] Providing Feedback

[0935] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[0936] Specific examples

[0937] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches a generating AI executive. The generating AI executive reviews the proposal's contents in detail and provides comments and questions to the user in real time. Through dialogue with the virtual executive, the user understands which parts of their document need improvement. In addition, an emotion engine grasps the user's emotional state in real time, and if the user is nervous, the generating AI executive will respond flexibly. Finally, the server generates a feedback report and provides it to the user.

[0938] As a concrete example, let's input the following prompt to the generative AI model:

[0939] "Present your proposal and get feedback from virtual executives."

[0940] The above is a detailed description of an embodiment of the present invention. Based on this embodiment, users can perform multiple simulations to improve the quality of their reviews. The emotion engine provides feedback tailored to the user's individual emotional state, which can lead to more effective learning and improvement.

[0941] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0942] Step 1: Data collection

[0943] The server collects past audio and video data from internal review meetings. Input includes data captured from internal recorders and cameras, or data files exported from existing conference systems. Data processing involves saving the audio data (e.g., .wav files) and video data (e.g., .mp4 files) in a specified directory. The output is the audio and video data saved in storage.

[0944] Step 2: Analyzing the audio data

[0945] The voice data collected by the server is processed by a voice recognition engine (e.g., Google Speech-to-Text API) and converted into text data. The input includes voice data. Data calculations include the process of performing voice recognition and converting the data into text data with time tags and speaker information attached. Specific operations include making an API call and saving the results as a text file. The output is text data with time tags attached.

[0946] Step 3: Analyzing the video data

[0947] The video data collected by the server is processed using a video analysis engine (e.g., Microsoft Azure Face API) to extract facial and movement characteristics. The input includes video data. Data calculations include the process of analyzing the video and extracting facial expression data (e.g., smiling, sadness, anger, etc.) and movement data (e.g., hand movements, eye focus, etc.). Specific operations include making an API call and saving the results in JSON format. Analyzed facial expression and movement data are obtained as output.

[0948] Step 4: Integrate the data

[0949] The server integrates the analyzed audio and video data and stores it as a single unified dataset. The input includes time-tagged text data and analyzed facial and movement data. The data processing involves merging the audio and video data based on the time tags to create a single unified dataset. The specific operation is to store it as a unified dataset in a database. The unified dataset is obtained as the output.

[0950] Step 5: Training the generative AI model

[0951] The server uses a unified dataset to train the generative AI model. The input includes the unified dataset. Data processing involves using a machine learning framework (e.g., TensorFlow) to learn the tone, pronunciation, and emotional patterns of a specific executive. Specifically, the dataset is fed into a neural network model, training is performed, and the optimal model parameters are saved. The output is a trained generative AI model.

[0952] Step 6: Receiving a simulation request

[0953] The user sends a request for an internal review simulation to the server from their own device. The input includes the user's login information and the desired materials for the simulation. Specific operations include uploading the materials through the user interface and sending the request to the server. The output shows the status of the request being sent to the server.

[0954] Step 7: Analyze the data and perform simulations

[0955] The server analyzes the materials received from the user, and the virtual executive reviews them using a generative AI model. Inputs include the user's materials and the generative AI model. Data calculations include analyzing the content of the materials using a text analysis engine (e.g., SpaCy or NLTK) and extracting important points and keywords. This data is passed to the generative AI model, and the virtual executive begins the review. Specifically, the model's output is fed back to the user's device in real time. Real-time comments and questions are provided to the user as output.

[0956] Step 8: Run the Emotion Engine

[0957] The server uses an emotion engine to analyze the user's video data and identify their emotional state in real time. The input includes the user's video data. The data calculation process involves analyzing the data using an emotion recognition algorithm and extracting the user's emotional state (e.g., tension, surprise, relief). Specific operations include calling the emotion engine's API and adjusting the feedback content based on the analysis results. The output is feedback corresponding to the user's emotional state.

[0958] Step 9: Provide feedback

[0959] The generated AI manager evaluates the user's materials and provides appropriate comments and feedback in real time. Specifically, it responds flexibly to the user's tension or anxiety and adjusts the content of the feedback. The output is the comments and advice provided as feedback that is displayed to the user.

[0960] Step 10: Generate a feedback report

[0961] The server analyzes the feedback generated after the simulation has finished and provides a comprehensive report to the user. The input includes a record of all interactions during the simulation. Data processing involves analyzing the log data and summarizing areas for improvement, positives, and specific action items. Specific operations include applying a report generation algorithm to create a report in a format that the user can download (e.g., PDF). The output is a comprehensive feedback report provided to the user.

[0962] The above is a detailed description of the specific processing steps of the system. By following this procedure, users can effectively conduct a review simulation and prepare for an actual review meeting.

[0963] (Application example 2)

[0964] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0965] In the design and development of conventional autonomous vehicles, there is a problem that the review simulations performed by engineers and designers often deviate significantly from the feedback given in actual review meetings. In addition, there is a lack of a means to provide feedback that takes into account the user's emotional state, which hinders more effective improvements.

[0966] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes: means for collecting and processing audio data and video data, which are past data of internal review meetings; means for training a generative AI model based on the processed audio data and video data; means for simulating a review of materials submitted by a user using the generative AI model; means for adjusting feedback content, including an emotion engine that analyzes the user's emotional state; and means for providing the simulation results to the user as feedback. This enables a virtual review simulation using the generative AI model, and can provide feedback based on the user's emotional state, enabling effective improvements in an environment that more closely resembles an actual meeting.

[0967] "Audio data" refers to digital data that records the sounds generated during a meeting, conversation, or the like.

[0968] "Video data" is digital data that records visual information generated during meetings, conversations, etc.

[0969] A "generative AI model" is an artificial intelligence model that is trained using machine learning techniques based on past data and can simulate judgments and actions for specific tasks.

[0970] "Review simulation" is the process of virtually evaluating and examining the materials submitted by the user and providing the results as feedback.

[0971] "Feedback" refers to evaluations and advice obtained as a result of the simulation and is provided to the user.

[0972] An "emotion engine" is a technology that analyzes a user's emotional state from facial expressions, tone of voice, etc.

[0973] "Server" means a computer system for collecting, analyzing data, and training generative AI models.

[0974] "Training" is the process by which a generative AI model learns using past data to improve its performance on a specific task.

[0975] A "user" is a person who uses the system to undergo a review simulation.

[0976] The following describes an embodiment of the present invention.

[0977] System Configuration

[0978] The system of this invention includes a server, a user terminal, an emotion engine, and a generative AI model. The server collects and analyzes data, trains the generative AI model, and sends the generated feedback to the user terminal. The user terminal provides an interface for the user to review the simulation. The emotion engine analyzes the user's video data and detects their emotional state in real time.

[0979] Program processing

[0980] The server collects audio and video data from past internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in sound wave format, and the video data is stored in video format.

[0981] The server then processes the collected voice data using a speech recognition engine (e.g., Google Speech-to-Text API) and converts it into text data. During this process, it also identifies the speaker and tags the speech over time. The video data is then processed using a video analysis engine (e.g., OpenCV) to analyze facial and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is then integrated with the text data and saved as a single unified dataset.

[0982] Using the unified dataset, the server trains a generative AI model (such as GPT-3 or BERT), which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[0983] To conduct an internal review simulation, a user sends a simulation request to the server from their own device. The user uploads the materials they wish to simulate (e.g., design documents or prototypes). These materials are transferred to the server and become the subject of review by the generative AI model. The server analyzes the materials, and the virtual head of the department conducts the review using the generative AI model.

[0984] Furthermore, the server uses an emotion engine to analyze the user's video data and identify their emotional state in real time. The Generative AI Headquarters can then adjust the feedback content based on the user's emotional state, enabling more flexible responses. For example, if the user is nervous, the server will provide feedback in a softer tone.

[0985] Specific examples

[0986] An engineer has prepared a design document for a new autonomous vehicle and wants to have it reviewed. The engineer accesses the "Automotive Development Simulation Trainer" from his or her PC and logs into the system. Next, he or she uploads the design document file to the server. The server receives the file and launches the Generative AI Headquarters. The Generative AI Headquarters reviews the contents of the design document in detail and provides comments and questions to the engineer in real time. Through dialogue with the virtual headquarters, the engineer understands which parts of his or her document need improvement. In addition, the emotion engine grasps the engineer's emotional state in real time, and if the engineer seems nervous, the Generative AI Headquarters will respond flexibly. Finally, the server generates a feedback report and provides it to the engineer, allowing him or her to prepare for the next actual review.

[0987] Prompt Sentence Examples

[0988] "Please upload the design document below. We will start the simulation and flexibly adjust the feedback based on the user's emotional state."

[0989] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0990] Step 1:

[0991] The server collects audio and video data from past internal review meetings. The audio and video data is acquired from recorders and cameras, or files exported from existing conference systems are used. This data is saved in the server's storage. The input is audio and video data, and the output is the saved raw data.

[0992] Step 2:

[0993] The server processes the collected voice data using a voice recognition engine (e.g., Google Speech-to-Text API) and converts it into text data. This process also identifies the speaker and tags the speech with time. The input is voice data, and the output is text data with speaker identification and time tags.

[0994] Step 3:

[0995] The server processes the video data using a video analysis engine (e.g., OpenCV) to analyze facial expressions and movement features. This makes it possible to identify the speaker's emotions and changes in facial expressions. The input is video data, and the output is data containing emotional states and facial expression features.

[0996] Step 4:

[0997] The server integrates the analyzed audio and video data and stores them as a unified dataset. The input is text data and facial expression data, and the output is the integrated dataset.

[0998] Step 5:

[0999] The server uses the unified dataset to train a generative AI model (e.g., GPT-3 or BERT). This training uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief executive. The input is the unified dataset, and the output is the trained generative AI model.

[1000] Step 6:

[1001] A user sends a request to the server from their own device to run an internal review simulation. The user uploads materials such as design documents and prototypes. The input is the user request and the uploaded materials, and the output is the transfer of the materials to the server.

[1002] Step 7:

[1003] The server analyzes the uploaded materials, and the virtual director reviews them using a generative AI model. The input is the uploaded materials, and the output is the review comments and feedback of the generative AI director.

[1004] Step 8:

[1005] The server uses an emotion engine to analyze the user's video data and identify the user's emotional state in real time. The input is the user's video data, and the output is the user's emotional state.

[1006] Step 9:

[1007] The Generative AI Director adjusts the feedback content based on the user's emotional state, enabling flexible responses. For example, if the user is nervous, the feedback will be given in a softer tone. The input is the user's emotional state and the Generative AI Director's review comments, and the output is the adjusted feedback.

[1008] Step 10:

[1009] The server generates a feedback report and provides it to the user. The report includes areas for improvement, good points, and specific advice. The input is the tailored feedback, and the output is the feedback report.

[1010] Through these steps, the system enables users to make effective improvements in an environment similar to an actual review meeting.

[1011] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

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

[1013] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1014] [Fourth embodiment]

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

[1016] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1017] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1018] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

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

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

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

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

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

[1024] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

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

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

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

[1028] MODE FOR CARRYING OUT THE INVENTION

[1029] The present invention provides a system that allows users to virtually perform review simulations using a generative AI model by using past data from internal review meetings. The following describes an embodiment of this system.

[1030] Data Collection Phase

[1031] The server collects past data (audio and video data) from internal review meetings. This data can be acquired using audio and video recording devices such as microphones and cameras. It is also possible to export data from existing conference systems. The collected audio data is stored in the server's storage in sound wave format, and the video data is stored in video format.

[1032] Data analysis and processing

[1033] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it also identifies the speaker and tags the speech content temporally. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is integrated with the text data and saved as a single unified data set.

[1034] Training a generative AI model

[1035] The server uses the unified dataset to train a generative AI model, which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[1036] Simulation Phase

[1037] The user sends a request to the server from their device to conduct an internal review simulation. The user uploads the document they wish to simulate (for example, a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual head of the department reviews it using the generative AI model.

[1038] Conducting a review

[1039] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the situation in an interactive format. The generative AI model reproduces tone and emotional patterns, providing an atmosphere similar to that of a conversation with a real director.

[1040] Providing Feedback

[1041] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[1042] Specific examples

[1043] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI Director. The Generative AI Director reviews the proposal in detail and provides comments and questions to the user in real time. Through dialogue with the virtual Director, the user understands which parts of their materials need improvement. Finally, the server generates a feedback report and provides it to the user, allowing them to prepare for the next actual review.

[1044] The above is a specific description of the embodiment of the present invention, which allows the user to perform simulations multiple times and improve the quality of the review.

[1045] The processing flow will be explained below.

[1046] Program processing steps

[1047] Step 1: Data collection

[1048] The server collects historical audio and video data from internal review meetings, either from recorders and cameras or using files exported from existing conferencing systems.

[1049] Step 2: Initial saving of data

[1050] The server temporarily stores the collected audio and video data in storage, where it is used in subsequent processing steps.

[1051] Step 3: Process the audio data

[1052] The server uses a speech recognition engine to convert the voice data into text data, thereby obtaining the spoken content in text format.

[1053] Step 4: Video data processing

[1054] The server uses a video analysis engine to extract facial expressions and movement characteristics from the video data, and through this process, emotion and movement patterns are analyzed.

[1055] Step 5: Integrate the data

[1056] The server combines the text data obtained from the audio data with the features obtained from the video data, creating a single dataset that includes who said what, and what emotions and facial expressions they showed.

[1057] Step 6: Training the generative AI model

[1058] The server uses the combined dataset to train a generative AI model, using machine learning techniques to create a model capable of reproducing the general manager's tone, pronunciation, and emotional patterns.

[1059] Step 7: Saving the generated AI model

[1060] The server stores the trained generative AI model in storage, which is then used in subsequent simulation phases.

[1061] Step 8: Request a simulation

[1062] A user accesses the simulation tool from his / her own terminal and sends a review simulation request to the server.

[1063] Step 9: Upload your materials

[1064] The user uploads the document (e.g., proposal) they wish to review and simulate. The uploaded document is transferred to the server.

[1065] Step 10: Initializing the generative AI model

[1066] The server analyzes the uploaded materials and prepares the simulation. The head of the AI ​​generation department loads the information necessary for reviewing the user's materials.

[1067] Step 11: Conduct a review simulation

[1068] The server launches the Generative AI Headquarters and begins reviewing the content of the materials. The Generative AI Headquarters analyzes the user's materials in detail and provides appropriate comments and questions. The user receives this feedback in real time on their device.

[1069] Step 12: Save the review results

[1070] The server saves the review results of the generated AI chief as a log, which records all review content.

[1071] Step 13: Generate a feedback report

[1072] The server analyzes the reviews and generates a comprehensive feedback report for the user, including suggestions for improvement.

[1073] Step 14: Provide feedback

[1074] The server will provide the generated feedback report to the user, who can then review the feedback on their own device and make any necessary improvements.

[1075] The above are the specific processing steps of this system.

[1076] Example 1

[1077] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1078] Current internal review meetings present difficulties in efficiently reviewing materials, particularly when evaluating new proposals and project plans, as they require time and effort. Furthermore, the quality of the review is heavily dependent on subjective factors, making it difficult to ensure consistency and objectivity. This leaves employees preparing for the review with few opportunities to simulate the actual opinions and comments of the general manager in advance. To address these issues, there is a need to develop a system that efficiently utilizes generative AI models using data from past review meetings, allowing users to effectively perform review simulations.

[1079] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1080] In this invention, the server includes means for collecting and processing audio and video data, means for training a generative AI model based on the processed audio and video data, means for using the generative AI model to review materials submitted by users, means for providing interactive interaction with the users during the review, and means for analyzing the results of the simulation and generating and providing a comprehensive feedback report, thereby enabling users to virtually simulate a review and understand and prepare the review content in advance.

[1081] "Audio data" is a recording format of audio acquired using an audio input device such as a microphone.

[1082] "Video data" refers to the recording format of video captured using a video capture device such as a camera.

[1083] A "generative AI model" is an artificial intelligence model trained using machine learning techniques to reproduce specific speech patterns and emotions.

[1084] "User" refers to a person who uses this system to conduct a document review simulation.

[1085] "Materials" refers to documents such as proposals and project plans uploaded by users and subject to review.

[1086] "Review" refers to the process of feedback and comments provided by the generative AI model.

[1087] "Dialogue format" refers to a format in which the user and the generative AI model interact through dialogue.

[1088] "Feedback" refers to the evaluation and advice that the generative AI model provides to the user on their material.

[1089] A "feedback report" is a document containing a comprehensive evaluation and advice generated based on the simulation results.

[1090] "Interaction" refers to the interaction or communication that takes place between a user and a system.

[1091] "Analysis" refers to the process of processing collected audio and video data and extracting useful information.

[1092] The present invention provides a system that allows users to virtually perform review simulations using a generative AI model based on past data from internal review meetings. The following describes an embodiment of this system.

[1093] Data Collection Phase

[1094] The server collects audio and video data from past internal review meetings. This data is collected using audio and video recording devices such as microphones and cameras, or an existing conference system. The collected audio data is saved in sound wave format, and the video data is saved in video format in the server's storage. This allows the contents of review meetings to be recorded in detail and used for later analysis.

[1095] Data analysis and processing

[1096] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it identifies the speaker and temporally tags the speech content. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. In particular, it is possible to identify the speaker's emotions and changes in facial expression. The analyzed voice data and video data are integrated and saved as a single unified data set.

[1097] Training a generative AI model

[1098] The server uses this unified dataset to train a generative AI model using machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[1099] Simulation Phase

[1100] The user sends a request to the server from their device to conduct a review simulation. The user uploads the document they wish to simulate (for example, a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual head of the headquarters reviews it using the generative AI model.

[1101] Conducting reviews and providing feedback

[1102] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to interactively simulate the situation. The generative AI model reproduces tone and emotional patterns to create an atmosphere similar to a conversation with a real director. After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report. This report includes points for improvement, good points, and specific advice. The user can use this as a reference to prepare for the actual review meeting.

[1103] Specific examples

[1104] An employee has prepared a proposal for a new project and wants to have it reviewed. The user accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI Director. The Generative AI Director reviews the proposal's contents in detail and provides comments and questions to the user in real time. Through dialogue with the virtual Director, the user understands which parts of their materials need improvement. Finally, the server generates a feedback report and provides it to the user.

[1105] An example of a prompt is as follows:

[1106] "Please evaluate the following proposal. The title is 'Proposal for a Next-Generation Marketing Strategy.' The contents of the proposal are as follows."

[1107] The above is a specific description of the embodiment of the present invention. Based on this embodiment, the user can perform simulations multiple times to improve the quality of the review.

[1108] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1109] Step 1: Data collection

[1110] The server collects audio and video data from internal review meetings. Specifically, it records the audio and video of the meetings using a microphone and camera. Data exported from existing conference systems can also be used. The input is audio and video data, and the output is audio files in sound wave format and video files in video format. These are stored in the server's storage.

[1111] Step 2: Analyzing the audio data

[1112] The server analyzes the collected voice data using a speech recognition engine. This engine receives an audio file as input and converts the speech into text data. Specifically, it identifies the speaker and adds time tags to the speech. The output is text data that includes speaker information and time tags.

[1113] Step 3: Analyzing the video data

[1114] The server analyzes the collected video data using a video analysis engine. This engine receives video files as input and extracts facial and movement features. Specifically, it identifies changes in the speaker's emotions and adds time tags to the analysis results of facial expressions and movements. The output is the analyzed video data, which includes emotion and movement tags.

[1115] Step 4: Generate a unified dataset

[1116] The server integrates the text data of the audio data with the results of video analysis to generate a unified dataset. The input is the text data and the analyzed video data, and the output is a unified dataset. This dataset includes information such as audio content, speaker identification, time tags, facial expression changes, and motion analysis.

[1117] Step 5: Training the generative AI model

[1118] The server trains the generative AI model using the unified dataset. The input is the unified dataset, and the output is the trained generative AI model. Specifically, it uses a machine learning algorithm to learn the tone, pronunciation, and emotional patterns of the director and generates a trained model.

[1119] Step 6: Processing the User Request

[1120] A user sends a review simulation request from their own device to the server. The input is the user's desired document (proposal or project plan), and the output is the document uploaded to the server.

[1121] Step 7: Analyze and review the materials

[1122] The server analyzes the materials received from the user using a generative AI model and conducts a review. The input is the uploaded materials, and the output is comments and questions from the head of the generative AI department. The generative AI model analyzes the contents of the materials in detail and generates feedback in the form of a statement from the head of the department.

[1123] Step 8: Provide feedback

[1124] The server provides the generated feedback to the user. Specifically, the head of the Generating AI displays comments and questions on the user's device in real time. The input is the generated feedback, and the output is the feedback displayed on the user's device. The user receives this in an interactive format.

[1125] Step 9: Generate a comprehensive report

[1126] After the simulation is complete, the server analyzes the generated feedback and generates a comprehensive report. The input is the feedback data, and the output is a comprehensive feedback report. This report includes points for improvement, good points, and specific advice. Users can use this as a reference to prepare for the actual review meeting.

[1127] (Application example 1)

[1128] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1129] When reviewing operation plans or maintenance reports for industrial equipment, it is difficult to efficiently simulate the opinions of personnel with specialized knowledge and experience. Furthermore, conventional review methods require physical meetings to apply for reviews and receive feedback, which places significant constraints on time and location. This makes it difficult to conduct reviews quickly and accurately.

[1130] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1131] In this invention, the server includes means for collecting and processing audio and video data that are past data from internal review meetings, means for training a generative AI model based on the processed audio and video data, means for simulating a review of materials submitted by a user using the generative AI model, means for providing the simulation results to the user as feedback, means for reviewing operation plans and maintenance reports for industrial equipment using the generative AI model, and means for requesting reviews and uploading materials from a smartphone, thereby enabling fast and accurate reviews without relying on physical meetings.

[1132] A "generative AI model" is an artificial intelligence that is trained based on past data from internal review meetings and can simulate specific speech content and emotional patterns.

[1133] "Audio data" refers to data that has been recorded and saved in digital format, including speech and sound during a meeting.

[1134] "Video data" refers to data that has been recorded and saved in digital format, including the state of a meeting, the facial expressions and movements of speakers, etc.

[1135] "Industrial equipment" refers to machinery and equipment used in factories to process and assemble products, such as robots and production line equipment.

[1136] An "operation plan" is a document that defines the operating procedures and timing of industrial equipment and production lines.

[1137] A "maintenance report" is a report that records the maintenance and repair status of industrial equipment.

[1138] "Feedback" refers to providing evaluation, suggestions for improvement, and advice regarding materials submitted by users.

[1139] A "smartphone" is a portable information terminal equipped with Internet connectivity and application execution capabilities.

[1140] To implement this invention, it is desirable to build the following system, which trains a generative AI model using past data from internal review meetings and performs a virtual review of documents submitted by users.

[1141] System configuration

[1142] The system consists of the following elements:

[1143] 1. Server:

[1144] The server collects, processes, and stores audio and video data, including a speech recognition engine and a video analysis engine. Specifically, it uses the speech_recognition library to convert audio data into text data and the moviepy library to analyze video.

[1145] 2. Generative AI Model:

[1146] The generative AI model reviews user-submitted materials based on the trained data. The generative AI model implements the GPT-3 model using the transformers library.

[1147] 3. User Device:

[1148] Users can access the system from their smartphone or computer and upload operation plans and maintenance reports.

[1149] 4. Building a network:

[1150] A network is required to exchange data between the user's device and the server. To ensure a secure connection, the HTTPS protocol is used.

[1151] Data collection and processing

[1152] The server collects audio and video data from past internal review meetings. The audio data is converted to text using the speech_recognition library. The video data is analyzed using the moviepy library to analyze the speaker's facial expressions and behavior.

[1153] Training a generative AI model

[1154] Based on the audio and video data, a generative AI model was trained using the transformers library to build a GPT-3 model that reproduced the Director's tone and emotional patterns.

[1155] System operation example

[1156] Users access the system from their smartphones and upload operation plans and maintenance reports for industrial equipment. The server receives the documents and performs a virtual review using a generative AI model. Feedback is displayed on the user's device in real time.

[1157] Specific examples

[1158] For example, suppose a user wants to review the following motion plan:

[1159] Please review the following plans:

[1160] The plan states:

[1161] 1. Basic robot movements

[1162] 2. Maintenance Procedures

[1163] 3. Safety Measures

[1164] Please point out the positive and negative aspects of this plan.

[1165] The server inputs this prompt into a generative AI model to generate specific feedback, allowing the user to understand how to improve their motion plan.

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

[1167] Step 1:

[1168] The server collects audio and video data from past internal review meetings. Audio data is recorded via a microphone, and video data is recorded via a camera. These data are stored in the server's storage. The input is from the microphone and camera feed, and the output is from audio data files and video data files.

[1169] Step 2:

[1170] The server processes the collected voice data using a voice recognition engine and converts it into text data. This process uses the speech_recognition library. The server inputs the voice file and outputs text data. Specifically, the voice data is passed to the recognition engine, which analyzes the speech and converts it into text format along with timestamps and speaker information.

[1171] Step 3:

[1172] The server processes the video data using a video analysis engine to analyze the speaker's facial expressions and behavioral features. This process uses the moviepy library. It inputs the video file and outputs the analysis results. Specifically, it extracts facial expressions and behaviors from each video frame, identifies the speaker's emotional state and gestures, and digitizes them.

[1173] Step 4:

[1174] The server integrates the audio and video data to generate a unified dataset. It also concatenates the audio-text data and video analysis data and stores them as a single dataset. The input is the audio-text and video analysis results, and the output is a unified dataset.

[1175] Step 5:

[1176] The server trains a generative AI model using a unified dataset. It uses the Transformers library and the GPT-3 model for training. It inputs the unified dataset and outputs a trained generative AI model. Specifically, the dataset is passed to the model to learn specific patterns and features.

[1177] Step 6:

[1178] Users access the system from their own devices and upload the materials they wish to review. The materials are sent as input to the server. Users can then select and upload materials using their smartphones or PCs.

[1179] Step 7:

[1180] The server analyzes the materials uploaded by the user and reviews them using a generative AI model. The server analyzes the document file as input and passes it to the generative AI model as a prompt. Specifically, it analyzes the content of the document, forms an appropriate prompt, and gives it to the generative AI model.

[1181] Step 8:

[1182] The generative AI model reviews materials based on prompts and generates feedback. The prompts are input and the review results are output. Specifically, the model interprets the prompts and generates a textual evaluation of the materials and suggestions for improvement.

[1183] Step 9:

[1184] The server provides the generated feedback to the user in real time by sending the feedback in text format to the user's device. Specifically, the output of the generative AI model is displayed on the user interface so that the user can check the feedback.

[1185] Step 10:

[1186] After the simulation is complete, the server generates a comprehensive report and provides it to the user. The server then aggregates the feedback results and creates a final report. Specifically, it compiles all the feedback, generates a report with specific improvements and advice, and sends it to the user.

[1187] Through the above steps, users can receive a quick and accurate review without relying on a physical meeting.

[1188] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1189] MODE FOR CARRYING OUT THE INVENTION

[1190] The present invention relates to a system that creates a generative AI model using past data from internal review meetings and uses the model to virtually simulate a review. Furthermore, by adding an emotion engine that recognizes the user's emotions, the system aims to provide feedback based on the user's emotional state. The following describes in detail an embodiment of this system.

[1191] Data Collection Phase

[1192] The server collects past audio and video data from internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in the form of sound waves, and the video data is stored in the server's storage in the form of video.

[1193] Data analysis and processing

[1194] The server processes the collected voice data using a speech recognition engine and converts it into text data. During this process, it also identifies the speaker and tags the speech content temporally. Next, the server processes the video data using a video analysis engine to analyze facial expressions and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is integrated with the text data and saved as a single unified data set.

[1195] Training a generative AI model

[1196] The server uses the unified dataset to train a generative AI model, which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[1197] Simulation Phase

[1198] The user sends a request to the server from their device to conduct an internal review simulation. The user uploads the document they wish to simulate (e.g., a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and the virtual general manager uses the generative AI model to conduct the review. Furthermore, the server uses an emotion engine to analyze the user's video data and identify their emotional state in real time.

[1199] Conducting a review

[1200] The generative AI director reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the situation interactively. The generative AI model reproduces tone and emotional patterns, providing an atmosphere similar to that of a conversation with a real director. Furthermore, the emotion engine recognizes the user's emotional state and adjusts the content of the generative AI director's feedback. For example, if the user is nervous, the generative AI director can provide feedback in a softer tone.

[1201] Providing Feedback

[1202] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[1203] Specific examples

[1204] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches the Generative AI General Manager. The Generative AI General Manager reviews the proposal in detail and provides comments and questions to the user in real time. Through dialogue with the virtual General Manager, the user understands which parts of their document need improvement. In addition, an emotion engine grasps the user's emotional state in real time, and if the user is nervous, the Generative AI General Manager will respond flexibly. Finally, the server generates a feedback report and provides it to the user, allowing them to prepare for the next actual review.

[1205] The above is a detailed description of an embodiment of the present invention. Based on this embodiment, users can perform multiple simulations to improve the quality of their reviews. The emotion engine provides feedback tailored to the user's individual emotional state, which can lead to more effective learning and improvement.

[1206] The processing flow will be explained below.

[1207] Program processing steps

[1208] Step 1: Data collection

[1209] The server collects past audio and video data from internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in the form of sound waves, and the video data is stored in the server's storage in the form of video.

[1210] Step 2: Initial saving of data

[1211] The server temporarily stores the collected audio and video data in storage, where it is used in subsequent processing steps.

[1212] Step 3: Process the audio data

[1213] The server uses a speech recognition engine to convert the voice data into text data, which allows the spoken content to be obtained in text format. It also identifies the speaker and tags the timing of the speech.

[1214] Step 4: Video data processing

[1215] The server uses a video analysis engine to extract facial expressions and movement characteristics from the video data, allowing the speaker's emotions and facial expressions to be analyzed.

[1216] Step 5: Integrate the data

[1217] The server combines the text data obtained from the audio data with the features obtained from the video data, creating a single dataset that includes who said what, and what emotions and facial expressions they showed.

[1218] Step 6: Training the generative AI model

[1219] The server uses the combined dataset to train a generative AI model, using machine learning techniques to create a model capable of reproducing the general manager's tone, pronunciation, and emotional patterns.

[1220] Step 7: Saving the generated AI model

[1221] The server stores the trained generative AI model in storage, which is then used in subsequent simulation phases.

[1222] Step 8: Request a simulation

[1223] A user accesses the simulation tool from his / her own terminal and sends a review simulation request to the server.

[1224] Step 9: Upload your materials

[1225] The user uploads the document (e.g., proposal) they wish to review and simulate. The uploaded document is transferred to the server.

[1226] Step 10: Initializing the generative AI model

[1227] The server analyzes the uploaded materials and prepares the simulation. The head of the AI ​​generation department loads the information necessary for reviewing the user's materials.

[1228] Step 11: Activating the Emotion Engine

[1229] The server activates the emotion engine and acquires video data from the user's device in real time. The emotion engine analyzes this data and identifies the user's emotional state.

[1230] Step 12: Conduct a review simulation

[1231] The server launches the Generative AI Director, who begins reviewing the materials based on their content. The Generative AI Director analyzes the user's materials in detail and provides appropriate comments and questions in real time. The user receives this feedback in real time on their own device. The emotion engine monitors the user's emotional state, and if it detects tension or stress, it adjusts the tone of the Generative AI Director's feedback.

[1232] Step 13: Save the review results

[1233] The server stores the review results of the generated AI chief as a log, which records all review content and emotional state.

[1234] Step 14: Generate a feedback report

[1235] The server analyzes the review results and emotional state and generates a comprehensive feedback report for the user, including suggestions for improvement and recommendations.

[1236] Step 15: Provide feedback

[1237] The server will provide the generated feedback report to the user, who can then review the feedback on their own device and make any necessary improvements.

[1238] The above are the specific processing steps of this system.

[1239] Example 2

[1240] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1241] Conventional internal review meeting systems do not fully utilize past review results, limiting opportunities for users to practice for actual review meetings or receive effective feedback. Furthermore, since feedback does not take into account the user's emotional state, it is difficult to realize true improvements. To solve these issues, a system that can utilize past data and provide feedback that matches the user's emotional state is needed.

[1242] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1243] In this invention, the server includes a means for collecting and processing audio and video data, which are past data from internal review meetings; a means for training a generative AI model using machine learning technology based on the processed audio and video data; and a means including an emotion engine for recognizing the user's emotional state and adjusting the feedback content. This enables simulations based on data from past review meetings, allowing users to receive more effective feedback when preparing for an actual review. Furthermore, the emotion engine enables appropriate responses that take into account the user's tension and anxiety, allowing users to perform review simulations more effectively.

[1244] "Audio data" refers to data in which audio from meetings, reviews, etc. is recorded in digital format.

[1245] "Video data" refers to data that has been recorded in digital format, such as footage of meetings or reviews.

[1246] "Machine learning technology" is a technology that allows a computer to extract patterns from given data and learn automatically.

[1247] A "generative AI model" is an artificial intelligence model trained using machine learning technology to reproduce the tone and emotional patterns of a specific executive.

[1248] An "emotion engine" is a system that recognizes a user's emotional state and adjusts the feedback content based on that state.

[1249] "Review simulation" is a process in which documents submitted by users are virtually reviewed using a generative AI model.

[1250] "Feedback" refers to evaluations and advice provided to users based on the results of the review simulation.

[1251] "Materials" are documents such as proposals and project plans that users submit for review purposes.

[1252] "Server" is the central computing system that performs data collection, analysis, model training and simulation.

[1253] MODE FOR CARRYING OUT THE INVENTION

[1254] The present invention relates to a system that creates a generative AI model using past data from internal review meetings and uses the model to virtually simulate a review. Furthermore, by adding an emotion engine that recognizes the user's emotions, the system aims to provide feedback based on the user's emotional state. The following describes in detail an embodiment of this system.

[1255] Data Collection Phase

[1256] The server collects past audio and video data from internal review meetings. This data is obtained from recorders and cameras, or files exported from existing conference systems are used. For example, audio data is collected in .wav file format, and video data is collected in .mp4 file format. The collected audio and video data is stored in the server's storage.

[1257] Data analysis and processing

[1258] The server processes the collected voice data using a speech recognition engine such as the Google Speech-to-Text API and converts it into text data. This process also identifies the speaker and tags the speech over time. The server then processes the video data using a video analysis engine such as the Microsoft Azure Face API to analyze facial and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is then integrated with the text data and saved as a single unified dataset.

[1259] Training a generative AI model

[1260] The server uses the unified dataset to train the generative AI model. This training uses machine learning techniques such as TensorFlow. The dataset is used to train the model to reproduce the tone, pronunciation, and emotional patterns of a specific executive. Once trained, the generative AI model can faithfully simulate the executive's speech and facial expressions.

[1261] Simulation Phase

[1262] A user sends a request for an internal review simulation to the server from their device. The user uploads the document they wish to simulate (e.g., a proposal or project plan). This document is transferred to the server and becomes the subject of review by the generative AI model. The server analyzes the document, and a virtual executive reviews it using the generative AI model. The server then uses an emotion engine to analyze the user's video data and identify their emotional state in real time.

[1263] Conducting a review

[1264] The generated AI executive reviews the user's materials and provides appropriate comments and feedback in real time. This feedback is displayed on the user's device, allowing the user to simulate the interaction in an interactive format. The generative AI model reproduces the executive's tone and emotional patterns, providing an atmosphere similar to that of a conversation with an actual executive. Furthermore, the emotion engine recognizes the user's emotional state and adjusts the content of the generated AI executive's feedback. For example, if the user is nervous, the generated AI executive can provide feedback in a softer tone.

[1265] Providing Feedback

[1266] After the simulation is complete, the server analyzes the generated feedback and provides the user with a comprehensive report that includes points for improvement, good points, and specific advice, which the user can use to prepare for the actual review meeting.

[1267] Specific examples

[1268] An employee has prepared a proposal for a new project and wants to have it reviewed. The employee accesses an internal review simulation tool from their PC and logs into the system. Next, they upload the proposal file to the server. The server receives the file and launches a generating AI executive. The generating AI executive reviews the proposal's contents in detail and provides comments and questions to the user in real time. Through dialogue with the virtual executive, the user understands which parts of their document need improvement. In addition, an emotion engine grasps the user's emotional state in real time, and if the user is nervous, the generating AI executive will respond flexibly. Finally, the server generates a feedback report and provides it to the user.

[1269] As a concrete example, let's input the following prompt to the generative AI model:

[1270] "Present your proposal and get feedback from virtual executives."

[1271] The above is a detailed description of an embodiment of the present invention. Based on this embodiment, users can perform multiple simulations to improve the quality of their reviews. The emotion engine provides feedback tailored to the user's individual emotional state, which can lead to more effective learning and improvement.

[1272] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1273] Step 1: Data collection

[1274] The server collects past audio and video data from internal review meetings. Input includes data captured from internal recorders and cameras, or data files exported from existing conference systems. Data processing involves saving the audio data (e.g., .wav files) and video data (e.g., .mp4 files) in a specified directory. The output is the audio and video data saved in storage.

[1275] Step 2: Analyzing the audio data

[1276] The voice data collected by the server is processed by a voice recognition engine (e.g., Google Speech-to-Text API) and converted into text data. The input includes voice data. Data calculations include the process of performing voice recognition and converting the data into text data with time tags and speaker information attached. Specific operations include making an API call and saving the results as a text file. The output is text data with time tags attached.

[1277] Step 3: Analyzing the video data

[1278] The video data collected by the server is processed using a video analysis engine (e.g., Microsoft Azure Face API) to extract facial and movement characteristics. The input includes video data. Data calculations include the process of analyzing the video and extracting facial expression data (e.g., smiling, sadness, anger, etc.) and movement data (e.g., hand movements, eye focus, etc.). Specific operations include making an API call and saving the results in JSON format. Analyzed facial expression and movement data are obtained as output.

[1279] Step 4: Integrate the data

[1280] The server integrates the analyzed audio and video data and stores it as a single unified dataset. The input includes time-tagged text data and analyzed facial and movement data. The data processing involves merging the audio and video data based on the time tags to create a single unified dataset. The specific operation is to store it as a unified dataset in a database. The unified dataset is obtained as the output.

[1281] Step 5: Training the generative AI model

[1282] The server uses a unified dataset to train the generative AI model. The input includes the unified dataset. Data processing involves using a machine learning framework (e.g., TensorFlow) to learn the tone, pronunciation, and emotional patterns of a specific executive. Specifically, the dataset is fed into a neural network model, training is performed, and the optimal model parameters are saved. The output is a trained generative AI model.

[1283] Step 6: Receiving a simulation request

[1284] The user sends a request for an internal review simulation to the server from their own device. The input includes the user's login information and the desired materials for the simulation. Specific operations include uploading the materials through the user interface and sending the request to the server. The output shows the status of the request being sent to the server.

[1285] Step 7: Analyze the data and perform simulations

[1286] The server analyzes the materials received from the user, and the virtual executive reviews them using a generative AI model. Inputs include the user's materials and the generative AI model. Data calculations include analyzing the content of the materials using a text analysis engine (e.g., SpaCy or NLTK) and extracting important points and keywords. This data is passed to the generative AI model, and the virtual executive begins the review. Specifically, the model's output is fed back to the user's device in real time. Real-time comments and questions are provided to the user as output.

[1287] Step 8: Run the Emotion Engine

[1288] The server uses an emotion engine to analyze the user's video data and identify their emotional state in real time. The input includes the user's video data. The data calculation process involves analyzing the data using an emotion recognition algorithm and extracting the user's emotional state (e.g., tension, surprise, relief). Specific operations include calling the emotion engine's API and adjusting the feedback content based on the analysis results. The output is feedback corresponding to the user's emotional state.

[1289] Step 9: Provide feedback

[1290] The generated AI manager evaluates the user's materials and provides appropriate comments and feedback in real time. Specifically, it responds flexibly to the user's tension or anxiety and adjusts the content of the feedback. The output is the comments and advice provided as feedback that is displayed to the user.

[1291] Step 10: Generate a feedback report

[1292] The server analyzes the feedback generated after the simulation has finished and provides a comprehensive report to the user. The input includes a record of all interactions during the simulation. Data processing involves analyzing the log data and summarizing areas for improvement, positives, and specific action items. Specific operations include applying a report generation algorithm to create a report in a format that the user can download (e.g., PDF). The output is a comprehensive feedback report provided to the user.

[1293] The above is a detailed description of the specific processing steps of the system. By following this procedure, users can effectively conduct a review simulation and prepare for an actual review meeting.

[1294] (Application example 2)

[1295] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1296] In the design and development of conventional autonomous vehicles, there is a problem that the review simulations performed by engineers and designers often deviate significantly from the feedback given in actual review meetings. In addition, there is a lack of a means to provide feedback that takes into account the user's emotional state, which hinders more effective improvements.

[1297] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes: means for collecting and processing audio data and video data, which are past data of internal review meetings; means for training a generative AI model based on the processed audio data and video data; means for simulating a review of materials submitted by a user using the generative AI model; means for adjusting feedback content, including an emotion engine that analyzes the user's emotional state; and means for providing the simulation results to the user as feedback. This enables a virtual review simulation using the generative AI model, and can provide feedback based on the user's emotional state, enabling effective improvements in an environment that more closely resembles an actual meeting.

[1298] "Audio data" refers to digital data that records the sounds generated during a meeting, conversation, or the like.

[1299] "Video data" is digital data that records visual information generated during meetings, conversations, etc.

[1300] A "generative AI model" is an artificial intelligence model that is trained using machine learning techniques based on past data and can simulate judgments and actions for specific tasks.

[1301] "Review simulation" is the process of virtually evaluating and examining the materials submitted by the user and providing the results as feedback.

[1302] "Feedback" refers to evaluations and advice obtained as a result of the simulation and is provided to the user.

[1303] An "emotion engine" is a technology that analyzes a user's emotional state from facial expressions, tone of voice, etc.

[1304] "Server" means a computer system for collecting, analyzing data, and training generative AI models.

[1305] "Training" is the process by which a generative AI model learns using past data to improve its performance on a specific task.

[1306] A "user" is a person who uses the system to undergo a review simulation.

[1307] The following describes an embodiment of the present invention.

[1308] System Configuration

[1309] The system of this invention includes a server, a user terminal, an emotion engine, and a generative AI model. The server collects and analyzes data, trains the generative AI model, and sends the generated feedback to the user terminal. The user terminal provides an interface for the user to review the simulation. The emotion engine analyzes the user's video data and detects their emotional state in real time.

[1310] Program processing

[1311] The server collects audio and video data from past internal review meetings. This data is acquired from recorders and cameras, or files exported from existing conference systems. The collected audio data is stored in the server's storage in sound wave format, and the video data is stored in video format.

[1312] The server then processes the collected voice data using a speech recognition engine (e.g., Google Speech-to-Text API) and converts it into text data. During this process, it also identifies the speaker and tags the speech over time. The video data is then processed using a video analysis engine (e.g., OpenCV) to analyze facial and movement characteristics. This makes it possible to identify the speaker's emotions and changes in facial expression. The analyzed data is then integrated with the text data and saved as a single unified dataset.

[1313] Using the unified dataset, the server trains a generative AI model (such as GPT-3 or BERT), which uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief. Once trained, the generative AI model can faithfully simulate the chief's speech and facial expressions.

[1314] To conduct an internal review simulation, a user sends a simulation request to the server from their own device. The user uploads the materials they wish to simulate (e.g., design documents or prototypes). These materials are transferred to the server and become the subject of review by the generative AI model. The server analyzes the materials, and the virtual head of the department conducts the review using the generative AI model.

[1315] Furthermore, the server uses an emotion engine to analyze the user's video data and identify their emotional state in real time. The Generative AI Headquarters can then adjust the feedback content based on the user's emotional state, enabling more flexible responses. For example, if the user is nervous, the server will provide feedback in a softer tone.

[1316] Specific examples

[1317] An engineer has prepared a design document for a new autonomous vehicle and wants to have it reviewed. The engineer accesses the "Automotive Development Simulation Trainer" from his or her PC and logs into the system. Next, he or she uploads the design document file to the server. The server receives the file and launches the Generative AI Headquarters. The Generative AI Headquarters reviews the contents of the design document in detail and provides comments and questions to the engineer in real time. Through dialogue with the virtual headquarters, the engineer understands which parts of his or her document need improvement. In addition, the emotion engine grasps the engineer's emotional state in real time, and if the engineer seems nervous, the Generative AI Headquarters will respond flexibly. Finally, the server generates a feedback report and provides it to the engineer, allowing him or her to prepare for the next actual review.

[1318] Prompt Sentence Examples

[1319] "Please upload the design document below. We will start the simulation and flexibly adjust the feedback based on the user's emotional state."

[1320] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1321] Step 1:

[1322] The server collects audio and video data from past internal review meetings. The audio and video data is acquired from recorders and cameras, or files exported from existing conference systems are used. This data is saved in the server's storage. The input is audio and video data, and the output is the saved raw data.

[1323] Step 2:

[1324] The server processes the collected voice data using a voice recognition engine (e.g., Google Speech-to-Text API) and converts it into text data. This process also identifies the speaker and tags the speech with time. The input is voice data, and the output is text data with speaker identification and time tags.

[1325] Step 3:

[1326] The server processes the video data using a video analysis engine (e.g., OpenCV) to analyze facial expressions and movement features. This makes it possible to identify the speaker's emotions and changes in facial expressions. The input is video data, and the output is data containing emotional states and facial expression features.

[1327] Step 4:

[1328] The server integrates the analyzed audio and video data and stores them as a unified dataset. The input is text data and facial expression data, and the output is the integrated dataset.

[1329] Step 5:

[1330] The server uses the unified dataset to train a generative AI model (e.g., GPT-3 or BERT). This training uses machine learning techniques to replicate the tone, pronunciation, and emotional patterns of a specific chief executive. The input is the unified dataset, and the output is the trained generative AI model.

[1331] Step 6:

[1332] A user sends a request to the server from their own device to run an internal review simulation. The user uploads materials such as design documents and prototypes. The input is the user request and the uploaded materials, and the output is the transfer of the materials to the server.

[1333] Step 7:

[1334] The server analyzes the uploaded materials, and the virtual director reviews them using a generative AI model. The input is the uploaded materials, and the output is the review comments and feedback of the generative AI director.

[1335] Step 8:

[1336] The server uses an emotion engine to analyze the user's video data and identify the user's emotional state in real time. The input is the user's video data, and the output is the user's emotional state.

[1337] Step 9:

[1338] The Generative AI Director adjusts the feedback content based on the user's emotional state, enabling flexible responses. For example, if the user is nervous, the feedback will be given in a softer tone. The input is the user's emotional state and the Generative AI Director's review comments, and the output is the adjusted feedback.

[1339] Step 10:

[1340] The server generates a feedback report and provides it to the user. The report includes areas for improvement, good points, and specific advice. The input is the tailored feedback, and the output is the feedback report.

[1341] Through these steps, the system enables users to make effective improvements in an environment similar to an actual review meeting.

[1342] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1343] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1344] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1345] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1346] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1347] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1348] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1349] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1350] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1351] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1352] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1353] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1354] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1355] 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.

[1356] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1357] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1358] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1359] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1360] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1361] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1362] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1363] The following is further disclosed regarding the above embodiment.

[1364] (Claim 1)

[1365] A means for collecting and processing audio and video data from past internal review meetings;

[1366] means for training a generative AI model based on the processed audio and video data;

[1367] means for simulating a review of a document submitted by a user using the generative AI model;

[1368] means for providing the simulation results to a user as feedback;

[1369] A system including:

[1370] (Claim 2)

[1371] 2. The system according to claim 1, wherein the review simulation means performs the review in an interactive format with the user.

[1372] (Claim 3)

[1373] 10. The system of claim 1, wherein the generative AI model replicates the tone and emotional patterns of a chief of staff.

[1374] "Example 1"

[1375] (Claim 1)

[1376] means for collecting and processing audio and video data;

[1377] means for training a generative AI model based on the processed audio and video data;

[1378] means for using the generative AI model to review user-submitted materials;

[1379] means for providing interactive interaction with a user during said review;

[1380] means for analyzing the simulation results and generating and providing a comprehensive feedback report;

[1381] A system including:

[1382] (Claim 2)

[1383] 10. The system of claim 1, wherein the generative AI model reproduces tone and emotional patterns.

[1384] (Claim 3)

[1385] The system of claim 1, characterized in that the materials submitted by the user are analyzed and reviewed by a virtual head of department using a generative AI model.

[1386] "Application Example 1"

[1387] (Claim 1)

[1388] A means for collecting and processing audio and video data from past internal review meetings;

[1389] means for training a generative AI model based on the processed audio and video data;

[1390] means for simulating a review of a document submitted by a user using the generative AI model;

[1391] means for providing the simulation results to a user as feedback;

[1392] A means for reviewing an operation plan or a maintenance report for industrial equipment using the generative AI model;

[1393] A means to request reviews and upload materials from your smartphone;

[1394] A system including:

[1395] (Claim 2)

[1396] 2. The system according to claim 1, wherein the review simulation means performs the review in an interactive format with the user.

[1397] (Claim 3)

[1398] 10. The system of claim 1, wherein the generative AI model replicates the tone and emotional patterns of a chief of staff.

[1399] "Example 2: Combining Emotion Engines"

[1400] (Claim 1)

[1401] A means for collecting and processing audio and video data from past internal review meetings;

[1402] means for training a generative AI model using machine learning techniques based on the processed audio and video data;

[1403] means for recognizing the emotional state of the user and adjusting the feedback content;

[1404] means for simulating a review of a document submitted by a user using the generative AI model;

[1405] means for providing the simulation results to a user as feedback;

[1406] A system including:

[1407] (Claim 2)

[1408] 2. The system according to claim 1, wherein the review simulation means interactively conducts a review with the user and adjusts feedback according to the user's emotional state.

[1409] (Claim 3)

[1410] 10. The system of claim 1, wherein the generative AI model reproduces the tone and emotional patterns of a specific executive.

[1411] "Application example 2 when combining emotion engines"

[1412] (Claim 1)

[1413] A means for collecting and processing audio and video data from past internal review meetings;

[1414] means for training a generative AI model based on the processed audio and video data;

[1415] means for simulating a review of a document submitted by a user using the generative AI model;

[1416] means for providing the simulation results to a user as feedback;

[1417] a means for adjusting the feedback content, the means including an emotion engine for analyzing the user's emotional state;

[1418] A system including:

[1419] (Claim 2)

[1420] 2. The system according to claim 1, wherein the review simulation means performs the review in an interactive format with the user.

[1421] (Claim 3)

[1422] 10. The system of claim 1, wherein the generative AI model replicates the tone and emotional patterns of a chief of staff. [Explanation of symbols]

[1423] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for collecting and processing audio and video data from past internal review meetings; means for training a generative AI model based on the processed audio and video data; means for simulating a review of a document submitted by a user using the generative AI model; means for providing the simulation results to a user as feedback; A system including:

2. 2. The system according to claim 1, wherein said review simulation means performs a review in an interactive format with a user.

3. 10. The system of claim 1, wherein the generative AI model replicates the tone and emotional patterns of a chief of staff.

Images