System

Abstract

A system is provided.SOLUTION: A system comprising: means for parsing an input prompt and selecting an optimal generative artificial intelligence engine and version thereof; means for transmitting the prompt to the selected generative artificial intelligence engine and receiving a generated response; means for transmitting the generated response to a user; and means for receiving feedback from the user and updating the selection algorithm of the generative artificial intelligence engine.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] Systems using generative artificial intelligence models have become extremely convenient in recent years, but the use of advanced generative AI models poses challenges, such as increased computational costs and processing time. Furthermore, users must manually select the appropriate AI model and version each time, which presents a hurdle to their use. The present invention aims to solve this challenge by providing a system that automatically selects the optimal generative AI engine and its version for each prompt entered by the user, thereby reducing computational costs and time. [Means for solving the problem]

[0005] The present invention solves the above-mentioned problem by providing a system including: means for analyzing an input prompt and selecting an optimal generative AI engine and its version; means for sending the prompt to the selected generative AI engine and receiving a generated response; means for sending the generated response to a user; and means for receiving feedback from the user and updating the selection algorithm of the generative AI engine. This system automatically selects the optimal generative AI engine and its version based on the content of the prompt and by referring to a preset evaluation score and feedback data. Furthermore, by aggregating user feedback and updating the selection algorithm, it is possible to achieve more appropriate engine selection in the future.

[0006] An "input prompt" is a sentence that refers to a question or request that a user enters into a system.

[0007] "Parsing" refers to the operation of analyzing the content of an input prompt to identify its characteristics and the type of information requested.

[0008] A "generative artificial intelligence engine" is an AI model that uses natural language processing technology to generate appropriate answers and information based on user input prompts.

[0009] "Version" refers to different releases or updated states of a particular Generative AI Engine.

[0010] "Selection" is the process of determining the optimal generative AI engine and its version based on the analysis results.

[0011] "Send" refers to the operation of transferring an input prompt to the selected generative AI engine.

[0012] "Receiving" refers to the operation in which the system obtains the response sent back from the generating AI engine.

[0013] "Response" refers to the answer or information generated by the generative AI engine in response to an input prompt.

[0014] "User" refers to any person or end user who uses the System.

[0015] "Feedback" refers to a user's evaluation or provision of opinion on a generated response.

[0016] "Selection algorithm" refers to a calculation method for automatically selecting the optimal generative AI engine and version based on analysis results and feedback.

[0017] "Profile updating" refers to the process of adjusting the parameters and settings of the selection algorithm based on user feedback. [Brief explanation of the drawings]

[0018] [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

[0019] 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.

[0020] First, the terms used in the following description will be explained.

[0021] 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).

[0022] 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.

[0023] 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.

[0024] 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.

[0025] 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."

[0026] [First embodiment]

[0027] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0028] 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.

[0029] 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).

[0030] 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.

[0031] 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.

[0032] 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.

[0033] 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.

[0034] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0035] 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.

[0036] 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.

[0037] 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.

[0038] 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."

[0039] The present invention relates to a system that automatically selects the optimal generation AI engine and its version for a prompt entered by a user, thereby reducing calculation costs and time. This system is implemented through interactions between a server, a terminal, and a user.

[0040] System Overview

[0041] 1. Enter and submit the prompt

[0042] The user types any question or request into an input field on the device, and this prompt is sent from the device to the server.

[0043] 2. Accepting and parsing prompts

[0044] The server receives the prompts sent from the device and uses text analysis algorithms to analyze the prompt content, including the type of question, the depth of information required, and the complexity of the vocabulary.

[0045] 3. Selection of the optimal engine

[0046] The server selects the optimal generative AI engine and its version based on the analysis results and references pre-set evaluation scores and feedback data collected from past dialogue history.

[0047] 4. Sending prompts and generating responses

[0048] The server sends prompts to the selected generative AI engine, which generates an appropriate response based on the received prompt and sends it back to the server.

[0049] 5. Sending the response and displaying it to the user

[0050] The server sends the generated response to the terminal, and the terminal displays the received response to the user.

[0051] 6. Gathering feedback and updating the algorithm

[0052] Users can input their feedback on the displayed responses into the device, which then sends it to the server, which receives the feedback and updates the selection algorithm. This feedback is collected as user satisfaction and additional comments.

[0053] Specific examples

[0054] For simple questions

[0055] 1. The user enters the prompt

[0056] User: Type "What's the weather like today?"

[0057] 2. The device sends a prompt to the server

[0058] The terminal sends the entered prompt to the server.

[0059] 3. The server parses the prompt and selects an engine

[0060] The server analyzes the prompt and determines that it is a simple question. To keep costs down, it selects Generative AI Engine 3.5.

[0061] 4. Send a prompt to the engine and receive a response

[0062] The server sends a prompt to engine 3.5 and receives the response "The weather is sunny today."

[0063] 5. Send the response to the device and display it to the user

[0064] The server sends the response to the terminal, which displays the response to the user.

[0065] 6. Users provide feedback

[0066] User: Provides feedback saying "Very satisfied." The device sends the feedback to the server.

[0067] For complex technical questions

[0068] 1. The user enters the prompt

[0069] User: Type "Tell me more about regularization techniques for neural networks."

[0070] 2. The device sends a prompt to the server

[0071] The terminal sends the entered prompt to the server.

[0072] 3. The server parses the prompt and selects an engine

[0073] The server analyzes the prompt, understands that it is requesting detailed technical content, and selects a highly accurate generative AI engine4.

[0074] 4. Send a prompt to the engine and receive a response

[0075] The server sends a prompt to Engine 4 and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[0076] 5. Send the response to the device and display it to the user

[0077] The server sends the response to the terminal, which displays the response to the user.

[0078] 6. Users provide feedback

[0079] User: Provides feedback such as "somewhat satisfied." The device sends the feedback to the server.

[0080] This system allows users to receive fast and effective information without having to worry about the optimal AI engine, simply by entering prompts, and the system's accuracy is continuously improved through feedback.

[0081] The processing flow will be explained below.

[0082] Step 1:

[0083] A user types a prompt into an input field on the device. For example, the user types "Tell me about the latest AI technologies."

[0084] Step 2:

[0085] The terminal sends the entered prompt to the server, specifically as JSON-formatted data using an HTTP POST request.

[0086] Step 3:

[0087] The server receives the prompt sent from the device and parses it into the appropriate data format (e.g., JSON).

[0088] Step 4:

[0089] The server analyzes the prompt content using text analysis algorithms, including the type of question, the depth of information requested, and the complexity of the vocabulary.

[0090] Step 5:

[0091] Based on the analysis results, the server refers to a pre-set evaluation score and past feedback data to select the optimal generative AI engine and its version. For example, generative AI4 is selected for technical questions.

[0092] Step 6:

[0093] The server sends a prompt to the selected generative AI engine using an API request.

[0094] Step 7:

[0095] A generative AI engine receives the prompt and generates an appropriate response, which is sent back to the server.

[0096] Step 8:

[0097] The server then formats the response as needed, specifically converting the JSON data into HTML.

[0098] Step 9:

[0099] The server sends the formatted response to the device, which then sends it to the device as an HTTP response.

[0100] Step 10:

[0101] The device receives the response from the server and displays it to the user.

[0102] Step 11:

[0103] The user provides feedback on the generated response, including satisfaction and any additional comments.

[0104] Step 12:

[0105] The device sends the user feedback to the server, and the feedback data is sent in JSON format again.

[0106] Step 13:

[0107] The server receives user feedback and updates the selection algorithm, which improves the accuracy of engine selection from next time onwards.

[0108] Example 1

[0109] 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."

[0110] Currently, there are many generative AI engines available, each with different characteristics and performance. However, manually selecting which engine is best for a particular prompt is inefficient, and if an appropriate selection is not made, the quality of the generated response may decline. Furthermore, a method is needed to effectively utilize user feedback to update the engine selection algorithm and improve the system's accuracy. This will enable the provision of optimal responses while reducing computational costs and time.

[0111] 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.

[0112] In this invention, the server includes means for analyzing an input prompt and selecting an optimal generative AI engine and its version, means for sending the prompt to the selected generative AI engine and receiving a generated response, means for sending the generated response to a terminal and displaying it to the user, means for receiving feedback from the user and updating the selection algorithm of the generative AI engine, and user interface means for displaying the generated response to the user. This allows the user to quickly and effectively receive information simply by entering the prompt, without having to worry about selecting the optimal AI engine. Furthermore, the system's accuracy can be continuously improved by utilizing feedback.

[0113] An "input prompt" is text data that a user inputs into a terminal in the form of a question, request, or the like.

[0114] "Parsing" is the process of classifying and interpreting the content of the received input prompt using natural language processing algorithms.

[0115] A "generative artificial intelligence engine" is an artificial intelligence system capable of generating natural language responses based on a given input prompt.

[0116] A "version" is identification information that indicates different revisions or update states of a particular generative AI engine.

[0117] A "response" is a reply message in natural language generated by the selected artificial intelligence engine.

[0118] A "user" is a person who uses the system to enter questions or requests and receive responses generated by the AI ​​engine.

[0119] A "terminal" is a device (e.g., a PC, smartphone, etc.) on which a user enters prompts and receives and displays responses from the server.

[0120] A "text analysis algorithm" is an algorithm that uses natural language processing technology to analyze the meaning and content of input prompts.

[0121] The "evaluation score" is a numerical value used to evaluate the performance and appropriateness of the generative AI engine based on past dialogue history.

[0122] "Feedback data" refers to data that includes evaluations and comments on responses provided by users.

[0123] The "selection algorithm" is a method for selecting the optimal generative AI engine and version based on analysis results and feedback data.

[0124] "User interface means" refers to an interface (e.g., a web page, application screen, etc.) through which a user inputs a prompt and displays a response.

[0125] This invention relates to a system that automatically selects the optimal AI engine and its version for a prompt entered by a user, thereby reducing calculation costs and time. This system is implemented through interactions between a server, a terminal, and a user.

[0126] System configuration

[0127] 1. Hardware and Software Configuration

[0128] Server: Use a server or cloud service with powerful computing resources (e.g., AWS (registered trademark), Google (registered trademark) Cloud).

[0129] Terminal: Refers to the device used by the user (e.g., PC, smartphone, tablet).

[0130] Natural Language Processing Software: Use natural language processing libraries such as NLTK or SpaCy for text analysis algorithms.

[0131] Generative artificial intelligence engine: Uses generative AI models such as GPT-3 (registered trademark) and GPT-4 (registered trademark).

[0132] Processing flow

[0133] 1. Enter and submit the prompt

[0134] The user types a question or request into an input field on the terminal, and this prompt is sent from the terminal to the server.

[0135] Example: A user types, "What's the weather like today?" or "Tell me more about regularization techniques for neural networks."

[0136] 2. Accepting and parsing prompts

[0137] The server receives the prompts sent from the device and uses text analysis algorithms to analyze the prompt content, including the type of question, the depth of information required, and the complexity of the vocabulary.

[0138] 3. Selection of the optimal engine

[0139] The server refers to the evaluation scores and feedback data collected from past dialogue history and selects the optimal generative AI engine and its version based on the analysis results.

[0140] 4. Sending prompts and generating responses

[0141] The server sends a prompt to the selected AI engine and receives the generated response. For example, the response "It's sunny today" is generated in response to the question "What's the weather like today?"

[0142] 5. Sending the response and displaying it to the user

[0143] The server sends the generated response to the terminal, which displays the response to the user.

[0144] 6. Gathering Feedback and Updating the Algorithm

[0145] The user inputs feedback on the displayed response into the terminal, which then sends the feedback to the server, which receives the feedback and uses it to update the selection algorithm.

[0146] Specific examples

[0147] For simple questions

[0148] 1. The user enters the prompt "What's the weather like today?"

[0149] 2. The terminal sends a prompt to the server.

[0150] 3. The server parses the prompt, determines that it is a simple question, and then selects the 3.5 version of the engine to keep costs down.

[0151] 4. The server sends a prompt to the engine and receives a response saying, "The weather is sunny today."

[0152] 5. The server sends the response to the terminal, which displays it to the user.

[0153] 6. The user gives feedback that they are "very satisfied," and the device sends the feedback to the server.

[0154] For complex technical questions

[0155] 1. The user types the prompt "Tell me more about regularization techniques in neural networks."

[0156] 2. The terminal sends a prompt to the server.

[0157] 3. The server parses the prompt, understands that you are requesting detailed technical information, and selects one of four high-precision engine versions.

[0158] 4. The server sends a prompt to the engine and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[0159] 5. The server sends the response to the terminal, which displays it to the user.

[0160] 6. The user gives feedback saying "somewhat satisfied," and the device sends the feedback to the server.

[0161] This system provides users with fast and accurate information without them having to worry about which AI engine is best suited to their query, and the system's accuracy can be improved by continuously utilizing user feedback.

[0162] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0163] Step 1: Enter the prompts

[0164] The user enters a question or request into an input field on the terminal.

[0165] As a concrete example, a user inputs text such as "What's the weather like today?" or "Please tell me more about regularization techniques for neural networks."

[0166] Input: Prompt sentence (e.g. "What's the weather like today?")

[0167] Output: prompt text formatted as text data

[0168] Step 2: Sending the prompt

[0169] The device sends the prompt entered by the user to the server by sending the prompt data to the server's API endpoint using an HTTP request.

[0170] Input: prompt text formatted as text data

[0171] Output: Prompt data sent to the server

[0172] Specific behavior: Issues an HTTP request over a network connection and sends prompt data.

[0173] Step 3: Accept the prompt

[0174] The server receives the prompt sent from the device and prepares it for analysis.

[0175] Input: Prompt data sent from the terminal

[0176] Output: Prompt data waiting to be parsed

[0177] Step 4: Parse the prompt

[0178] The server analyzes the prompt using text analysis algorithms (e.g., NLTK, SpaCy) to determine the type of question, the depth of information required, and lexical complexity.

[0179] Input: Prompt data waiting to be parsed

[0180] Output: Analysis results (e.g., question characteristics, difficulty, topic)

[0181] Specific operation: Tokenize the text, perform grammatical analysis, and semantic analysis, and extract each attribute.

[0182] Step 5: Selecting the optimal engine

[0183] Based on the analysis results, the server refers to the evaluation scores and feedback data collected from past dialogue history and selects the optimal generative AI engine and its version.

[0184] Input: Analysis results, evaluation scores, feedback data

[0185] Output: Selected generative AI engine and its version

[0186] What it does: Searches for suitable engines and versions in the database, and compares the rating scores and feedback to select the best engine.

[0187] Step 6: Sending prompts

[0188] The server sends a prompt to the selected AI engine and sends the prompt data using the AI ​​engine's API.

[0189] Input: Selected generation AI engine and its version, prompt data

[0190] Output: Prompt data sent to the generative AI engine

[0191] Specific behavior: Issues an API call and sends a prompt to the engine.

[0192] Step 7: Generate the Response

[0193] The generative artificial intelligence engine generates appropriate responses based on the prompts and sends them back to the server.

[0194] Input: prompt data

[0195] Output: The generated response

[0196] Specific behavior: Generates a natural language response using the internal model. Example: In response to the prompt "What's the weather like today?", generates "The weather is sunny today."

[0197] Step 8: Sending the Response

[0198] The server sends the generated response to the terminal.

[0199] Input: The generated response

[0200] Output: Response data sent to the device

[0201] Specific operation: The generated text is sent back to the terminal as an HTTP response.

[0202] Step 9: View the response

[0203] The terminal displays the received response to the user.

[0204] Input: Response data received from the server

[0205] Output: Response displayed on screen

[0206] Specific behavior: Display text on the user interface. Example: Display "Today's weather is sunny" on the screen.

[0207] Step 10: Provide feedback

[0208] The user inputs feedback on the displayed response into the terminal.

[0209] Input: Rating and comments on the response

[0210] Output: Input feedback data

[0211] Specific actions: Clicking the rating button or typing in the comment box and submitting.

[0212] Step 11: Submit your feedback

[0213] The terminal sends the feedback to the server.

[0214] Input: Feedback data

[0215] Output: Feedback data sent to the server

[0216] Specific operation: Sends feedback data to the server as an HTTP POST request.

[0217] Step 12: Update the algorithm

[0218] The server updates its selection algorithm based on the feedback it receives.

[0219] Input: Feedback data

[0220] Output: Updated election algorithm

[0221] What it does: The feedback data is stored in a database and used as training data for machine learning models, including periodic retraining.

[0222] (Application example 1)

[0223] 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."

[0224] Modern content delivery services require the ability to quickly provide optimal responses to a variety of prompts entered by users. Conventional systems often use the same generative AI engine for all requests, resulting in problems such as high-cost engines always being used or an inability to select an appropriate engine. Another issue is that user feedback is rarely reflected in the system, making it difficult to improve service quality. The present invention aims to solve these problems and provide a system for effectively providing customized content to users.

[0225] 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.

[0226] In this invention, the server includes means for analyzing an input prompt and selecting an optimal generating AI engine and its version, means for sending the prompt to the selected generating AI engine and receiving a generated response, means for sending the generated response to the user, means for receiving feedback from the user and updating the selection algorithm of the generating AI engine, and means for providing customized content based on the generated response. This makes it possible to select an optimal engine for the user's input prompt and provide customized content efficiently and effectively.

[0227] An "input prompt" is a request or question that a user sends to a system.

[0228] A "generative artificial intelligence engine" is an artificial intelligence algorithm that generates appropriate responses based on user prompts.

[0229] "Version" is a number or state that represents a different release or improvement of the Generative AI Engine.

[0230] "Server" is the central system that processes user requests, selects the appropriate engine, and generates and sends the response.

[0231] "User" means an end user who utilizes the system to enter prompts and receive responses and content.

[0232] A "response" is a response generated by a generative artificial intelligence engine based on a prompt.

[0233] "Feedback" refers to the opinions and ratings that users provide in response to a response.

[0234] A "selection algorithm" is a process or method for selecting the optimal generative AI engine and its version.

[0235] "Customized Content" means information and materials that are optimized for a particular user based on the user's prompts and feedback.

[0236] The present invention is a system that automatically selects the optimal generative AI engine and its version for a prompt input by a user, and provides customized content based on the selection result. This system is implemented through interactions between a server, a terminal, and a user.

[0237] The system program is structured as follows: First, the user uses the terminal to input a prompt. This prompt is in text format, such as "I would like to read an article about the latest AI technology." The terminal then sends this input prompt to the server.

[0238] The server then analyzes the received prompt using a text analysis algorithm that evaluates the prompt's content, vocabulary complexity, and required information depth. Based on the analysis results, the server selects the optimal generative AI engine and its version. The selection is made using an algorithm that references pre-set evaluation scores and past feedback data.

[0239] A prompt is sent to the selected generative AI engine, causing it to generate an appropriate response. For example, this response might be, "The following are recommended articles about the latest AI technology: 1. The cutting edge of AI research 2. The evolution of deep learning 3. The future of coexistence between AI and humans." The server receives this generated response and sends it to the device.

[0240] The terminal displays the received response to the user, who can then provide feedback on the displayed response, such as "very satisfied" or "somewhat satisfied." This feedback is then sent back to the server from the terminal.

[0241] The server updates the generative AI engine selection algorithm based on the received feedback. This algorithm update will enable more accurate engine selection in the future. For example, if users provide a large number of "very satisfied" feedback, the server recognizes that the engine version is highly rated and reflects this information in the next selection.

[0242] The hardware required to realize this system includes the user's device (smartphone, PC, etc.), server, and internet communication environment. The software includes a text analysis algorithm, an evaluation score and feedback data management algorithm, a generative AI engine, and a server program that integrates them.

[0243] In this way, the present invention allows users to select the most appropriate engine for their input prompts and receive customized content efficiently and effectively, which not only improves user satisfaction but also contributes to cost and resource optimization.

[0244] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0245] Step 1:

[0246] The user inputs a prompt using the terminal. For example, the user inputs text such as "I would like to read an article about the latest AI technology." The input prompt is sent from the terminal to the server.

[0247] Input: A prompt from the user (e.g., "I want to read an article about the latest AI technology")

[0248] Output: Prompt sent from terminal to server

[0249] Step 2:

[0250] The server receives the prompt text sent from the device, inputs it into a text analysis algorithm, and analyzes the content of the prompt. As a result of the analysis, the type of question, the depth of information required, the complexity of vocabulary, etc. are extracted.

[0251] Input: Prompt sent from the terminal

[0252] Output: Analysis of the prompt (type of question, depth of information required, vocabulary complexity, etc.)

[0253] Step 3:

[0254] The server selects the optimal generative AI engine and its version based on the analyzed prompts, taking into account pre-set evaluation scores and past feedback data.

[0255] Input: Parsed result of prompt statement

[0256] Output: Selected generative AI engine and its version

[0257] Step 4:

[0258] The server sends the prompt sentence to the selected AI engine to generate an appropriate response. The AI ​​engine generates a response based on the prompt sentence and sends it back to the server.

[0259] Input: Selected artificial intelligence engine, prompt

[0260] Output: The generated response

[0261] Step 5:

[0262] The server receives the generated response and sends it to the terminal, which displays the received response to the user.

[0263] Input: The generated response

[0264] Output: Sending the response to the terminal and displaying it to the user

[0265] Step 6:

[0266] The user provides feedback on the displayed response. For example, the user inputs an evaluation such as "very satisfied" into the terminal. The terminal then transmits the feedback to the server.

[0267] Input: User feedback (e.g. "Very satisfied")

[0268] Output: Sending feedback from the device to the server

[0269] Step 7:

[0270] The server updates the generative AI engine's selection algorithm based on the received feedback, allowing it to select a more appropriate engine and version for the next prompt.

[0271] Input: User feedback

[0272] Output: Updated election algorithm

[0273] In this way, the present invention realizes a system that selects the most suitable generative artificial intelligence engine in response to a user's prompts and provides customized content efficiently and effectively.

[0274] 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.

[0275] This invention relates to a system that automatically selects the optimal generative AI engine and its version for a prompt entered by a user, and further combines it with an emotion engine that recognizes the user's emotions, thereby reducing calculation costs and time and providing a more advanced response. This system is implemented through interactions between four parties: a server, a terminal, an emotion engine, and a user.

[0276] System Overview

[0277] 1. Input prompts and emotion recognition

[0278] Users can enter any question or request into the device's input field, and the emotion engine will simultaneously recognize emotions from the user's input, facial expressions, and voice.

[0279] 2. Sending prompts and emotion data

[0280] The device sends the entered prompt and recognized emotion data to the server, which is sent as JSON formatted data using an HTTP POST request.

[0281] 3. Accepting and parsing prompts

[0282] The server receives the prompt and emotion data sent from the device, analyzes the received data in an appropriate data format, and recognizes the prompt content and the user's emotion.

[0283] 4. Selection of the optimal engine

[0284] The server selects the optimal AI engine and its version based on the prompt analysis results and emotional data, referring to pre-set evaluation scores and past feedback data. Emotional data is also taken into consideration, so the engine that best suits the user's emotional state is selected.

[0285] 5. Sending prompts and generating responses

[0286] The server sends a prompt to the selected generative AI engine, which generates an appropriate response based on the received prompt and emotion data and sends it back to the server.

[0287] 6. Sending the response and displaying it to the user

[0288] The server then sends the generated response to the device, which then displays the received response to the user, including feedback based on the emotion data.

[0289] 7. Gathering feedback and updating the algorithm

[0290] Users can provide feedback on the displayed responses to their devices, which then send it to the server. The server receives the feedback and updates the selection algorithm. This feedback is collected as user satisfaction, additional comments, and emotional data.

[0291] Specific examples

[0292] For simple questions

[0293] 1. The user enters the prompt

[0294] User: As soon as the user types "What's the weather like today?", the device detects the emotion "happiness" from the user's facial expression.

[0295] 2. The device sends prompts and emotion data to the server

[0296] The terminal transmits the input prompt and the emotion data of "joy" to the server.

[0297] 3. The server analyzes the prompt and emotion data and selects the engine.

[0298] The server analyzes the prompt and emotion data, determines that it is a simple question, and selects Generative AI Engine 3.5.

[0299] 4. Send a prompt to the engine and receive a response

[0300] The server sends a prompt to engine 3.5 and receives the response "The weather is sunny today."

[0301] 5. Send the response to the device and display it to the user

[0302] The server sends the response to the terminal, which displays the response to the user.

[0303] 6. Users provide feedback

[0304] The user provides feedback such as "very satisfied," and the device sends this feedback to the server.

[0305] For complex technical questions

[0306] 1. The user enters the prompt

[0307] User: Type "Please tell me more about neural network regularization techniques" and in the process detect the emotion "interested" from the user's voice.

[0308] 2. The device sends prompts and emotion data to the server

[0309] The terminal transmits the input prompt and the emotion data of "interest" to the server.

[0310] 3. The server analyzes the prompt and emotion data and selects the engine.

[0311] The server analyzes the prompt and emotion data, determines that detailed technical content is required, and then selects a generative artificial intelligence engine 4.

[0312] 4. Send a prompt to the engine and receive a response

[0313] The server sends a prompt to Engine 4 and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[0314] 5. Send the response to the device and display it to the user

[0315] The server sends the response to the terminal, which displays the response to the user.

[0316] 6. Users provide feedback

[0317] The user provides feedback such as "somewhat satisfied," and the device sends this feedback to the server.

[0318] This system allows users to receive prompt and emotional information quickly and effectively without having to worry about the optimal AI engine, and the system's accuracy is continuously improved through feedback.

[0319] The processing flow will be explained below.

[0320] Step 1:

[0321] A user types a prompt into an input field on the device. For example, the user types "Tell me about the latest AI technologies."

[0322] Step 2:

[0323] The emotion engine analyzes the user's input, facial expressions, voice, etc. to recognize the user's emotions. For example, it can detect "interest" as an emotion.

[0324] Step 3:

[0325] The device sends the input prompt and emotion data to the server as JSON format data using an HTTP POST request.

[0326] Step 4:

[0327] The server receives the prompt and emotion data sent from the device and parses the received data into an appropriate data format (e.g., JSON).

[0328] Step 5:

[0329] The server analyzes the prompt content and emotional data using text analysis and emotion recognition algorithms, including the type of question, the depth of information requested, vocabulary complexity, and the user's emotions.

[0330] Step 6:

[0331] Based on the analysis results, the server refers to the preset evaluation score, past feedback data, and emotion data to select the optimal generative AI engine and its version. For example, if the question is technical and the emotion of "interest" is detected, generative AI engine 4 will be selected.

[0332] Step 7:

[0333] The server sends the prompt and emotion data to the selected generative AI engine using an API request.

[0334] Step 8:

[0335] A generative AI engine receives the prompt and emotion data and generates an appropriate response, which is then sent back to the server.

[0336] Step 9:

[0337] The server then formats the response as needed, specifically converting the JSON data into HTML.

[0338] Step 10:

[0339] The server sends the formatted response to the device, which then sends it to the device as an HTTP response.

[0340] Step 11:

[0341] The device receives the response from the server and displays it to the user.

[0342] Step 12:

[0343] Users provide feedback on the generated responses, including satisfaction, additional comments, and sentiment data.

[0344] Step 13:

[0345] The device sends the user feedback to the server, and the feedback data is sent in JSON format again.

[0346] Step 14:

[0347] The server receives user feedback and updates the selection algorithm, which improves the accuracy of engine selection from next time onwards.

[0348] Example 2

[0349] 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."

[0350] Conventional generative AI systems require manual selection of engines and versions to generate optimal responses to input prompts, resulting in increased computational costs and time for data processing and selection. Furthermore, it is difficult to automatically generate responses based on user emotions, and systems for improving user satisfaction are lacking. Furthermore, there is an insufficient mechanism for continuous improvement by aggregating and incorporating user feedback.

[0351] 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.

[0352] In this invention, the server includes means for receiving an input prompt from a user and recognizing the user's emotional data, means for analyzing the input prompt and the emotional data and selecting an optimal generating AI engine and its version, and means for sending the prompt to the selected generating AI engine and receiving a generated response, thereby enabling the selection of an optimal engine according to the user's emotional state and the rapid generation of a response.

[0353] "User" means an individual or entity that utilizes the System to enter prompts and receive information or services.

[0354] An "input prompt" is text data such as a question or request that a user inputs to a system.

[0355] "Emotion data" refers to data that indicates the emotional state of the user as recognized from facial expressions, voice, etc.

[0356] A "generative artificial intelligence engine" is an artificial intelligence model for generating appropriate responses based on user input prompts.

[0357] "Version" refers to different releases or revisions of the Generative Artificial Intelligence Engine.

[0358] "Analysis" refers to the process of verifying input prompts, emotional data, past feedback data, etc., and processing the data to select the optimal generative AI engine and its version.

[0359] A "response" is response data generated by a generative artificial intelligence engine in response to a user's input prompt.

[0360] "Feedback" refers to reaction data such as ratings and comments provided by users in response to generated responses.

[0361] A "selection algorithm" is a computational method or set of rules for analyzing prompts and emotional data and selecting the optimal generative AI engine and its version.

[0362] The "system" is a collection of components that executes a series of processes: receiving prompts and emotional data from users, selecting the optimal generative artificial intelligence engine to generate a response, aggregating feedback, and updating the algorithm.

[0363] This invention is a system that automatically selects the optimal generative AI engine and its version for the prompt entered by the user, and further combines it with an emotion engine that recognizes the user's emotions, thereby reducing calculation costs and time and providing a more advanced response. This system is implemented through interactions between elements including the user, terminal, server, and emotion engine.

[0364] Explanation of program processing

[0365] 1. User prompt input and emotion recognition

[0366] Users enter any question or request into an input field via a device such as a smartphone or computer. At the same time, the emotion engine uses the device's camera and microphone to analyze the user's facial expressions and voice and recognize their emotions. For example, when a user types "What's the weather like today?", the device detects the user's smile and identifies it as the emotion of "joy."

[0367] 2. Sending prompts and emotion data

[0368] The device compiles the input prompt and recognized emotion data in JSON format and sends it to the server using an HTTP POST request, specifically using the fetch API or XMLHttpRequest.

[0369] 3. Accepting and parsing prompts

[0370] The server receives the prompt and emotion data sent from the device and parses it into the appropriate format. It uses the JSON.parse method to parse the data and stores the prompt and emotion data in separate variables. During this process, it also references past feedback data and evaluation scores stored in a database (e.g., MySQL (registered trademark) or PostgreSQL).

[0371] 4. Selection of the optimal engine

[0372] The server selects the optimal AI engine and its version based on the prompt analysis results and emotion data, along with evaluation scores and feedback data. For example, Engine 3.5 is selected for simple questions, while Engine 4 is selected for detailed technical questions.

[0373] 5. Sending prompts and generating responses

[0374] The server sends a prompt to the selected generative AI engine, which then generates an appropriate response. The generative AI engine takes into account the received prompt and emotional data to create an appropriate response and sends it back to the server. For example, a response might be generated: "The weather is sunny today."

[0375] 6. Sending the response and displaying it to the user

[0376] The server sends the generated response to the device, which then displays it to the user. The device screen displays "Today's weather is sunny."

[0377] 7. Gathering feedback and updating the algorithm

[0378] The user inputs feedback on the displayed response into the terminal, including options such as "very satisfied" and "somewhat satisfied." The terminal then transmits the user's feedback to the server.

[0379] The server stores the received feedback in a database and updates the selection algorithm based on that feedback, which will result in better engine selection in the future.

[0380] Specific examples

[0381] For simple questions

[0382] 1. User: Enter "What's the weather like today?" The device detects "happiness" from the user's facial expression.

[0383] 2. Device: Sends the prompt and the "happiness" emotion data to the server.

[0384] 3. Server: Analyzes prompt and sentiment data and selects engine 3.5.

[0385] 4. Server: Sends a prompt to Engine 3.5 and receives the response "The weather is sunny today."

[0386] 5. Server: Sends the response to the terminal, which displays it to the user.

[0387] 6. User: Gives feedback that he is "very satisfied," and the device sends the feedback to the server.

[0388] For complex technical questions

[0389] 1. User: Type "Please tell me more about neural network regularization techniques." The device detects "interests" from the user's voice.

[0390] 2. Device: Sends prompt and "interest" emotion data to the server.

[0391] 3. Server: Analyzes prompt and emotion data and selects Engine 4.

[0392] 4. Server: Sends a prompt to Engine 4 and receives a response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[0393] 5. Server: Sends the response to the terminal, which displays it to the user.

[0394] 6. User: Gives feedback that they are "somewhat satisfied," and the device sends that feedback to the server.

[0395] The system selects the most suitable AI engine and provides fast and effective information by simply inputting prompts and emotional data, while the accuracy of the system is continuously improved through feedback.

[0396] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0397] Divide the processing flow of the system program into processing steps

[0398] Step 1: User prompt input and emotion recognition

[0399] 1.1 Prompt Input

[0400] 1.2 Emotion recognition

[0401] Step 2: Sending prompts and emotion data

[0402] Step 3: Server accepts and parses prompt

[0403] 3.1 Receiving prompts and emotion data

[0404] 3.2 Prompt and Emotion Data Analysis

[0405] Step 4: Selecting the optimal engine

[0406] 4.1 Referencing evaluation scores and feedback data

[0407] 4.2 Selecting the optimal engine and version

[0408] Step 5: Sending prompts and generating responses

[0409] 5.1 Sending Prompts to the Selection Engine

[0410] 5.2 Engine Generates and Returns Response

[0411] Step 6: Sending the Response and Displaying it to the User

[0412] Step 7: Gather feedback and update the algorithm

[0413] 7.1 Entering and Submitting Feedback

[0414] 7.2 Receiving Feedback and Updating the Algorithm

[0415] Specific explanation of each processing step

[0416] Step 1: User prompt input and emotion recognition

[0417] 1.1 Prompt Input

[0418] The user writes a question or request into the input field of the terminal. For example, "What's the weather like today?" Input: The text data entered by the user into the terminal. Output: The prompt text entered.

[0419] 1.2 Emotion recognition

[0420] The device uses the built-in camera and microphone to collect the user's facial expressions and voice, and the emotion engine analyzes their emotions in real time. For example, if the user smiles, the emotion "joy" is detected. Input: User's facial and voice data. Output: Analyzed emotion data (e.g., "joy").

[0421] Step 2: Sending prompts and emotion data

[0422] The device collects the input prompt and emotion data and sends them to the server in JSON format. Specifically, it makes an HTTP POST request using the fetch API or XMLHttpRequest. Input: Prompt text and emotion data. Output: JSON data sent to the server.

[0423] Step 3: Server accepts and parses prompt

[0424] 3.1 Receiving prompts and emotion data

[0425] The server receives the prompt and emotion data sent from the device. Input: JSON data sent from the device. Output: Received prompt and emotion data.

[0426] 3.2 Prompt and Emotion Data Analysis

[0427] The server parses the prompt and emotion data using the JSON.parse method and stores them in separate variables. Input: Received prompt and emotion data. Output: Parsed prompt and emotion data.

[0428] Step 4: Selecting the optimal engine

[0429] 4.1 Referencing evaluation scores and feedback data

[0430] The server references past feedback data and evaluation scores from a database. For example, queries are performed using MySQL or PostgreSQL. Input: prompt text and emotion data. Output: evaluation scores and feedback data required to select the optimal engine.

[0431] 4.2 Selecting the optimal engine and version

[0432] The server selects the optimal generative AI engine and its version based on the prompt, emotion data, evaluation scores, and feedback data. Input: Evaluation scores and feedback data. Output: Selected optimal engine and its version.

[0433] Step 5: Sending prompts and generating responses

[0434] 5.1 Sending Prompts to the Selection Engine

[0435] The server sends a prompt to the selected engine. Specifically, it sends an HTTP request to the engine's API endpoint. Input: Selected engine and prompt text. Output: Sending prompt data to the generative AI engine.

[0436] 5.2 Engine Generates and Returns Response

[0437] The generative AI model analyzes the prompt and emotion data, generates an appropriate response, and sends it back to the server. For example, the response "Today's weather is sunny." Input: Prompt sentence and emotion data. Output: Generated response sentence.

[0438] Step 6: Sending the Response and Displaying it to the User

[0439] The server sends the generated response to the device. The device receives the response and displays it to the user. For example, "Today's weather is sunny" is displayed on a smartphone screen. Input: The generated response text. Output: The response text displayed to the user.

[0440] Step 7: Gather feedback and update the algorithm

[0441] 7.1 Entering and Submitting Feedback

[0442] The user inputs feedback about the displayed response into the terminal. For example, the user inputs feedback such as "very satisfied." The terminal sends the feedback to the server. Input: User feedback. Output: Feedback data sent to the server.

[0443] 7.2 Receiving Feedback and Updating the Algorithm

[0444] The server stores the received feedback in a database and updates the selection algorithm, which will result in better prompt selection in future. Input: Submitted feedback data. Output: Updated selection algorithm.

[0445] (Application example 2)

[0446] 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."

[0447] Conventional generative AI systems did not take user emotions into consideration, limiting the quality of the user experience. Furthermore, selecting the optimal AI engine was difficult and required significant time and computational costs. This often led to dissatisfaction with the quality and timeliness of information provided.

[0448] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0449] In this invention, the server includes means for analyzing the input prompt and selecting the optimal generative AI engine and its version, means for recognizing the user's emotions and acquiring emotion data, and means for transmitting the acquired emotion data to the server and selecting the optimal generative AI engine together with the prompt analysis result, thereby making it possible to provide a more appropriate response that takes the user's emotions into consideration and improve the quality of the user experience.

[0450] An "input prompt" is a question or request that a user provides to a system.

[0451] A "generative artificial intelligence engine" refers to an algorithm or system for generating responses in natural language based on input prompts.

[0452] "Emotion data" refers to information about emotions extracted from a user's facial expressions, voice, text input, etc.

[0453] "Server" refers to a central processing unit for analyzing input data and providing generated responses to users.

[0454] "Prompt analysis result" refers to information obtained by the server analyzing the input prompt.

[0455] "Optimal engine selection" refers to the process of selecting the appropriate generative AI engine and its version based on prompt analysis results and emotion data.

[0456] "Feedback" refers to the opinions and ratings that users provide in response to responses they receive.

[0457] "Selection algorithm" refers to a set of calculation procedures and rules for selecting the optimal generative artificial intelligence engine and its version.

[0458] System Configuration

[0459] The system based on the present invention allows users to communicate with emotion recognition using generative AI models. This system includes the following elements:

[0460] 1. Parsing the input prompt

[0461] The server analyzes the prompts entered by the user through the terminal. The prompts contain questions or requests from the user, and performs syntactic and semantic analysis using natural language analysis techniques.

[0462] 2. Acquiring Emotion Data

[0463] The device acquires the user's emotion data related to the input prompt. The emotion data is recognized from the user's facial expressions, voice, text input, etc. using an emotion engine (e.g., a model using OpenCV or TENSORFLOW®).

[0464] 3. Selection of the optimal engine

[0465] The server selects the optimal generative AI engine and its version based on the prompt analysis results and emotion data. The server references past evaluation scores and feedback data and determines the optimal engine using an engine selection algorithm.

[0466] 4. Retrieving the generated response

[0467] The server sends prompts to the selected generative artificial intelligence engine and receives a generated response, which is provided in natural language format and includes an appropriate response to the user's prompt.

[0468] 5. Displaying the response to the user

[0469] The device displays the received responses to the user, and the responses are generated taking into account the user's emotional state, resulting in a satisfying interaction experience for the user.

[0470] 6. Gathering feedback and updating the algorithm

[0471] The device sends user feedback to the server, which then updates the engine selection algorithm based on the feedback to achieve better engine selection in the future.

[0472] Hardware and software used

[0473] Hardware

[0474] Devices (smartphones, smart glasses, head-mounted displays, etc.)

[0475] Server (high-performance data analysis server)

[0476] software

[0477] Emotion engine (OpenCV, TensorFlow, Keras)

[0478] HTTP request processing library (requests)

[0479] Generative AI engine (general-purpose generative AI model)

[0480] Specific examples

[0481] If a user types "What's the latest trending item?" and makes a surprised expression, the emotion engine will detect the "surprise" emotion. This prompt and emotion data will be sent to the server. The server will analyze the prompt and emotion data, select the optimal generative AI engine, and generate a response from the engine saying, "The latest trending item is a new smartwatch," which will be displayed to the user via their device.

[0482] Prompt Sentence Examples

[0483] What's the latest trending item?

[0484] Thus, the present invention is a system that enhances the user experience by recognizing the user's emotions and providing optimal responses based on them.

[0485] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0486] Step 1:

[0487] The user inputs a prompt into a device (such as a smartphone, smart glasses, or head-mounted display). The device then captures the user's input as text data. It also simultaneously records the user's facial expressions and voice, capturing them as emotion data. This allows the device to collect the prompt and emotion data as input.

[0488] Step 2:

[0489] The device sends the collected prompt sentences and emotion data to the server. This data is converted into JSON format and sent using an HTTP POST request. Specifically, the device converts the collected data into JSON and sends it to the server using the HTTP request library (requests).

[0490] Step 3:

[0491] The server analyzes the received prompt and emotion data. It parses the JSON data received as input and extracts the prompt and emotion data. It then performs syntactic and semantic analysis of the prompt using a natural language processing library (e.g., spaCy, NLTK). Meanwhile, it uses an emotion engine (e.g., TensorFlow, Keras) to classify the emotion data.

[0492] Step 4:

[0493] The server selects the optimal generative AI engine and its version based on the analysis results and emotional data. It uses past evaluation scores and feedback data to select the optimal engine using an algorithm. The algorithm uses the prompt type, the user's emotional state, and past engine performance data as inputs, and outputs the selection result.

[0494] Step 5:

[0495] The server sends a prompt to the selected generative AI engine and receives the generated response. Specifically, the server calls the selected engine's API and sends the prompt. The engine analyzes the prompt, generates an appropriate natural language response, and sends it back to the server. This response becomes the server's output.

[0496] Step 6:

[0497] The server sends the generated response to the terminal. The response data is converted to JSON format and sent to the terminal using an HTTP POST request. Specifically, the server converts the generated response to JSON and sends it to the terminal using the HTTP request library (requests).

[0498] Step 7:

[0499] The device displays the received response to the user. The user can then enter feedback such as satisfaction or additional comments. This feedback is also captured as text data and used in a later step. The device then converts the captured feedback into JSON format.

[0500] Step 8:

[0501] The device sends user feedback to the server. This feedback data is sent using an HTTP POST request. Specifically, the device converts the feedback data into JSON and sends it to the server using the HTTP request library (requests).

[0502] Step 9:

[0503] The server analyzes the received feedback and updates the engine selection algorithm. The analysis includes semantic analysis of the feedback comments and weighted averaging of the evaluation scores. Based on the analysis results, the engine selection algorithm is tuned to achieve more accurate selection in the future.

[0504] 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.

[0505] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0506] 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.

[0507] [Second embodiment]

[0508] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0509] 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.

[0510] 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).

[0511] 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.

[0512] 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.

[0513] 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).

[0514] 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. 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.

[0515] 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.

[0516] 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.

[0517] 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.

[0518] In the smart glasses 214, 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.

[0519] 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."

[0520] The present invention relates to a system that automatically selects the optimal generation AI engine and its version for a prompt entered by a user, thereby reducing calculation costs and time. This system is implemented through interactions between a server, a terminal, and a user.

[0521] System Overview

[0522] 1. Enter and submit the prompt

[0523] The user types any question or request into an input field on the device, and this prompt is sent from the device to the server.

[0524] 2. Accepting and parsing prompts

[0525] The server receives the prompts sent from the device and uses text analysis algorithms to analyze the prompt content, including the type of question, the depth of information required, and the complexity of the vocabulary.

[0526] 3. Selection of the optimal engine

[0527] The server selects the optimal generative AI engine and its version based on the analysis results and references pre-set evaluation scores and feedback data collected from past dialogue history.

[0528] 4. Sending prompts and generating responses

[0529] The server sends prompts to the selected generative AI engine, which generates an appropriate response based on the received prompt and sends it back to the server.

[0530] 5. Sending the response and displaying it to the user

[0531] The server sends the generated response to the terminal, and the terminal displays the received response to the user.

[0532] 6. Gathering feedback and updating the algorithm

[0533] Users can input their feedback on the displayed responses into the device, which then sends it to the server, which receives the feedback and updates the selection algorithm. This feedback is collected as user satisfaction and additional comments.

[0534] Specific examples

[0535] For simple questions

[0536] 1. The user enters the prompt

[0537] User: Type "What's the weather like today?"

[0538] 2. The device sends a prompt to the server

[0539] The terminal sends the entered prompt to the server.

[0540] 3. The server parses the prompt and selects an engine

[0541] The server analyzes the prompt and determines that it is a simple question. To keep costs down, it selects Generative AI Engine 3.5.

[0542] 4. Send a prompt to the engine and receive a response

[0543] The server sends a prompt to engine 3.5 and receives the response "The weather is sunny today."

[0544] 5. Send the response to the device and display it to the user

[0545] The server sends the response to the terminal, which displays the response to the user.

[0546] 6. Users provide feedback

[0547] User: Provides feedback saying "Very satisfied." The device sends the feedback to the server.

[0548] For complex technical questions

[0549] 1. The user enters the prompt

[0550] User: Type "Tell me more about regularization techniques for neural networks."

[0551] 2. The device sends a prompt to the server

[0552] The terminal sends the entered prompt to the server.

[0553] 3. The server parses the prompt and selects an engine

[0554] The server analyzes the prompt, understands that it is requesting detailed technical content, and selects a highly accurate generative AI engine4.

[0555] 4. Send a prompt to the engine and receive a response

[0556] The server sends a prompt to Engine 4 and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[0557] 5. Send the response to the device and display it to the user

[0558] The server sends the response to the terminal, which displays the response to the user.

[0559] 6. Users provide feedback

[0560] User: Provides feedback such as "somewhat satisfied." The device sends the feedback to the server.

[0561] This system allows users to receive fast and effective information without having to worry about the optimal AI engine, simply by entering prompts, and the system's accuracy is continuously improved through feedback.

[0562] The processing flow will be explained below.

[0563] Step 1:

[0564] A user types a prompt into an input field on the device. For example, the user types "Tell me about the latest AI technologies."

[0565] Step 2:

[0566] The terminal sends the entered prompt to the server, specifically as JSON-formatted data using an HTTP POST request.

[0567] Step 3:

[0568] The server receives the prompt sent from the device and parses it into the appropriate data format (e.g., JSON).

[0569] Step 4:

[0570] The server analyzes the prompt content using text analysis algorithms, including the type of question, the depth of information requested, and the complexity of the vocabulary.

[0571] Step 5:

[0572] Based on the analysis results, the server refers to a pre-set evaluation score and past feedback data to select the optimal generative AI engine and its version. For example, generative AI4 is selected for technical questions.

[0573] Step 6:

[0574] The server sends a prompt to the selected generative AI engine using an API request.

[0575] Step 7:

[0576] A generative AI engine receives the prompt and generates an appropriate response, which is sent back to the server.

[0577] Step 8:

[0578] The server then formats the response as needed, specifically converting the JSON data into HTML.

[0579] Step 9:

[0580] The server sends the formatted response to the device, which then sends it to the device as an HTTP response.

[0581] Step 10:

[0582] The device receives the response from the server and displays it to the user.

[0583] Step 11:

[0584] The user provides feedback on the generated response, including satisfaction and any additional comments.

[0585] Step 12:

[0586] The device sends the user feedback to the server, and the feedback data is sent in JSON format again.

[0587] Step 13:

[0588] The server receives user feedback and updates the selection algorithm, which improves the accuracy of engine selection from next time onwards.

[0589] Example 1

[0590] 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."

[0591] Currently, there are many generative AI engines available, each with different characteristics and performance. However, manually selecting which engine is best for a particular prompt is inefficient, and if an appropriate selection is not made, the quality of the generated response may decline. Furthermore, a method is needed to effectively utilize user feedback to update the engine selection algorithm and improve the system's accuracy. This will enable the provision of optimal responses while reducing computational costs and time.

[0592] 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.

[0593] In this invention, the server includes means for analyzing an input prompt and selecting an optimal generative AI engine and its version, means for sending the prompt to the selected generative AI engine and receiving a generated response, means for sending the generated response to a terminal and displaying it to the user, means for receiving feedback from the user and updating the selection algorithm of the generative AI engine, and user interface means for displaying the generated response to the user. This allows the user to quickly and effectively receive information simply by entering the prompt, without having to worry about selecting the optimal AI engine. Furthermore, the system's accuracy can be continuously improved by utilizing feedback.

[0594] An "input prompt" is text data that a user inputs into a terminal in the form of a question, request, or the like.

[0595] "Parsing" is the process of classifying and interpreting the content of the received input prompt using natural language processing algorithms.

[0596] A "generative artificial intelligence engine" is an artificial intelligence system capable of generating natural language responses based on a given input prompt.

[0597] A "version" is identification information that indicates different revisions or update states of a particular generative AI engine.

[0598] A "response" is a reply message in natural language generated by the selected artificial intelligence engine.

[0599] A "user" is a person who uses the system to enter questions or requests and receive responses generated by the AI ​​engine.

[0600] A "terminal" is a device (e.g., a PC, smartphone, etc.) on which a user enters prompts and receives and displays responses from the server.

[0601] A "text analysis algorithm" is an algorithm that uses natural language processing technology to analyze the meaning and content of input prompts.

[0602] The "evaluation score" is a numerical value used to evaluate the performance and appropriateness of the generative AI engine based on past dialogue history.

[0603] "Feedback data" refers to data that includes evaluations and comments on responses provided by users.

[0604] The "selection algorithm" is a method for selecting the optimal generative AI engine and version based on analysis results and feedback data.

[0605] "User interface means" refers to an interface (e.g., a web page, application screen, etc.) through which a user inputs a prompt and displays a response.

[0606] This invention relates to a system that automatically selects the optimal AI engine and its version for a prompt entered by a user, thereby reducing calculation costs and time. This system is implemented through interactions between a server, a terminal, and a user.

[0607] System configuration

[0608] 1. Hardware and Software Configuration

[0609] Server: Use a server or cloud service with powerful computing resources (e.g., AWS, Google Cloud).

[0610] Terminal: Refers to the device used by the user (e.g., PC, smartphone, tablet).

[0611] Natural Language Processing Software: Use natural language processing libraries such as NLTK or SpaCy for text analysis algorithms.

[0612] Generative artificial intelligence engine: Uses generative AI models such as GPT-3 and GPT-4.

[0613] Processing flow

[0614] 1. Enter and submit the prompt

[0615] The user types a question or request into an input field on the terminal, and this prompt is sent from the terminal to the server.

[0616] Example: A user types, "What's the weather like today?" or "Tell me more about regularization techniques for neural networks."

[0617] 2. Accepting and parsing prompts

[0618] The server receives the prompts sent from the device and uses text analysis algorithms to analyze the prompt content, including the type of question, the depth of information required, and the complexity of the vocabulary.

[0619] 3. Selection of the optimal engine

[0620] The server refers to the evaluation scores and feedback data collected from past dialogue history and selects the optimal generative AI engine and its version based on the analysis results.

[0621] 4. Sending prompts and generating responses

[0622] The server sends a prompt to the selected AI engine and receives the generated response. For example, the response "It's sunny today" is generated in response to the question "What's the weather like today?"

[0623] 5. Sending the response and displaying it to the user

[0624] The server sends the generated response to the terminal, which displays the response to the user.

[0625] 6. Gathering Feedback and Updating the Algorithm

[0626] The user inputs feedback on the displayed response into the terminal, which then sends the feedback to the server, which receives the feedback and uses it to update the selection algorithm.

[0627] Specific examples

[0628] For simple questions

[0629] 1. The user enters the prompt "What's the weather like today?"

[0630] 2. The terminal sends a prompt to the server.

[0631] 3. The server parses the prompt, determines that it is a simple question, and then selects the 3.5 version of the engine to keep costs down.

[0632] 4. The server sends a prompt to the engine and receives a response saying, "The weather is sunny today."

[0633] 5. The server sends the response to the terminal, which displays it to the user.

[0634] 6. The user gives feedback that they are "very satisfied," and the device sends the feedback to the server.

[0635] For complex technical questions

[0636] 1. The user types the prompt "Tell me more about regularization techniques in neural networks."

[0637] 2. The terminal sends a prompt to the server.

[0638] 3. The server parses the prompt, understands that you are requesting detailed technical information, and selects one of four high-precision engine versions.

[0639] 4. The server sends a prompt to the engine and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[0640] 5. The server sends the response to the terminal, which displays it to the user.

[0641] 6. The user gives feedback saying "somewhat satisfied," and the device sends the feedback to the server.

[0642] This system provides users with fast and accurate information without them having to worry about which AI engine is best suited to their query, and the system's accuracy can be improved by continuously utilizing user feedback.

[0643] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0644] Step 1: Enter the prompts

[0645] The user enters a question or request into an input field on the terminal.

[0646] As a concrete example, a user inputs text such as "What's the weather like today?" or "Please tell me more about regularization techniques for neural networks."

[0647] Input: Prompt sentence (e.g. "What's the weather like today?")

[0648] Output: prompt text formatted as text data

[0649] Step 2: Sending the prompt

[0650] The device sends the prompt entered by the user to the server by sending the prompt data to the server's API endpoint using an HTTP request.

[0651] Input: prompt text formatted as text data

[0652] Output: Prompt data sent to the server

[0653] Specific behavior: Issues an HTTP request over a network connection and sends prompt data.

[0654] Step 3: Accept the prompt

[0655] The server receives the prompt sent from the device and prepares it for analysis.

[0656] Input: Prompt data sent from the terminal

[0657] Output: Prompt data waiting to be parsed

[0658] Step 4: Parse the prompt

[0659] The server analyzes the prompt using text analysis algorithms (e.g., NLTK, SpaCy) to determine the type of question, the depth of information required, and lexical complexity.

[0660] Input: Prompt data waiting to be parsed

[0661] Output: Analysis results (e.g., question characteristics, difficulty, topic)

[0662] Specific operation: Tokenize the text, perform grammatical analysis, and semantic analysis, and extract each attribute.

[0663] Step 5: Selecting the optimal engine

[0664] Based on the analysis results, the server refers to the evaluation scores and feedback data collected from past dialogue history and selects the optimal generative AI engine and its version.

[0665] Input: Analysis results, evaluation scores, feedback data

[0666] Output: Selected generative AI engine and its version

[0667] What it does: Searches for suitable engines and versions in the database, and compares the rating scores and feedback to select the best engine.

[0668] Step 6: Sending prompts

[0669] The server sends a prompt to the selected AI engine and sends the prompt data using the AI ​​engine's API.

[0670] Input: Selected generation AI engine and its version, prompt data

[0671] Output: Prompt data sent to the generative AI engine

[0672] Specific behavior: Issues an API call and sends a prompt to the engine.

[0673] Step 7: Generate the Response

[0674] The generative artificial intelligence engine generates appropriate responses based on the prompts and sends them back to the server.

[0675] Input: prompt data

[0676] Output: The generated response

[0677] Specific behavior: Generates a natural language response using the internal model. Example: In response to the prompt "What's the weather like today?", generates "The weather is sunny today."

[0678] Step 8: Sending the Response

[0679] The server sends the generated response to the terminal.

[0680] Input: The generated response

[0681] Output: Response data sent to the device

[0682] Specific operation: The generated text is sent back to the terminal as an HTTP response.

[0683] Step 9: View the response

[0684] The terminal displays the received response to the user.

[0685] Input: Response data received from the server

[0686] Output: Response displayed on screen

[0687] Specific behavior: Display text on the user interface. Example: Display "Today's weather is sunny" on the screen.

[0688] Step 10: Provide feedback

[0689] The user inputs feedback on the displayed response into the terminal.

[0690] Input: Rating and comments on the response

[0691] Output: Input feedback data

[0692] Specific actions: Clicking the rating button or typing in the comment box and submitting.

[0693] Step 11: Submit your feedback

[0694] The terminal sends the feedback to the server.

[0695] Input: Feedback data

[0696] Output: Feedback data sent to the server

[0697] Specific operation: Sends feedback data to the server as an HTTP POST request.

[0698] Step 12: Update the algorithm

[0699] The server updates its selection algorithm based on the feedback it receives.

[0700] Input: Feedback data

[0701] Output: Updated election algorithm

[0702] What it does: The feedback data is stored in a database and used as training data for machine learning models, including periodic retraining.

[0703] (Application example 1)

[0704] 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."

[0705] Modern content delivery services require the ability to quickly provide optimal responses to a variety of prompts entered by users. Conventional systems often use the same generative AI engine for all requests, resulting in problems such as high-cost engines always being used or an inability to select an appropriate engine. Another issue is that user feedback is rarely reflected in the system, making it difficult to improve service quality. The present invention aims to solve these problems and provide a system for effectively providing customized content to users.

[0706] 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.

[0707] In this invention, the server includes means for analyzing an input prompt and selecting an optimal generating AI engine and its version, means for sending the prompt to the selected generating AI engine and receiving a generated response, means for sending the generated response to the user, means for receiving feedback from the user and updating the selection algorithm of the generating AI engine, and means for providing customized content based on the generated response. This makes it possible to select an optimal engine for the user's input prompt and provide customized content efficiently and effectively.

[0708] An "input prompt" is a request or question that a user sends to a system.

[0709] A "generative artificial intelligence engine" is an artificial intelligence algorithm that generates appropriate responses based on user prompts.

[0710] "Version" is a number or state that represents a different release or improvement of the Generative AI Engine.

[0711] "Server" is the central system that processes user requests, selects the appropriate engine, and generates and sends the response.

[0712] "User" means an end user who utilizes the system to enter prompts and receive responses and content.

[0713] A "response" is a response generated by a generative artificial intelligence engine based on a prompt.

[0714] "Feedback" refers to the opinions and ratings that users provide in response to a response.

[0715] A "selection algorithm" is a process or method for selecting the optimal generative AI engine and its version.

[0716] "Customized Content" means information and materials that are optimized for a particular user based on the user's prompts and feedback.

[0717] The present invention is a system that automatically selects the optimal generative AI engine and its version for a prompt input by a user, and provides customized content based on the selection result. This system is implemented through interactions between a server, a terminal, and a user.

[0718] The system program is structured as follows: First, the user uses the terminal to input a prompt. This prompt is in text format, such as "I would like to read an article about the latest AI technology." The terminal then sends this input prompt to the server.

[0719] The server then analyzes the received prompt using a text analysis algorithm that evaluates the prompt's content, vocabulary complexity, and required information depth. Based on the analysis results, the server selects the optimal generative AI engine and its version. The selection is made using an algorithm that references pre-set evaluation scores and past feedback data.

[0720] A prompt is sent to the selected generative AI engine, causing it to generate an appropriate response. For example, this response might be, "The following are recommended articles about the latest AI technology: 1. The cutting edge of AI research 2. The evolution of deep learning 3. The future of coexistence between AI and humans." The server receives this generated response and sends it to the device.

[0721] The terminal displays the received response to the user, who can then provide feedback on the displayed response, such as "very satisfied" or "somewhat satisfied." This feedback is then sent back to the server from the terminal.

[0722] The server updates the generative AI engine selection algorithm based on the received feedback. This algorithm update will enable more accurate engine selection in the future. For example, if users provide a large number of "very satisfied" feedback, the server recognizes that the engine version is highly rated and reflects this information in the next selection.

[0723] The hardware required to realize this system includes the user's device (smartphone, PC, etc.), server, and internet communication environment. The software includes a text analysis algorithm, an evaluation score and feedback data management algorithm, a generative AI engine, and a server program that integrates them.

[0724] In this way, the present invention allows users to select the most appropriate engine for their input prompts and receive customized content efficiently and effectively, which not only improves user satisfaction but also contributes to cost and resource optimization.

[0725] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0726] Step 1:

[0727] The user inputs a prompt using the terminal. For example, the user inputs text such as "I would like to read an article about the latest AI technology." The input prompt is sent from the terminal to the server.

[0728] Input: A prompt from the user (e.g., "I want to read an article about the latest AI technology")

[0729] Output: Prompt sent from terminal to server

[0730] Step 2:

[0731] The server receives the prompt text sent from the device, inputs it into a text analysis algorithm, and analyzes the content of the prompt. As a result of the analysis, the type of question, the depth of information required, the complexity of vocabulary, etc. are extracted.

[0732] Input: Prompt sent from the terminal

[0733] Output: Analysis of the prompt (type of question, depth of information required, vocabulary complexity, etc.)

[0734] Step 3:

[0735] The server selects the optimal generative AI engine and its version based on the analyzed prompts, taking into account pre-set evaluation scores and past feedback data.

[0736] Input: Parsed result of prompt statement

[0737] Output: Selected generative AI engine and its version

[0738] Step 4:

[0739] The server sends the prompt sentence to the selected AI engine to generate an appropriate response. The AI ​​engine generates a response based on the prompt sentence and sends it back to the server.

[0740] Input: Selected artificial intelligence engine, prompt

[0741] Output: The generated response

[0742] Step 5:

[0743] The server receives the generated response and sends it to the terminal, which displays the received response to the user.

[0744] Input: The generated response

[0745] Output: Sending the response to the terminal and displaying it to the user

[0746] Step 6:

[0747] The user provides feedback on the displayed response. For example, the user inputs an evaluation such as "very satisfied" into the terminal. The terminal then transmits the feedback to the server.

[0748] Input: User feedback (e.g. "Very satisfied")

[0749] Output: Sending feedback from the device to the server

[0750] Step 7:

[0751] The server updates the generative AI engine's selection algorithm based on the received feedback, allowing it to select a more appropriate engine and version for the next prompt.

[0752] Input: User feedback

[0753] Output: Updated election algorithm

[0754] In this way, the present invention realizes a system that selects the most suitable generative artificial intelligence engine in response to a user's prompts and provides customized content efficiently and effectively.

[0755] 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.

[0756] This invention relates to a system that automatically selects the optimal generative AI engine and its version for a prompt entered by a user, and further combines it with an emotion engine that recognizes the user's emotions, thereby reducing calculation costs and time and providing a more advanced response. This system is implemented through interactions between four parties: a server, a terminal, an emotion engine, and a user.

[0757] System Overview

[0758] 1. Input prompts and emotion recognition

[0759] Users can enter any question or request into the device's input field, and the emotion engine will simultaneously recognize emotions from the user's input, facial expressions, and voice.

[0760] 2. Sending prompts and emotion data

[0761] The device sends the entered prompt and recognized emotion data to the server, which is sent as JSON formatted data using an HTTP POST request.

[0762] 3. Accepting and parsing prompts

[0763] The server receives the prompt and emotion data sent from the device, analyzes the received data in an appropriate data format, and recognizes the prompt content and the user's emotion.

[0764] 4. Selection of the optimal engine

[0765] The server selects the optimal AI engine and its version based on the prompt analysis results and emotional data, referring to pre-set evaluation scores and past feedback data. Emotional data is also taken into consideration, so the engine that best suits the user's emotional state is selected.

[0766] 5. Sending prompts and generating responses

[0767] The server sends a prompt to the selected generative AI engine, which generates an appropriate response based on the received prompt and emotion data and sends it back to the server.

[0768] 6. Sending the response and displaying it to the user

[0769] The server then sends the generated response to the device, which then displays the received response to the user, including feedback based on the emotion data.

[0770] 7. Gathering feedback and updating the algorithm

[0771] Users can provide feedback on the displayed responses to their devices, which then send it to the server. The server receives the feedback and updates the selection algorithm. This feedback is collected as user satisfaction, additional comments, and emotional data.

[0772] Specific examples

[0773] For simple questions

[0774] 1. The user enters the prompt

[0775] User: As soon as the user types "What's the weather like today?", the device detects the emotion "happiness" from the user's facial expression.

[0776] 2. The device sends prompts and emotion data to the server

[0777] The terminal transmits the input prompt and the emotion data of "joy" to the server.

[0778] 3. The server analyzes the prompt and emotion data and selects the engine.

[0779] The server analyzes the prompt and emotion data, determines that it is a simple question, and selects Generative AI Engine 3.5.

[0780] 4. Send a prompt to the engine and receive a response

[0781] The server sends a prompt to engine 3.5 and receives the response "The weather is sunny today."

[0782] 5. Send the response to the device and display it to the user

[0783] The server sends the response to the terminal, which displays the response to the user.

[0784] 6. Users provide feedback

[0785] The user provides feedback such as "very satisfied," and the device sends this feedback to the server.

[0786] For complex technical questions

[0787] 1. The user enters the prompt

[0788] User: Type "Please tell me more about neural network regularization techniques" and in the process detect the emotion "interested" from the user's voice.

[0789] 2. The device sends prompts and emotion data to the server

[0790] The terminal transmits the input prompt and the emotion data of "interest" to the server.

[0791] 3. The server analyzes the prompt and emotion data and selects the engine.

[0792] The server analyzes the prompt and emotion data, determines that detailed technical content is required, and then selects a generative artificial intelligence engine 4.

[0793] 4. Send a prompt to the engine and receive a response

[0794] The server sends a prompt to Engine 4 and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[0795] 5. Send the response to the device and display it to the user

[0796] The server sends the response to the terminal, which displays the response to the user.

[0797] 6. Users provide feedback

[0798] The user provides feedback such as "somewhat satisfied," and the device sends this feedback to the server.

[0799] This system allows users to receive prompt and emotional information quickly and effectively without having to worry about the optimal AI engine, and the system's accuracy is continuously improved through feedback.

[0800] The processing flow will be explained below.

[0801] Step 1:

[0802] A user types a prompt into an input field on the device. For example, the user types "Tell me about the latest AI technologies."

[0803] Step 2:

[0804] The emotion engine analyzes the user's input, facial expressions, voice, etc. to recognize the user's emotions. For example, it can detect "interest" as an emotion.

[0805] Step 3:

[0806] The device sends the input prompt and emotion data to the server as JSON format data using an HTTP POST request.

[0807] Step 4:

[0808] The server receives the prompt and emotion data sent from the device and parses the received data into an appropriate data format (e.g., JSON).

[0809] Step 5:

[0810] The server analyzes the prompt content and emotional data using text analysis and emotion recognition algorithms, including the type of question, the depth of information requested, vocabulary complexity, and the user's emotions.

[0811] Step 6:

[0812] Based on the analysis results, the server refers to the preset evaluation score, past feedback data, and emotion data to select the optimal generative AI engine and its version. For example, if the question is technical and the emotion of "interest" is detected, generative AI engine 4 will be selected.

[0813] Step 7:

[0814] The server sends the prompt and emotion data to the selected generative AI engine using an API request.

[0815] Step 8:

[0816] A generative AI engine receives the prompt and emotion data and generates an appropriate response, which is then sent back to the server.

[0817] Step 9:

[0818] The server then formats the response as needed, specifically converting the JSON data into HTML.

[0819] Step 10:

[0820] The server sends the formatted response to the device, which then sends it to the device as an HTTP response.

[0821] Step 11:

[0822] The device receives the response from the server and displays it to the user.

[0823] Step 12:

[0824] Users provide feedback on the generated responses, including satisfaction, additional comments, and sentiment data.

[0825] Step 13:

[0826] The device sends the user feedback to the server, and the feedback data is sent in JSON format again.

[0827] Step 14:

[0828] The server receives user feedback and updates the selection algorithm, which improves the accuracy of engine selection from next time onwards.

[0829] Example 2

[0830] 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."

[0831] Conventional generative AI systems require manual selection of engines and versions to generate optimal responses to input prompts, resulting in increased computational costs and time for data processing and selection. Furthermore, it is difficult to automatically generate responses based on user emotions, and systems for improving user satisfaction are lacking. Furthermore, there is an insufficient mechanism for continuous improvement by aggregating and incorporating user feedback.

[0832] 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.

[0833] In this invention, the server includes means for receiving an input prompt from a user and recognizing the user's emotional data, means for analyzing the input prompt and the emotional data and selecting an optimal generating AI engine and its version, and means for sending the prompt to the selected generating AI engine and receiving a generated response, thereby enabling the selection of an optimal engine according to the user's emotional state and the rapid generation of a response.

[0834] "User" means an individual or entity that utilizes the System to enter prompts and receive information or services.

[0835] An "input prompt" is text data such as a question or request that a user inputs to a system.

[0836] "Emotion data" refers to data that indicates the emotional state of the user as recognized from facial expressions, voice, etc.

[0837] A "generative artificial intelligence engine" is an artificial intelligence model for generating appropriate responses based on user input prompts.

[0838] "Version" refers to different releases or revisions of the Generative Artificial Intelligence Engine.

[0839] "Analysis" refers to the process of verifying input prompts, emotional data, past feedback data, etc., and processing the data to select the optimal generative AI engine and its version.

[0840] A "response" is response data generated by a generative artificial intelligence engine in response to a user's input prompt.

[0841] "Feedback" refers to reaction data such as ratings and comments provided by users in response to generated responses.

[0842] A "selection algorithm" is a computational method or set of rules for analyzing prompts and emotional data and selecting the optimal generative AI engine and its version.

[0843] The "system" is a collection of components that executes a series of processes: receiving prompts and emotional data from users, selecting the optimal generative artificial intelligence engine to generate a response, aggregating feedback, and updating the algorithm.

[0844] This invention is a system that automatically selects the optimal generative AI engine and its version for the prompt entered by the user, and further combines it with an emotion engine that recognizes the user's emotions, thereby reducing calculation costs and time and providing a more advanced response. This system is implemented through interactions between elements including the user, terminal, server, and emotion engine.

[0845] Explanation of program processing

[0846] 1. User prompt input and emotion recognition

[0847] Users enter any question or request into an input field via a device such as a smartphone or computer. At the same time, the emotion engine uses the device's camera and microphone to analyze the user's facial expressions and voice and recognize their emotions. For example, when a user types "What's the weather like today?", the device detects the user's smile and identifies it as the emotion of "joy."

[0848] 2. Sending prompts and emotion data

[0849] The device compiles the input prompt and recognized emotion data in JSON format and sends it to the server using an HTTP POST request, specifically using the fetch API or XMLHttpRequest.

[0850] 3. Accepting and parsing prompts

[0851] The server receives the prompt and emotion data sent from the device and parses it into the appropriate format. It uses the JSON.parse method to parse the data and stores the prompt and emotion data in separate variables. During this process, it also references past feedback data and evaluation scores stored in a database (e.g., MySQL or PostgreSQL).

[0852] 4. Selection of the optimal engine

[0853] The server selects the optimal AI engine and its version based on the prompt analysis results and emotion data, along with evaluation scores and feedback data. For example, Engine 3.5 is selected for simple questions, while Engine 4 is selected for detailed technical questions.

[0854] 5. Sending prompts and generating responses

[0855] The server sends a prompt to the selected generative AI engine, which then generates an appropriate response. The generative AI engine takes into account the received prompt and emotional data to create an appropriate response and sends it back to the server. For example, a response might be generated: "The weather is sunny today."

[0856] 6. Sending the response and displaying it to the user

[0857] The server sends the generated response to the device, which then displays it to the user. The device screen displays "Today's weather is sunny."

[0858] 7. Gathering feedback and updating the algorithm

[0859] The user inputs feedback on the displayed response into the terminal, including options such as "very satisfied" and "somewhat satisfied." The terminal then transmits the user's feedback to the server.

[0860] The server stores the received feedback in a database and updates the selection algorithm based on that feedback, which will result in better engine selection in the future.

[0861] Specific examples

[0862] For simple questions

[0863] 1. User: Enter "What's the weather like today?" The device detects "happiness" from the user's facial expression.

[0864] 2. Device: Sends the prompt and the "happiness" emotion data to the server.

[0865] 3. Server: Analyzes prompt and sentiment data and selects engine 3.5.

[0866] 4. Server: Sends a prompt to Engine 3.5 and receives the response "The weather is sunny today."

[0867] 5. Server: Sends the response to the terminal, which displays it to the user.

[0868] 6. User: Gives feedback that he is "very satisfied," and the device sends the feedback to the server.

[0869] For complex technical questions

[0870] 1. User: Type "Please tell me more about neural network regularization techniques." The device detects "interests" from the user's voice.

[0871] 2. Device: Sends prompt and "interest" emotion data to the server.

[0872] 3. Server: Analyzes prompt and emotion data and selects Engine 4.

[0873] 4. Server: Sends a prompt to Engine 4 and receives a response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[0874] 5. Server: Sends the response to the terminal, which displays it to the user.

[0875] 6. User: Gives feedback that they are "somewhat satisfied," and the device sends that feedback to the server.

[0876] The system selects the most suitable AI engine and provides fast and effective information by simply inputting prompts and emotional data, while the accuracy of the system is continuously improved through feedback.

[0877] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0878] Divide the processing flow of the system program into processing steps

[0879] Step 1: User prompt input and emotion recognition

[0880] 1.1 Prompt Input

[0881] 1.2 Emotion recognition

[0882] Step 2: Sending prompts and emotion data

[0883] Step 3: Server accepts and parses prompt

[0884] 3.1 Receiving prompts and emotion data

[0885] 3.2 Prompt and Emotion Data Analysis

[0886] Step 4: Selecting the optimal engine

[0887] 4.1 Referencing evaluation scores and feedback data

[0888] 4.2 Selecting the optimal engine and version

[0889] Step 5: Sending prompts and generating responses

[0890] 5.1 Sending Prompts to the Selection Engine

[0891] 5.2 Engine Generates and Returns Response

[0892] Step 6: Sending the Response and Displaying it to the User

[0893] Step 7: Gather feedback and update the algorithm

[0894] 7.1 Entering and Submitting Feedback

[0895] 7.2 Receiving Feedback and Updating the Algorithm

[0896] Specific explanation of each processing step

[0897] Step 1: User prompt input and emotion recognition

[0898] 1.1 Prompt Input

[0899] The user writes a question or request into the input field of the terminal. For example, "What's the weather like today?" Input: The text data entered by the user into the terminal. Output: The prompt text entered.

[0900] 1.2 Emotion recognition

[0901] The device uses the built-in camera and microphone to collect the user's facial expressions and voice, and the emotion engine analyzes their emotions in real time. For example, if the user smiles, the emotion "joy" is detected. Input: User's facial and voice data. Output: Analyzed emotion data (e.g., "joy").

[0902] Step 2: Sending prompts and emotion data

[0903] The device collects the input prompt and emotion data and sends them to the server in JSON format. Specifically, it makes an HTTP POST request using the fetch API or XMLHttpRequest. Input: Prompt text and emotion data. Output: JSON data sent to the server.

[0904] Step 3: Server accepts and parses prompt

[0905] 3.1 Receiving prompts and emotion data

[0906] The server receives the prompt and emotion data sent from the device. Input: JSON data sent from the device. Output: Received prompt and emotion data.

[0907] 3.2 Prompt and Emotion Data Analysis

[0908] The server parses the prompt and emotion data using the JSON.parse method and stores them in separate variables. Input: Received prompt and emotion data. Output: Parsed prompt and emotion data.

[0909] Step 4: Selecting the optimal engine

[0910] 4.1 Referencing evaluation scores and feedback data

[0911] The server references past feedback data and evaluation scores from a database. For example, queries are performed using MySQL or PostgreSQL. Input: prompt text and emotion data. Output: evaluation scores and feedback data required to select the optimal engine.

[0912] 4.2 Selecting the optimal engine and version

[0913] The server selects the optimal generative AI engine and its version based on the prompt, emotion data, evaluation scores, and feedback data. Input: Evaluation scores and feedback data. Output: Selected optimal engine and its version.

[0914] Step 5: Sending prompts and generating responses

[0915] 5.1 Sending Prompts to the Selection Engine

[0916] The server sends a prompt to the selected engine. Specifically, it sends an HTTP request to the engine's API endpoint. Input: Selected engine and prompt text. Output: Sending prompt data to the generative AI engine.

[0917] 5.2 Engine Generates and Returns Response

[0918] The generative AI model analyzes the prompt and emotion data, generates an appropriate response, and sends it back to the server. For example, the response "Today's weather is sunny." Input: Prompt sentence and emotion data. Output: Generated response sentence.

[0919] Step 6: Sending the Response and Displaying it to the User

[0920] The server sends the generated response to the device. The device receives the response and displays it to the user. For example, "Today's weather is sunny" is displayed on a smartphone screen. Input: The generated response text. Output: The response text displayed to the user.

[0921] Step 7: Gather feedback and update the algorithm

[0922] 7.1 Entering and Submitting Feedback

[0923] The user inputs feedback about the displayed response into the terminal. For example, the user inputs feedback such as "very satisfied." The terminal sends the feedback to the server. Input: User feedback. Output: Feedback data sent to the server.

[0924] 7.2 Receiving Feedback and Updating the Algorithm

[0925] The server stores the received feedback in a database and updates the selection algorithm, which will result in better prompt selection in future. Input: Submitted feedback data. Output: Updated selection algorithm.

[0926] (Application example 2)

[0927] 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."

[0928] Conventional generative AI systems did not take user emotions into consideration, limiting the quality of the user experience. Furthermore, selecting the optimal AI engine was difficult and required significant time and computational costs. This often led to dissatisfaction with the quality and timeliness of information provided.

[0929] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0930] In this invention, the server includes means for analyzing the input prompt and selecting the optimal generative AI engine and its version, means for recognizing the user's emotions and acquiring emotion data, and means for transmitting the acquired emotion data to the server and selecting the optimal generative AI engine together with the prompt analysis result, thereby making it possible to provide a more appropriate response that takes the user's emotions into consideration and improve the quality of the user experience.

[0931] An "input prompt" is a question or request that a user provides to a system.

[0932] A "generative artificial intelligence engine" refers to an algorithm or system for generating responses in natural language based on input prompts.

[0933] "Emotion data" refers to information about emotions extracted from a user's facial expressions, voice, text input, etc.

[0934] "Server" refers to a central processing unit for analyzing input data and providing generated responses to users.

[0935] "Prompt analysis result" refers to information obtained by the server analyzing the input prompt.

[0936] "Optimal engine selection" refers to the process of selecting the appropriate generative AI engine and its version based on prompt analysis results and emotion data.

[0937] "Feedback" refers to the opinions and ratings that users provide in response to responses they receive.

[0938] "Selection algorithm" refers to a set of calculation procedures and rules for selecting the optimal generative artificial intelligence engine and its version.

[0939] System Configuration

[0940] The system based on the present invention allows users to communicate with emotion recognition using generative AI models. This system includes the following elements:

[0941] 1. Parsing the input prompt

[0942] The server analyzes the prompts entered by the user through the terminal. The prompts contain questions or requests from the user, and performs syntactic and semantic analysis using natural language analysis techniques.

[0943] 2. Acquiring Emotion Data

[0944] The device acquires the user's emotion data related to the input prompt. The emotion data is recognized from the user's facial expressions, voice, text input, etc. using an emotion engine (e.g., a model using OpenCV or TensorFlow).

[0945] 3. Selection of the optimal engine

[0946] The server selects the optimal generative AI engine and its version based on the prompt analysis results and emotion data. The server references past evaluation scores and feedback data and determines the optimal engine using an engine selection algorithm.

[0947] 4. Retrieving the generated response

[0948] The server sends prompts to the selected generative artificial intelligence engine and receives a generated response, which is provided in natural language format and includes an appropriate response to the user's prompt.

[0949] 5. Displaying the response to the user

[0950] The device displays the received responses to the user, and the responses are generated taking into account the user's emotional state, resulting in a satisfying interaction experience for the user.

[0951] 6. Gathering feedback and updating the algorithm

[0952] The device sends user feedback to the server, which then updates the engine selection algorithm based on the feedback to achieve better engine selection in the future.

[0953] Hardware and software used

[0954] Hardware

[0955] Devices (smartphones, smart glasses, head-mounted displays, etc.)

[0956] Server (high-performance data analysis server)

[0957] software

[0958] Emotion engine (OpenCV, TensorFlow, Keras)

[0959] HTTP request processing library (requests)

[0960] Generative AI engine (general-purpose generative AI model)

[0961] Specific examples

[0962] If a user types "What's the latest trending item?" and makes a surprised expression, the emotion engine will detect the "surprise" emotion. This prompt and emotion data will be sent to the server. The server will analyze the prompt and emotion data, select the optimal generative AI engine, and generate a response from the engine saying, "The latest trending item is a new smartwatch," which will be displayed to the user via their device.

[0963] Prompt Sentence Examples

[0964] What's the latest trending item?

[0965] Thus, the present invention is a system that enhances the user experience by recognizing the user's emotions and providing optimal responses based on them.

[0966] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0967] Step 1:

[0968] The user inputs a prompt into a device (such as a smartphone, smart glasses, or head-mounted display). The device then captures the user's input as text data. It also simultaneously records the user's facial expressions and voice, capturing them as emotion data. This allows the device to collect the prompt and emotion data as input.

[0969] Step 2:

[0970] The device sends the collected prompt sentences and emotion data to the server. This data is converted into JSON format and sent using an HTTP POST request. Specifically, the device converts the collected data into JSON and sends it to the server using the HTTP request library (requests).

[0971] Step 3:

[0972] The server analyzes the received prompt and emotion data. It parses the JSON data received as input and extracts the prompt and emotion data. It then performs syntactic and semantic analysis of the prompt using a natural language processing library (e.g., spaCy, NLTK). Meanwhile, it uses an emotion engine (e.g., TensorFlow, Keras) to classify the emotion data.

[0973] Step 4:

[0974] The server selects the optimal generative AI engine and its version based on the analysis results and emotional data. It uses past evaluation scores and feedback data to select the optimal engine using an algorithm. The algorithm uses the prompt type, the user's emotional state, and past engine performance data as inputs, and outputs the selection result.

[0975] Step 5:

[0976] The server sends a prompt to the selected generative AI engine and receives the generated response. Specifically, the server calls the selected engine's API and sends the prompt. The engine analyzes the prompt, generates an appropriate natural language response, and sends it back to the server. This response becomes the server's output.

[0977] Step 6:

[0978] The server sends the generated response to the terminal. The response data is converted to JSON format and sent to the terminal using an HTTP POST request. Specifically, the server converts the generated response to JSON and sends it to the terminal using the HTTP request library (requests).

[0979] Step 7:

[0980] The device displays the received response to the user. The user can then enter feedback such as satisfaction or additional comments. This feedback is also captured as text data and used in a later step. The device then converts the captured feedback into JSON format.

[0981] Step 8:

[0982] The device sends user feedback to the server. This feedback data is sent using an HTTP POST request. Specifically, the device converts the feedback data into JSON and sends it to the server using the HTTP request library (requests).

[0983] Step 9:

[0984] The server analyzes the received feedback and updates the engine selection algorithm. The analysis includes semantic analysis of the feedback comments and weighted averaging of the evaluation scores. Based on the analysis results, the engine selection algorithm is tuned to achieve more accurate selection in the future.

[0985] 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.

[0986] 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.

[0987] 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.

[0988] [Third embodiment]

[0989] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0990] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0991] 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).

[0992] 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.

[0993] 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.

[0994] 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).

[0995] 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. 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.

[0996] 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.

[0997] 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.

[0998] 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.

[0999] 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.

[1000] 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."

[1001] The present invention relates to a system that automatically selects the optimal generation AI engine and its version for a prompt entered by a user, thereby reducing calculation costs and time. This system is implemented through interactions between a server, a terminal, and a user.

[1002] System Overview

[1003] 1. Enter and submit the prompt

[1004] The user types any question or request into an input field on the device, and this prompt is sent from the device to the server.

[1005] 2. Accepting and parsing prompts

[1006] The server receives the prompts sent from the device and uses text analysis algorithms to analyze the prompt content, including the type of question, the depth of information required, and the complexity of the vocabulary.

[1007] 3. Selection of the optimal engine

[1008] The server selects the optimal generative AI engine and its version based on the analysis results and references pre-set evaluation scores and feedback data collected from past dialogue history.

[1009] 4. Sending prompts and generating responses

[1010] The server sends prompts to the selected generative AI engine, which generates an appropriate response based on the received prompt and sends it back to the server.

[1011] 5. Sending the response and displaying it to the user

[1012] The server sends the generated response to the terminal, and the terminal displays the received response to the user.

[1013] 6. Gathering feedback and updating the algorithm

[1014] Users can input their feedback on the displayed responses into the device, which then sends it to the server, which receives the feedback and updates the selection algorithm. This feedback is collected as user satisfaction and additional comments.

[1015] Specific examples

[1016] For simple questions

[1017] 1. The user enters the prompt

[1018] User: Type "What's the weather like today?"

[1019] 2. The device sends a prompt to the server

[1020] The terminal sends the entered prompt to the server.

[1021] 3. The server parses the prompt and selects an engine

[1022] The server analyzes the prompt and determines that it is a simple question. To keep costs down, it selects Generative AI Engine 3.5.

[1023] 4. Send a prompt to the engine and receive a response

[1024] The server sends a prompt to engine 3.5 and receives the response "The weather is sunny today."

[1025] 5. Send the response to the device and display it to the user

[1026] The server sends the response to the terminal, which displays the response to the user.

[1027] 6. Users provide feedback

[1028] User: Provides feedback saying "Very satisfied." The device sends the feedback to the server.

[1029] For complex technical questions

[1030] 1. The user enters the prompt

[1031] User: Type "Tell me more about regularization techniques for neural networks."

[1032] 2. The device sends a prompt to the server

[1033] The terminal sends the entered prompt to the server.

[1034] 3. The server parses the prompt and selects an engine

[1035] The server analyzes the prompt, understands that it is requesting detailed technical content, and selects a highly accurate generative AI engine4.

[1036] 4. Send a prompt to the engine and receive a response

[1037] The server sends a prompt to Engine 4 and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[1038] 5. Send the response to the device and display it to the user

[1039] The server sends the response to the terminal, which displays the response to the user.

[1040] 6. Users provide feedback

[1041] User: Provides feedback such as "somewhat satisfied." The device sends the feedback to the server.

[1042] This system allows users to receive fast and effective information without having to worry about the optimal AI engine, simply by entering prompts, and the system's accuracy is continuously improved through feedback.

[1043] The processing flow will be explained below.

[1044] Step 1:

[1045] A user types a prompt into an input field on the device. For example, the user types "Tell me about the latest AI technologies."

[1046] Step 2:

[1047] The terminal sends the entered prompt to the server, specifically as JSON-formatted data using an HTTP POST request.

[1048] Step 3:

[1049] The server receives the prompt sent from the device and parses it into the appropriate data format (e.g., JSON).

[1050] Step 4:

[1051] The server analyzes the prompt content using text analysis algorithms, including the type of question, the depth of information requested, and the complexity of the vocabulary.

[1052] Step 5:

[1053] Based on the analysis results, the server refers to a pre-set evaluation score and past feedback data to select the optimal generative AI engine and its version. For example, generative AI4 is selected for technical questions.

[1054] Step 6:

[1055] The server sends a prompt to the selected generative AI engine using an API request.

[1056] Step 7:

[1057] A generative AI engine receives the prompt and generates an appropriate response, which is sent back to the server.

[1058] Step 8:

[1059] The server then formats the response as needed, specifically converting the JSON data into HTML.

[1060] Step 9:

[1061] The server sends the formatted response to the device, which then sends it to the device as an HTTP response.

[1062] Step 10:

[1063] The device receives the response from the server and displays it to the user.

[1064] Step 11:

[1065] The user provides feedback on the generated response, including satisfaction and any additional comments.

[1066] Step 12:

[1067] The device sends the user feedback to the server, and the feedback data is sent in JSON format again.

[1068] Step 13:

[1069] The server receives user feedback and updates the selection algorithm, which improves the accuracy of engine selection from next time onwards.

[1070] Example 1

[1071] 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."

[1072] Currently, there are many generative AI engines available, each with different characteristics and performance. However, manually selecting which engine is best for a particular prompt is inefficient, and if an appropriate selection is not made, the quality of the generated response may decline. Furthermore, a method is needed to effectively utilize user feedback to update the engine selection algorithm and improve the system's accuracy. This will enable the provision of optimal responses while reducing computational costs and time.

[1073] 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.

[1074] In this invention, the server includes means for analyzing an input prompt and selecting an optimal generative AI engine and its version, means for sending the prompt to the selected generative AI engine and receiving a generated response, means for sending the generated response to a terminal and displaying it to the user, means for receiving feedback from the user and updating the selection algorithm of the generative AI engine, and user interface means for displaying the generated response to the user. This allows the user to quickly and effectively receive information simply by entering the prompt, without having to worry about selecting the optimal AI engine. Furthermore, the system's accuracy can be continuously improved by utilizing feedback.

[1075] An "input prompt" is text data that a user inputs into a terminal in the form of a question, request, or the like.

[1076] "Parsing" is the process of classifying and interpreting the content of the received input prompt using natural language processing algorithms.

[1077] A "generative artificial intelligence engine" is an artificial intelligence system capable of generating natural language responses based on a given input prompt.

[1078] A "version" is identification information that indicates different revisions or update states of a particular generative AI engine.

[1079] A "response" is a reply message in natural language generated by the selected artificial intelligence engine.

[1080] A "user" is a person who uses the system to enter questions or requests and receive responses generated by the AI ​​engine.

[1081] A "terminal" is a device (e.g., a PC, smartphone, etc.) on which a user enters prompts and receives and displays responses from the server.

[1082] A "text analysis algorithm" is an algorithm that uses natural language processing technology to analyze the meaning and content of input prompts.

[1083] The "evaluation score" is a numerical value used to evaluate the performance and appropriateness of the generative AI engine based on past dialogue history.

[1084] "Feedback data" refers to data that includes evaluations and comments on responses provided by users.

[1085] The "selection algorithm" is a method for selecting the optimal generative AI engine and version based on analysis results and feedback data.

[1086] "User interface means" refers to an interface (e.g., a web page, application screen, etc.) through which a user inputs a prompt and displays a response.

[1087] This invention relates to a system that automatically selects the optimal AI engine and its version for a prompt entered by a user, thereby reducing calculation costs and time. This system is implemented through interactions between a server, a terminal, and a user.

[1088] System configuration

[1089] 1. Hardware and Software Configuration

[1090] Server: Use a server or cloud service with powerful computing resources (e.g., AWS, Google Cloud).

[1091] Terminal: Refers to the device used by the user (e.g., PC, smartphone, tablet).

[1092] Natural Language Processing Software: Use natural language processing libraries such as NLTK or SpaCy for text analysis algorithms.

[1093] Generative artificial intelligence engine: Uses generative AI models such as GPT-3 and GPT-4.

[1094] Processing flow

[1095] 1. Enter and submit the prompt

[1096] The user types a question or request into an input field on the terminal, and this prompt is sent from the terminal to the server.

[1097] Example: A user types, "What's the weather like today?" or "Tell me more about regularization techniques for neural networks."

[1098] 2. Accepting and parsing prompts

[1099] The server receives the prompts sent from the device and uses text analysis algorithms to analyze the prompt content, including the type of question, the depth of information required, and the complexity of the vocabulary.

[1100] 3. Selection of the optimal engine

[1101] The server refers to the evaluation scores and feedback data collected from past dialogue history and selects the optimal generative AI engine and its version based on the analysis results.

[1102] 4. Sending prompts and generating responses

[1103] The server sends a prompt to the selected AI engine and receives the generated response. For example, the response "It's sunny today" is generated in response to the question "What's the weather like today?"

[1104] 5. Sending the response and displaying it to the user

[1105] The server sends the generated response to the terminal, which displays the response to the user.

[1106] 6. Gathering Feedback and Updating the Algorithm

[1107] The user inputs feedback on the displayed response into the terminal, which then sends the feedback to the server, which receives the feedback and uses it to update the selection algorithm.

[1108] Specific examples

[1109] For simple questions

[1110] 1. The user enters the prompt "What's the weather like today?"

[1111] 2. The terminal sends a prompt to the server.

[1112] 3. The server parses the prompt, determines that it is a simple question, and then selects the 3.5 version of the engine to keep costs down.

[1113] 4. The server sends a prompt to the engine and receives a response saying, "The weather is sunny today."

[1114] 5. The server sends the response to the terminal, which displays it to the user.

[1115] 6. The user gives feedback that they are "very satisfied," and the device sends the feedback to the server.

[1116] For complex technical questions

[1117] 1. The user types the prompt "Tell me more about regularization techniques in neural networks."

[1118] 2. The terminal sends a prompt to the server.

[1119] 3. The server parses the prompt, understands that you are requesting detailed technical information, and selects one of four high-precision engine versions.

[1120] 4. The server sends a prompt to the engine and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[1121] 5. The server sends the response to the terminal, which displays it to the user.

[1122] 6. The user gives feedback saying "somewhat satisfied," and the device sends the feedback to the server.

[1123] This system provides users with fast and accurate information without them having to worry about which AI engine is best suited to their query, and the system's accuracy can be improved by continuously utilizing user feedback.

[1124] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1125] Step 1: Enter the prompts

[1126] The user enters a question or request into an input field on the terminal.

[1127] As a concrete example, a user inputs text such as "What's the weather like today?" or "Please tell me more about regularization techniques for neural networks."

[1128] Input: Prompt sentence (e.g. "What's the weather like today?")

[1129] Output: prompt text formatted as text data

[1130] Step 2: Sending the prompt

[1131] The device sends the prompt entered by the user to the server by sending the prompt data to the server's API endpoint using an HTTP request.

[1132] Input: prompt text formatted as text data

[1133] Output: Prompt data sent to the server

[1134] Specific behavior: Issues an HTTP request over a network connection and sends prompt data.

[1135] Step 3: Accept the prompt

[1136] The server receives the prompt sent from the device and prepares it for analysis.

[1137] Input: Prompt data sent from the terminal

[1138] Output: Prompt data waiting to be parsed

[1139] Step 4: Parse the prompt

[1140] The server analyzes the prompt using text analysis algorithms (e.g., NLTK, SpaCy) to determine the type of question, the depth of information required, and lexical complexity.

[1141] Input: Prompt data waiting to be parsed

[1142] Output: Analysis results (e.g., question characteristics, difficulty, topic)

[1143] Specific operation: Tokenize the text, perform grammatical analysis, and semantic analysis, and extract each attribute.

[1144] Step 5: Selecting the optimal engine

[1145] Based on the analysis results, the server refers to the evaluation scores and feedback data collected from past dialogue history and selects the optimal generative AI engine and its version.

[1146] Input: Analysis results, evaluation scores, feedback data

[1147] Output: Selected generative AI engine and its version

[1148] What it does: Searches for suitable engines and versions in the database, and compares the rating scores and feedback to select the best engine.

[1149] Step 6: Sending prompts

[1150] The server sends a prompt to the selected AI engine and sends the prompt data using the AI ​​engine's API.

[1151] Input: Selected generation AI engine and its version, prompt data

[1152] Output: Prompt data sent to the generative AI engine

[1153] Specific behavior: Issues an API call and sends a prompt to the engine.

[1154] Step 7: Generate the Response

[1155] The generative artificial intelligence engine generates appropriate responses based on the prompts and sends them back to the server.

[1156] Input: prompt data

[1157] Output: The generated response

[1158] Specific behavior: Generates a natural language response using the internal model. Example: In response to the prompt "What's the weather like today?", generates "The weather is sunny today."

[1159] Step 8: Sending the Response

[1160] The server sends the generated response to the terminal.

[1161] Input: The generated response

[1162] Output: Response data sent to the device

[1163] Specific operation: The generated text is sent back to the terminal as an HTTP response.

[1164] Step 9: View the response

[1165] The terminal displays the received response to the user.

[1166] Input: Response data received from the server

[1167] Output: Response displayed on screen

[1168] Specific behavior: Display text on the user interface. Example: Display "Today's weather is sunny" on the screen.

[1169] Step 10: Provide feedback

[1170] The user inputs feedback on the displayed response into the terminal.

[1171] Input: Rating and comments on the response

[1172] Output: Input feedback data

[1173] Specific actions: Clicking the rating button or typing in the comment box and submitting.

[1174] Step 11: Submit your feedback

[1175] The terminal sends the feedback to the server.

[1176] Input: Feedback data

[1177] Output: Feedback data sent to the server

[1178] Specific operation: Sends feedback data to the server as an HTTP POST request.

[1179] Step 12: Update the algorithm

[1180] The server updates its selection algorithm based on the feedback it receives.

[1181] Input: Feedback data

[1182] Output: Updated election algorithm

[1183] What it does: The feedback data is stored in a database and used as training data for machine learning models, including periodic retraining.

[1184] (Application example 1)

[1185] 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."

[1186] Modern content delivery services require the ability to quickly provide optimal responses to a variety of prompts entered by users. Conventional systems often use the same generative AI engine for all requests, resulting in problems such as high-cost engines always being used or an inability to select an appropriate engine. Another issue is that user feedback is rarely reflected in the system, making it difficult to improve service quality. The present invention aims to solve these problems and provide a system for effectively providing customized content to users.

[1187] 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.

[1188] In this invention, the server includes means for analyzing an input prompt and selecting an optimal generating AI engine and its version, means for sending the prompt to the selected generating AI engine and receiving a generated response, means for sending the generated response to the user, means for receiving feedback from the user and updating the selection algorithm of the generating AI engine, and means for providing customized content based on the generated response. This makes it possible to select an optimal engine for the user's input prompt and provide customized content efficiently and effectively.

[1189] An "input prompt" is a request or question that a user sends to a system.

[1190] A "generative artificial intelligence engine" is an artificial intelligence algorithm that generates appropriate responses based on user prompts.

[1191] "Version" is a number or state that represents a different release or improvement of the Generative AI Engine.

[1192] "Server" is the central system that processes user requests, selects the appropriate engine, and generates and sends the response.

[1193] "User" means an end user who utilizes the system to enter prompts and receive responses and content.

[1194] A "response" is a response generated by a generative artificial intelligence engine based on a prompt.

[1195] "Feedback" refers to the opinions and ratings that users provide in response to a response.

[1196] A "selection algorithm" is a process or method for selecting the optimal generative AI engine and its version.

[1197] "Customized Content" means information and materials that are optimized for a particular user based on the user's prompts and feedback.

[1198] The present invention is a system that automatically selects the optimal generative AI engine and its version for a prompt input by a user, and provides customized content based on the selection result. This system is implemented through interactions between a server, a terminal, and a user.

[1199] The system program is structured as follows: First, the user uses the terminal to input a prompt. This prompt is in text format, such as "I would like to read an article about the latest AI technology." The terminal then sends this input prompt to the server.

[1200] The server then analyzes the received prompt using a text analysis algorithm that evaluates the prompt's content, vocabulary complexity, and required information depth. Based on the analysis results, the server selects the optimal generative AI engine and its version. The selection is made using an algorithm that references pre-set evaluation scores and past feedback data.

[1201] A prompt is sent to the selected generative AI engine, causing it to generate an appropriate response. For example, this response might be, "The following are recommended articles about the latest AI technology: 1. The cutting edge of AI research 2. The evolution of deep learning 3. The future of coexistence between AI and humans." The server receives this generated response and sends it to the device.

[1202] The terminal displays the received response to the user, who can then provide feedback on the displayed response, such as "very satisfied" or "somewhat satisfied." This feedback is then sent back to the server from the terminal.

[1203] The server updates the generative AI engine selection algorithm based on the received feedback. This algorithm update will enable more accurate engine selection in the future. For example, if users provide a large number of "very satisfied" feedback, the server recognizes that the engine version is highly rated and reflects this information in the next selection.

[1204] The hardware required to realize this system includes the user's device (smartphone, PC, etc.), server, and internet communication environment. The software includes a text analysis algorithm, an evaluation score and feedback data management algorithm, a generative AI engine, and a server program that integrates them.

[1205] In this way, the present invention allows users to select the most appropriate engine for their input prompts and receive customized content efficiently and effectively, which not only improves user satisfaction but also contributes to cost and resource optimization.

[1206] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1207] Step 1:

[1208] The user inputs a prompt using the terminal. For example, the user inputs text such as "I would like to read an article about the latest AI technology." The input prompt is sent from the terminal to the server.

[1209] Input: A prompt from the user (e.g., "I want to read an article about the latest AI technology")

[1210] Output: Prompt sent from terminal to server

[1211] Step 2:

[1212] The server receives the prompt text sent from the device, inputs it into a text analysis algorithm, and analyzes the content of the prompt. As a result of the analysis, the type of question, the depth of information required, the complexity of vocabulary, etc. are extracted.

[1213] Input: Prompt sent from the terminal

[1214] Output: Analysis of the prompt (type of question, depth of information required, vocabulary complexity, etc.)

[1215] Step 3:

[1216] The server selects the optimal generative AI engine and its version based on the analyzed prompts, taking into account pre-set evaluation scores and past feedback data.

[1217] Input: Parsed result of prompt statement

[1218] Output: Selected generative AI engine and its version

[1219] Step 4:

[1220] The server sends the prompt sentence to the selected AI engine to generate an appropriate response. The AI ​​engine generates a response based on the prompt sentence and sends it back to the server.

[1221] Input: Selected artificial intelligence engine, prompt

[1222] Output: The generated response

[1223] Step 5:

[1224] The server receives the generated response and sends it to the terminal, which displays the received response to the user.

[1225] Input: The generated response

[1226] Output: Sending the response to the terminal and displaying it to the user

[1227] Step 6:

[1228] The user provides feedback on the displayed response. For example, the user inputs an evaluation such as "very satisfied" into the terminal. The terminal then transmits the feedback to the server.

[1229] Input: User feedback (e.g. "Very satisfied")

[1230] Output: Sending feedback from the device to the server

[1231] Step 7:

[1232] The server updates the generative AI engine's selection algorithm based on the received feedback, allowing it to select a more appropriate engine and version for the next prompt.

[1233] Input: User feedback

[1234] Output: Updated election algorithm

[1235] In this way, the present invention realizes a system that selects the most suitable generative artificial intelligence engine in response to a user's prompts and provides customized content efficiently and effectively.

[1236] 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.

[1237] This invention relates to a system that automatically selects the optimal generative AI engine and its version for a prompt entered by a user, and further combines it with an emotion engine that recognizes the user's emotions, thereby reducing calculation costs and time and providing a more advanced response. This system is implemented through interactions between four parties: a server, a terminal, an emotion engine, and a user.

[1238] System Overview

[1239] 1. Input prompts and emotion recognition

[1240] Users can enter any question or request into the device's input field, and the emotion engine will simultaneously recognize emotions from the user's input, facial expressions, and voice.

[1241] 2. Sending prompts and emotion data

[1242] The device sends the entered prompt and recognized emotion data to the server, which is sent as JSON formatted data using an HTTP POST request.

[1243] 3. Accepting and parsing prompts

[1244] The server receives the prompt and emotion data sent from the device, analyzes the received data in an appropriate data format, and recognizes the prompt content and the user's emotion.

[1245] 4. Selection of the optimal engine

[1246] The server selects the optimal AI engine and its version based on the prompt analysis results and emotional data, referring to pre-set evaluation scores and past feedback data. Emotional data is also taken into consideration, so the engine that best suits the user's emotional state is selected.

[1247] 5. Sending prompts and generating responses

[1248] The server sends a prompt to the selected generative AI engine, which generates an appropriate response based on the received prompt and emotion data and sends it back to the server.

[1249] 6. Sending the response and displaying it to the user

[1250] The server then sends the generated response to the device, which then displays the received response to the user, including feedback based on the emotion data.

[1251] 7. Gathering feedback and updating the algorithm

[1252] Users can provide feedback on the displayed responses to their devices, which then send it to the server. The server receives the feedback and updates the selection algorithm. This feedback is collected as user satisfaction, additional comments, and emotional data.

[1253] Specific examples

[1254] For simple questions

[1255] 1. The user enters the prompt

[1256] User: As soon as the user types "What's the weather like today?", the device detects the emotion "happiness" from the user's facial expression.

[1257] 2. The device sends prompts and emotion data to the server

[1258] The terminal transmits the input prompt and the emotion data of "joy" to the server.

[1259] 3. The server analyzes the prompt and emotion data and selects the engine.

[1260] The server analyzes the prompt and emotion data, determines that it is a simple question, and selects Generative AI Engine 3.5.

[1261] 4. Send a prompt to the engine and receive a response

[1262] The server sends a prompt to engine 3.5 and receives the response "The weather is sunny today."

[1263] 5. Send the response to the device and display it to the user

[1264] The server sends the response to the terminal, which displays the response to the user.

[1265] 6. Users provide feedback

[1266] The user provides feedback such as "very satisfied," and the device sends this feedback to the server.

[1267] For complex technical questions

[1268] 1. The user enters the prompt

[1269] User: Type "Please tell me more about neural network regularization techniques" and in the process detect the emotion "interested" from the user's voice.

[1270] 2. The device sends prompts and emotion data to the server

[1271] The terminal transmits the input prompt and the emotion data of "interest" to the server.

[1272] 3. The server analyzes the prompt and emotion data and selects the engine.

[1273] The server analyzes the prompt and emotion data, determines that detailed technical content is required, and then selects a generative artificial intelligence engine 4.

[1274] 4. Send a prompt to the engine and receive a response

[1275] The server sends a prompt to Engine 4 and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[1276] 5. Send the response to the device and display it to the user

[1277] The server sends the response to the terminal, which displays the response to the user.

[1278] 6. Users provide feedback

[1279] The user provides feedback such as "somewhat satisfied," and the device sends this feedback to the server.

[1280] This system allows users to receive prompt and emotional information quickly and effectively without having to worry about the optimal AI engine, and the system's accuracy is continuously improved through feedback.

[1281] The processing flow will be explained below.

[1282] Step 1:

[1283] A user types a prompt into an input field on the device. For example, the user types "Tell me about the latest AI technologies."

[1284] Step 2:

[1285] The emotion engine analyzes the user's input, facial expressions, voice, etc. to recognize the user's emotions. For example, it can detect "interest" as an emotion.

[1286] Step 3:

[1287] The device sends the input prompt and emotion data to the server as JSON format data using an HTTP POST request.

[1288] Step 4:

[1289] The server receives the prompt and emotion data sent from the device and parses the received data into an appropriate data format (e.g., JSON).

[1290] Step 5:

[1291] The server analyzes the prompt content and emotional data using text analysis and emotion recognition algorithms, including the type of question, the depth of information requested, vocabulary complexity, and the user's emotions.

[1292] Step 6:

[1293] Based on the analysis results, the server refers to the preset evaluation score, past feedback data, and emotion data to select the optimal generative AI engine and its version. For example, if the question is technical and the emotion of "interest" is detected, generative AI engine 4 will be selected.

[1294] Step 7:

[1295] The server sends the prompt and emotion data to the selected generative AI engine using an API request.

[1296] Step 8:

[1297] A generative AI engine receives the prompt and emotion data and generates an appropriate response, which is then sent back to the server.

[1298] Step 9:

[1299] The server then formats the response as needed, specifically converting the JSON data into HTML.

[1300] Step 10:

[1301] The server sends the formatted response to the device, which then sends it to the device as an HTTP response.

[1302] Step 11:

[1303] The device receives the response from the server and displays it to the user.

[1304] Step 12:

[1305] Users provide feedback on the generated responses, including satisfaction, additional comments, and sentiment data.

[1306] Step 13:

[1307] The device sends the user feedback to the server, and the feedback data is sent in JSON format again.

[1308] Step 14:

[1309] The server receives user feedback and updates the selection algorithm, which improves the accuracy of engine selection from next time onwards.

[1310] Example 2

[1311] 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."

[1312] Conventional generative AI systems require manual selection of engines and versions to generate optimal responses to input prompts, resulting in increased computational costs and time for data processing and selection. Furthermore, it is difficult to automatically generate responses based on user emotions, and systems for improving user satisfaction are lacking. Furthermore, there is an insufficient mechanism for continuous improvement by aggregating and incorporating user feedback.

[1313] 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.

[1314] In this invention, the server includes means for receiving an input prompt from a user and recognizing the user's emotional data, means for analyzing the input prompt and the emotional data and selecting an optimal generating AI engine and its version, and means for sending the prompt to the selected generating AI engine and receiving a generated response, thereby enabling the selection of an optimal engine according to the user's emotional state and the rapid generation of a response.

[1315] "User" means an individual or entity that utilizes the System to enter prompts and receive information or services.

[1316] An "input prompt" is text data such as a question or request that a user inputs to a system.

[1317] "Emotion data" refers to data that indicates the emotional state of the user as recognized from facial expressions, voice, etc.

[1318] A "generative artificial intelligence engine" is an artificial intelligence model for generating appropriate responses based on user input prompts.

[1319] "Version" refers to different releases or revisions of the Generative Artificial Intelligence Engine.

[1320] "Analysis" refers to the process of verifying input prompts, emotional data, past feedback data, etc., and processing the data to select the optimal generative AI engine and its version.

[1321] A "response" is response data generated by a generative artificial intelligence engine in response to a user's input prompt.

[1322] "Feedback" refers to reaction data such as ratings and comments provided by users in response to generated responses.

[1323] A "selection algorithm" is a computational method or set of rules for analyzing prompts and emotional data and selecting the optimal generative AI engine and its version.

[1324] The "system" is a collection of components that executes a series of processes: receiving prompts and emotional data from users, selecting the optimal generative artificial intelligence engine to generate a response, aggregating feedback, and updating the algorithm.

[1325] This invention is a system that automatically selects the optimal generative AI engine and its version for the prompt entered by the user, and further combines it with an emotion engine that recognizes the user's emotions, thereby reducing calculation costs and time and providing a more advanced response. This system is implemented through interactions between elements including the user, terminal, server, and emotion engine.

[1326] Explanation of program processing

[1327] 1. User prompt input and emotion recognition

[1328] Users enter any question or request into an input field via a device such as a smartphone or computer. At the same time, the emotion engine uses the device's camera and microphone to analyze the user's facial expressions and voice and recognize their emotions. For example, when a user types "What's the weather like today?", the device detects the user's smile and identifies it as the emotion of "joy."

[1329] 2. Sending prompts and emotion data

[1330] The device compiles the input prompt and recognized emotion data in JSON format and sends it to the server using an HTTP POST request, specifically using the fetch API or XMLHttpRequest.

[1331] 3. Accepting and parsing prompts

[1332] The server receives the prompt and emotion data sent from the device and parses it into the appropriate format. It uses the JSON.parse method to parse the data and stores the prompt and emotion data in separate variables. During this process, it also references past feedback data and evaluation scores stored in a database (e.g., MySQL or PostgreSQL).

[1333] 4. Selection of the optimal engine

[1334] The server selects the optimal AI engine and its version based on the prompt analysis results and emotion data, along with evaluation scores and feedback data. For example, Engine 3.5 is selected for simple questions, while Engine 4 is selected for detailed technical questions.

[1335] 5. Sending prompts and generating responses

[1336] The server sends a prompt to the selected generative AI engine, which then generates an appropriate response. The generative AI engine takes into account the received prompt and emotional data to create an appropriate response and sends it back to the server. For example, a response might be generated: "The weather is sunny today."

[1337] 6. Sending the response and displaying it to the user

[1338] The server sends the generated response to the device, which then displays it to the user. The device screen displays "Today's weather is sunny."

[1339] 7. Gathering feedback and updating the algorithm

[1340] The user inputs feedback on the displayed response into the terminal, including options such as "very satisfied" and "somewhat satisfied." The terminal then transmits the user's feedback to the server.

[1341] The server stores the received feedback in a database and updates the selection algorithm based on that feedback, which will result in better engine selection in the future.

[1342] Specific examples

[1343] For simple questions

[1344] 1. User: Enter "What's the weather like today?" The device detects "happiness" from the user's facial expression.

[1345] 2. Device: Sends the prompt and the "happiness" emotion data to the server.

[1346] 3. Server: Analyzes prompt and sentiment data and selects engine 3.5.

[1347] 4. Server: Sends a prompt to Engine 3.5 and receives the response "The weather is sunny today."

[1348] 5. Server: Sends the response to the terminal, which displays it to the user.

[1349] 6. User: Gives feedback that he is "very satisfied," and the device sends the feedback to the server.

[1350] For complex technical questions

[1351] 1. User: Type "Please tell me more about neural network regularization techniques." The device detects "interests" from the user's voice.

[1352] 2. Device: Sends prompt and "interest" emotion data to the server.

[1353] 3. Server: Analyzes prompt and emotion data and selects Engine 4.

[1354] 4. Server: Sends a prompt to Engine 4 and receives a response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[1355] 5. Server: Sends the response to the terminal, which displays it to the user.

[1356] 6. User: Gives feedback that they are "somewhat satisfied," and the device sends that feedback to the server.

[1357] The system selects the most suitable AI engine and provides fast and effective information by simply inputting prompts and emotional data, while the accuracy of the system is continuously improved through feedback.

[1358] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1359] Divide the processing flow of the system program into processing steps

[1360] Step 1: User prompt input and emotion recognition

[1361] 1.1 Prompt Input

[1362] 1.2 Emotion recognition

[1363] Step 2: Sending prompts and emotion data

[1364] Step 3: Server accepts and parses prompt

[1365] 3.1 Receiving prompts and emotion data

[1366] 3.2 Prompt and Emotion Data Analysis

[1367] Step 4: Selecting the optimal engine

[1368] 4.1 Referencing evaluation scores and feedback data

[1369] 4.2 Selecting the optimal engine and version

[1370] Step 5: Sending prompts and generating responses

[1371] 5.1 Sending Prompts to the Selection Engine

[1372] 5.2 Engine Generates and Returns Response

[1373] Step 6: Sending the Response and Displaying it to the User

[1374] Step 7: Gather feedback and update the algorithm

[1375] 7.1 Entering and Submitting Feedback

[1376] 7.2 Receiving Feedback and Updating the Algorithm

[1377] Specific explanation of each processing step

[1378] Step 1: User prompt input and emotion recognition

[1379] 1.1 Prompt Input

[1380] The user writes a question or request into the input field of the terminal. For example, "What's the weather like today?" Input: The text data entered by the user into the terminal. Output: The prompt text entered.

[1381] 1.2 Emotion recognition

[1382] The device uses the built-in camera and microphone to collect the user's facial expressions and voice, and the emotion engine analyzes their emotions in real time. For example, if the user smiles, the emotion "joy" is detected. Input: User's facial and voice data. Output: Analyzed emotion data (e.g., "joy").

[1383] Step 2: Sending prompts and emotion data

[1384] The device collects the input prompt and emotion data and sends them to the server in JSON format. Specifically, it makes an HTTP POST request using the fetch API or XMLHttpRequest. Input: Prompt text and emotion data. Output: JSON data sent to the server.

[1385] Step 3: Server accepts and parses prompt

[1386] 3.1 Receiving prompts and emotion data

[1387] The server receives the prompt and emotion data sent from the device. Input: JSON data sent from the device. Output: Received prompt and emotion data.

[1388] 3.2 Prompt and Emotion Data Analysis

[1389] The server parses the prompt and emotion data using the JSON.parse method and stores them in separate variables. Input: Received prompt and emotion data. Output: Parsed prompt and emotion data.

[1390] Step 4: Selecting the optimal engine

[1391] 4.1 Referencing evaluation scores and feedback data

[1392] The server references past feedback data and evaluation scores from a database. For example, queries are performed using MySQL or PostgreSQL. Input: prompt text and emotion data. Output: evaluation scores and feedback data required to select the optimal engine.

[1393] 4.2 Selecting the optimal engine and version

[1394] The server selects the optimal generative AI engine and its version based on the prompt, emotion data, evaluation scores, and feedback data. Input: Evaluation scores and feedback data. Output: Selected optimal engine and its version.

[1395] Step 5: Sending prompts and generating responses

[1396] 5.1 Sending Prompts to the Selection Engine

[1397] The server sends a prompt to the selected engine. Specifically, it sends an HTTP request to the engine's API endpoint. Input: Selected engine and prompt text. Output: Sending prompt data to the generative AI engine.

[1398] 5.2 Engine Generates and Returns Response

[1399] The generative AI model analyzes the prompt and emotion data, generates an appropriate response, and sends it back to the server. For example, the response "Today's weather is sunny." Input: Prompt sentence and emotion data. Output: Generated response sentence.

[1400] Step 6: Sending the Response and Displaying it to the User

[1401] The server sends the generated response to the device. The device receives the response and displays it to the user. For example, "Today's weather is sunny" is displayed on a smartphone screen. Input: The generated response text. Output: The response text displayed to the user.

[1402] Step 7: Gather feedback and update the algorithm

[1403] 7.1 Entering and Submitting Feedback

[1404] The user inputs feedback about the displayed response into the terminal. For example, the user inputs feedback such as "very satisfied." The terminal sends the feedback to the server. Input: User feedback. Output: Feedback data sent to the server.

[1405] 7.2 Receiving Feedback and Updating the Algorithm

[1406] The server stores the received feedback in a database and updates the selection algorithm, which will result in better prompt selection in future. Input: Submitted feedback data. Output: Updated selection algorithm.

[1407] (Application example 2)

[1408] 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."

[1409] Conventional generative AI systems did not take user emotions into consideration, limiting the quality of the user experience. Furthermore, selecting the optimal AI engine was difficult and required significant time and computational costs. This often led to dissatisfaction with the quality and timeliness of information provided.

[1410] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1411] In this invention, the server includes means for analyzing the input prompt and selecting the optimal generative AI engine and its version, means for recognizing the user's emotions and acquiring emotion data, and means for transmitting the acquired emotion data to the server and selecting the optimal generative AI engine together with the prompt analysis result, thereby making it possible to provide a more appropriate response that takes the user's emotions into consideration and improve the quality of the user experience.

[1412] An "input prompt" is a question or request that a user provides to a system.

[1413] A "generative artificial intelligence engine" refers to an algorithm or system for generating responses in natural language based on input prompts.

[1414] "Emotion data" refers to information about emotions extracted from a user's facial expressions, voice, text input, etc.

[1415] "Server" refers to a central processing unit for analyzing input data and providing generated responses to users.

[1416] "Prompt analysis result" refers to information obtained by the server analyzing the input prompt.

[1417] "Optimal engine selection" refers to the process of selecting the appropriate generative AI engine and its version based on prompt analysis results and emotion data.

[1418] "Feedback" refers to the opinions and ratings that users provide in response to responses they receive.

[1419] "Selection algorithm" refers to a set of calculation procedures and rules for selecting the optimal generative artificial intelligence engine and its version.

[1420] System Configuration

[1421] The system based on the present invention allows users to communicate with emotion recognition using generative AI models. This system includes the following elements:

[1422] 1. Parsing the input prompt

[1423] The server analyzes the prompts entered by the user through the terminal. The prompts contain questions or requests from the user, and performs syntactic and semantic analysis using natural language analysis techniques.

[1424] 2. Acquiring Emotion Data

[1425] The device acquires the user's emotion data related to the input prompt. The emotion data is recognized from the user's facial expressions, voice, text input, etc. using an emotion engine (e.g., a model using OpenCV or TensorFlow).

[1426] 3. Selection of the optimal engine

[1427] The server selects the optimal generative AI engine and its version based on the prompt analysis results and emotion data. The server references past evaluation scores and feedback data and determines the optimal engine using an engine selection algorithm.

[1428] 4. Retrieving the generated response

[1429] The server sends prompts to the selected generative artificial intelligence engine and receives a generated response, which is provided in natural language format and includes an appropriate response to the user's prompt.

[1430] 5. Displaying the response to the user

[1431] The device displays the received responses to the user, and the responses are generated taking into account the user's emotional state, resulting in a satisfying interaction experience for the user.

[1432] 6. Gathering feedback and updating the algorithm

[1433] The device sends user feedback to the server, which then updates the engine selection algorithm based on the feedback to achieve better engine selection in the future.

[1434] Hardware and software used

[1435] Hardware

[1436] Devices (smartphones, smart glasses, head-mounted displays, etc.)

[1437] Server (high-performance data analysis server)

[1438] software

[1439] Emotion engine (OpenCV, TensorFlow, Keras)

[1440] HTTP request processing library (requests)

[1441] Generative AI engine (general-purpose generative AI model)

[1442] Specific examples

[1443] If a user types "What's the latest trending item?" and makes a surprised expression, the emotion engine will detect the "surprise" emotion. This prompt and emotion data will be sent to the server. The server will analyze the prompt and emotion data, select the optimal generative AI engine, and generate a response from the engine saying, "The latest trending item is a new smartwatch," which will be displayed to the user via their device.

[1444] Prompt Sentence Examples

[1445] What's the latest trending item?

[1446] Thus, the present invention is a system that enhances the user experience by recognizing the user's emotions and providing optimal responses based on them.

[1447] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1448] Step 1:

[1449] The user inputs a prompt into a device (such as a smartphone, smart glasses, or head-mounted display). The device then captures the user's input as text data. It also simultaneously records the user's facial expressions and voice, capturing them as emotion data. This allows the device to collect the prompt and emotion data as input.

[1450] Step 2:

[1451] The device sends the collected prompt sentences and emotion data to the server. This data is converted into JSON format and sent using an HTTP POST request. Specifically, the device converts the collected data into JSON and sends it to the server using the HTTP request library (requests).

[1452] Step 3:

[1453] The server analyzes the received prompt and emotion data. It parses the JSON data received as input and extracts the prompt and emotion data. It then performs syntactic and semantic analysis of the prompt using a natural language processing library (e.g., spaCy, NLTK). Meanwhile, it uses an emotion engine (e.g., TensorFlow, Keras) to classify the emotion data.

[1454] Step 4:

[1455] The server selects the optimal generative AI engine and its version based on the analysis results and emotional data. It uses past evaluation scores and feedback data to select the optimal engine using an algorithm. The algorithm uses the prompt type, the user's emotional state, and past engine performance data as inputs, and outputs the selection result.

[1456] Step 5:

[1457] The server sends a prompt to the selected generative AI engine and receives the generated response. Specifically, the server calls the selected engine's API and sends the prompt. The engine analyzes the prompt, generates an appropriate natural language response, and sends it back to the server. This response becomes the server's output.

[1458] Step 6:

[1459] The server sends the generated response to the terminal. The response data is converted to JSON format and sent to the terminal using an HTTP POST request. Specifically, the server converts the generated response to JSON and sends it to the terminal using the HTTP request library (requests).

[1460] Step 7:

[1461] The device displays the received response to the user. The user can then enter feedback such as satisfaction or additional comments. This feedback is also captured as text data and used in a later step. The device then converts the captured feedback into JSON format.

[1462] Step 8:

[1463] The device sends user feedback to the server. This feedback data is sent using an HTTP POST request. Specifically, the device converts the feedback data into JSON and sends it to the server using the HTTP request library (requests).

[1464] Step 9:

[1465] The server analyzes the received feedback and updates the engine selection algorithm. The analysis includes semantic analysis of the feedback comments and weighted averaging of the evaluation scores. Based on the analysis results, the engine selection algorithm is tuned to achieve more accurate selection in the future.

[1466] 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.

[1467] 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.

[1468] 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.

[1469] [Fourth embodiment]

[1470] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1471] 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.

[1472] 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).

[1473] 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.

[1474] 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.

[1475] 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).

[1476] 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. 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.

[1477] 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.

[1478] 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.

[1479] 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.

[1480] 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.

[1481] 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.

[1482] 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."

[1483] The present invention relates to a system that automatically selects the optimal generation AI engine and its version for a prompt entered by a user, thereby reducing calculation costs and time. This system is implemented through interactions between a server, a terminal, and a user.

[1484] System Overview

[1485] 1. Enter and submit the prompt

[1486] The user types any question or request into an input field on the device, and this prompt is sent from the device to the server.

[1487] 2. Accepting and parsing prompts

[1488] The server receives the prompts sent from the device and uses text analysis algorithms to analyze the prompt content, including the type of question, the depth of information required, and the complexity of the vocabulary.

[1489] 3. Selection of the optimal engine

[1490] The server selects the optimal generative AI engine and its version based on the analysis results and references pre-set evaluation scores and feedback data collected from past dialogue history.

[1491] 4. Sending prompts and generating responses

[1492] The server sends prompts to the selected generative AI engine, which generates an appropriate response based on the received prompt and sends it back to the server.

[1493] 5. Sending the response and displaying it to the user

[1494] The server sends the generated response to the terminal, and the terminal displays the received response to the user.

[1495] 6. Gathering feedback and updating the algorithm

[1496] Users can input their feedback on the displayed responses into the device, which then sends it to the server, which receives the feedback and updates the selection algorithm. This feedback is collected as user satisfaction and additional comments.

[1497] Specific examples

[1498] For simple questions

[1499] 1. The user enters the prompt

[1500] User: Type "What's the weather like today?"

[1501] 2. The device sends a prompt to the server

[1502] The terminal sends the entered prompt to the server.

[1503] 3. The server parses the prompt and selects an engine

[1504] The server analyzes the prompt and determines that it is a simple question. To keep costs down, it selects Generative AI Engine 3.5.

[1505] 4. Send a prompt to the engine and receive a response

[1506] The server sends a prompt to engine 3.5 and receives the response "The weather is sunny today."

[1507] 5. Send the response to the device and display it to the user

[1508] The server sends the response to the terminal, which displays the response to the user.

[1509] 6. Users provide feedback

[1510] User: Provides feedback saying "Very satisfied." The device sends the feedback to the server.

[1511] For complex technical questions

[1512] 1. The user enters the prompt

[1513] User: Type "Tell me more about regularization techniques for neural networks."

[1514] 2. The device sends a prompt to the server

[1515] The terminal sends the entered prompt to the server.

[1516] 3. The server parses the prompt and selects an engine

[1517] The server analyzes the prompt, understands that it is requesting detailed technical content, and selects a highly accurate generative AI engine4.

[1518] 4. Send a prompt to the engine and receive a response

[1519] The server sends a prompt to Engine 4 and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[1520] 5. Send the response to the device and display it to the user

[1521] The server sends the response to the terminal, which displays the response to the user.

[1522] 6. Users provide feedback

[1523] User: Provides feedback such as "somewhat satisfied." The device sends the feedback to the server.

[1524] This system allows users to receive fast and effective information without having to worry about the optimal AI engine, simply by entering prompts, and the system's accuracy is continuously improved through feedback.

[1525] The processing flow will be explained below.

[1526] Step 1:

[1527] A user types a prompt into an input field on the device. For example, the user types "Tell me about the latest AI technologies."

[1528] Step 2:

[1529] The terminal sends the entered prompt to the server, specifically as JSON-formatted data using an HTTP POST request.

[1530] Step 3:

[1531] The server receives the prompt sent from the device and parses it into the appropriate data format (e.g., JSON).

[1532] Step 4:

[1533] The server analyzes the prompt content using text analysis algorithms, including the type of question, the depth of information requested, and the complexity of the vocabulary.

[1534] Step 5:

[1535] Based on the analysis results, the server refers to a pre-set evaluation score and past feedback data to select the optimal generative AI engine and its version. For example, generative AI4 is selected for technical questions.

[1536] Step 6:

[1537] The server sends a prompt to the selected generative AI engine using an API request.

[1538] Step 7:

[1539] A generative AI engine receives the prompt and generates an appropriate response, which is sent back to the server.

[1540] Step 8:

[1541] The server then formats the response as needed, specifically converting the JSON data into HTML.

[1542] Step 9:

[1543] The server sends the formatted response to the device, which then sends it to the device as an HTTP response.

[1544] Step 10:

[1545] The device receives the response from the server and displays it to the user.

[1546] Step 11:

[1547] The user provides feedback on the generated response, including satisfaction and any additional comments.

[1548] Step 12:

[1549] The device sends the user feedback to the server, and the feedback data is sent in JSON format again.

[1550] Step 13:

[1551] The server receives user feedback and updates the selection algorithm, which improves the accuracy of engine selection from next time onwards.

[1552] Example 1

[1553] 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."

[1554] Currently, there are many generative AI engines available, each with different characteristics and performance. However, manually selecting which engine is best for a particular prompt is inefficient, and if an appropriate selection is not made, the quality of the generated response may decline. Furthermore, a method is needed to effectively utilize user feedback to update the engine selection algorithm and improve the system's accuracy. This will enable the provision of optimal responses while reducing computational costs and time.

[1555] 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.

[1556] In this invention, the server includes means for analyzing an input prompt and selecting an optimal generative AI engine and its version, means for sending the prompt to the selected generative AI engine and receiving a generated response, means for sending the generated response to a terminal and displaying it to the user, means for receiving feedback from the user and updating the selection algorithm of the generative AI engine, and user interface means for displaying the generated response to the user. This allows the user to quickly and effectively receive information simply by entering the prompt, without having to worry about selecting the optimal AI engine. Furthermore, the system's accuracy can be continuously improved by utilizing feedback.

[1557] An "input prompt" is text data that a user inputs into a terminal in the form of a question, request, or the like.

[1558] "Parsing" is the process of classifying and interpreting the content of the received input prompt using natural language processing algorithms.

[1559] A "generative artificial intelligence engine" is an artificial intelligence system capable of generating natural language responses based on a given input prompt.

[1560] A "version" is identification information that indicates different revisions or update states of a particular generative AI engine.

[1561] A "response" is a reply message in natural language generated by the selected artificial intelligence engine.

[1562] A "user" is a person who uses the system to enter questions or requests and receive responses generated by the AI ​​engine.

[1563] A "terminal" is a device (e.g., a PC, smartphone, etc.) on which a user enters prompts and receives and displays responses from the server.

[1564] A "text analysis algorithm" is an algorithm that uses natural language processing technology to analyze the meaning and content of input prompts.

[1565] The "evaluation score" is a numerical value used to evaluate the performance and appropriateness of the generative AI engine based on past dialogue history.

[1566] "Feedback data" refers to data that includes evaluations and comments on responses provided by users.

[1567] The "selection algorithm" is a method for selecting the optimal generative AI engine and version based on analysis results and feedback data.

[1568] "User interface means" refers to an interface (e.g., a web page, application screen, etc.) through which a user inputs a prompt and displays a response.

[1569] This invention relates to a system that automatically selects the optimal AI engine and its version for a prompt entered by a user, thereby reducing calculation costs and time. This system is implemented through interactions between a server, a terminal, and a user.

[1570] System configuration

[1571] 1. Hardware and Software Configuration

[1572] Server: Use a server or cloud service with powerful computing resources (e.g., AWS, Google Cloud).

[1573] Terminal: Refers to the device used by the user (e.g., PC, smartphone, tablet).

[1574] Natural Language Processing Software: Use natural language processing libraries such as NLTK or SpaCy for text analysis algorithms.

[1575] Generative artificial intelligence engine: Uses generative AI models such as GPT-3 and GPT-4.

[1576] Processing flow

[1577] 1. Enter and submit the prompt

[1578] The user types a question or request into an input field on the terminal, and this prompt is sent from the terminal to the server.

[1579] Example: A user types, "What's the weather like today?" or "Tell me more about regularization techniques for neural networks."

[1580] 2. Accepting and parsing prompts

[1581] The server receives the prompts sent from the device and uses text analysis algorithms to analyze the prompt content, including the type of question, the depth of information required, and the complexity of the vocabulary.

[1582] 3. Selection of the optimal engine

[1583] The server refers to the evaluation scores and feedback data collected from past dialogue history and selects the optimal generative AI engine and its version based on the analysis results.

[1584] 4. Sending prompts and generating responses

[1585] The server sends a prompt to the selected AI engine and receives the generated response. For example, the response "It's sunny today" is generated in response to the question "What's the weather like today?"

[1586] 5. Sending the response and displaying it to the user

[1587] The server sends the generated response to the terminal, which displays the response to the user.

[1588] 6. Gathering Feedback and Updating the Algorithm

[1589] The user inputs feedback on the displayed response into the terminal, which then sends the feedback to the server, which receives the feedback and uses it to update the selection algorithm.

[1590] Specific examples

[1591] For simple questions

[1592] 1. The user enters the prompt "What's the weather like today?"

[1593] 2. The terminal sends a prompt to the server.

[1594] 3. The server parses the prompt, determines that it is a simple question, and then selects the 3.5 version of the engine to keep costs down.

[1595] 4. The server sends a prompt to the engine and receives a response saying, "The weather is sunny today."

[1596] 5. The server sends the response to the terminal, which displays it to the user.

[1597] 6. The user gives feedback that they are "very satisfied," and the device sends the feedback to the server.

[1598] For complex technical questions

[1599] 1. The user types the prompt "Tell me more about regularization techniques in neural networks."

[1600] 2. The terminal sends a prompt to the server.

[1601] 3. The server parses the prompt, understands that you are requesting detailed technical information, and selects one of four high-precision engine versions.

[1602] 4. The server sends a prompt to the engine and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[1603] 5. The server sends the response to the terminal, which displays it to the user.

[1604] 6. The user gives feedback saying "somewhat satisfied," and the device sends the feedback to the server.

[1605] This system provides users with fast and accurate information without them having to worry about which AI engine is best suited to their query, and the system's accuracy can be improved by continuously utilizing user feedback.

[1606] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1607] Step 1: Enter the prompts

[1608] The user enters a question or request into an input field on the terminal.

[1609] As a concrete example, a user inputs text such as "What's the weather like today?" or "Please tell me more about regularization techniques for neural networks."

[1610] Input: Prompt sentence (e.g. "What's the weather like today?")

[1611] Output: prompt text formatted as text data

[1612] Step 2: Sending the prompt

[1613] The device sends the prompt entered by the user to the server by sending the prompt data to the server's API endpoint using an HTTP request.

[1614] Input: prompt text formatted as text data

[1615] Output: Prompt data sent to the server

[1616] Specific behavior: Issues an HTTP request over a network connection and sends prompt data.

[1617] Step 3: Accept the prompt

[1618] The server receives the prompt sent from the device and prepares it for analysis.

[1619] Input: Prompt data sent from the terminal

[1620] Output: Prompt data waiting to be parsed

[1621] Step 4: Parse the prompt

[1622] The server analyzes the prompt using text analysis algorithms (e.g., NLTK, SpaCy) to determine the type of question, the depth of information required, and lexical complexity.

[1623] Input: Prompt data waiting to be parsed

[1624] Output: Analysis results (e.g., question characteristics, difficulty, topic)

[1625] Specific operation: Tokenize the text, perform grammatical analysis, and semantic analysis, and extract each attribute.

[1626] Step 5: Selecting the optimal engine

[1627] Based on the analysis results, the server refers to the evaluation scores and feedback data collected from past dialogue history and selects the optimal generative AI engine and its version.

[1628] Input: Analysis results, evaluation scores, feedback data

[1629] Output: Selected generative AI engine and its version

[1630] What it does: Searches for suitable engines and versions in the database, and compares the rating scores and feedback to select the best engine.

[1631] Step 6: Sending prompts

[1632] The server sends a prompt to the selected AI engine and sends the prompt data using the AI ​​engine's API.

[1633] Input: Selected generation AI engine and its version, prompt data

[1634] Output: Prompt data sent to the generative AI engine

[1635] Specific behavior: Issues an API call and sends a prompt to the engine.

[1636] Step 7: Generate the Response

[1637] The generative artificial intelligence engine generates appropriate responses based on the prompts and sends them back to the server.

[1638] Input: prompt data

[1639] Output: The generated response

[1640] Specific behavior: Generates a natural language response using the internal model. Example: In response to the prompt "What's the weather like today?", generates "The weather is sunny today."

[1641] Step 8: Sending the Response

[1642] The server sends the generated response to the terminal.

[1643] Input: The generated response

[1644] Output: Response data sent to the device

[1645] Specific operation: The generated text is sent back to the terminal as an HTTP response.

[1646] Step 9: View the response

[1647] The terminal displays the received response to the user.

[1648] Input: Response data received from the server

[1649] Output: Response displayed on screen

[1650] Specific behavior: Display text on the user interface. Example: Display "Today's weather is sunny" on the screen.

[1651] Step 10: Provide feedback

[1652] The user inputs feedback on the displayed response into the terminal.

[1653] Input: Rating and comments on the response

[1654] Output: Input feedback data

[1655] Specific actions: Clicking the rating button or typing in the comment box and submitting.

[1656] Step 11: Submit your feedback

[1657] The terminal sends the feedback to the server.

[1658] Input: Feedback data

[1659] Output: Feedback data sent to the server

[1660] Specific operation: Sends feedback data to the server as an HTTP POST request.

[1661] Step 12: Update the algorithm

[1662] The server updates its selection algorithm based on the feedback it receives.

[1663] Input: Feedback data

[1664] Output: Updated election algorithm

[1665] What it does: The feedback data is stored in a database and used as training data for machine learning models, including periodic retraining.

[1666] (Application example 1)

[1667] 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."

[1668] Modern content delivery services require the ability to quickly provide optimal responses to a variety of prompts entered by users. Conventional systems often use the same generative AI engine for all requests, resulting in problems such as high-cost engines always being used or an inability to select an appropriate engine. Another issue is that user feedback is rarely reflected in the system, making it difficult to improve service quality. The present invention aims to solve these problems and provide a system for effectively providing customized content to users.

[1669] 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.

[1670] In this invention, the server includes means for analyzing an input prompt and selecting an optimal generating AI engine and its version, means for sending the prompt to the selected generating AI engine and receiving a generated response, means for sending the generated response to the user, means for receiving feedback from the user and updating the selection algorithm of the generating AI engine, and means for providing customized content based on the generated response. This makes it possible to select an optimal engine for the user's input prompt and provide customized content efficiently and effectively.

[1671] An "input prompt" is a request or question that a user sends to a system.

[1672] A "generative artificial intelligence engine" is an artificial intelligence algorithm that generates appropriate responses based on user prompts.

[1673] "Version" is a number or state that represents a different release or improvement of the Generative AI Engine.

[1674] "Server" is the central system that processes user requests, selects the appropriate engine, and generates and sends the response.

[1675] "User" means an end user who utilizes the system to enter prompts and receive responses and content.

[1676] A "response" is a response generated by a generative artificial intelligence engine based on a prompt.

[1677] "Feedback" refers to the opinions and ratings that users provide in response to a response.

[1678] A "selection algorithm" is a process or method for selecting the optimal generative AI engine and its version.

[1679] "Customized Content" means information and materials that are optimized for a particular user based on the user's prompts and feedback.

[1680] The present invention is a system that automatically selects the optimal generative AI engine and its version for a prompt input by a user, and provides customized content based on the selection result. This system is implemented through interactions between a server, a terminal, and a user.

[1681] The system program is structured as follows: First, the user uses the terminal to input a prompt. This prompt is in text format, such as "I would like to read an article about the latest AI technology." The terminal then sends this input prompt to the server.

[1682] The server then analyzes the received prompt using a text analysis algorithm that evaluates the prompt's content, vocabulary complexity, and required information depth. Based on the analysis results, the server selects the optimal generative AI engine and its version. The selection is made using an algorithm that references pre-set evaluation scores and past feedback data.

[1683] A prompt is sent to the selected generative AI engine, causing it to generate an appropriate response. For example, this response might be, "The following are recommended articles about the latest AI technology: 1. The cutting edge of AI research 2. The evolution of deep learning 3. The future of coexistence between AI and humans." The server receives this generated response and sends it to the device.

[1684] The terminal displays the received response to the user, who can then provide feedback on the displayed response, such as "very satisfied" or "somewhat satisfied." This feedback is then sent back to the server from the terminal.

[1685] The server updates the generative AI engine selection algorithm based on the received feedback. This algorithm update will enable more accurate engine selection in the future. For example, if users provide a large number of "very satisfied" feedback, the server recognizes that the engine version is highly rated and reflects this information in the next selection.

[1686] The hardware required to realize this system includes the user's device (smartphone, PC, etc.), server, and internet communication environment. The software includes a text analysis algorithm, an evaluation score and feedback data management algorithm, a generative AI engine, and a server program that integrates them.

[1687] In this way, the present invention allows users to select the most appropriate engine for their input prompts and receive customized content efficiently and effectively, which not only improves user satisfaction but also contributes to cost and resource optimization.

[1688] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1689] Step 1:

[1690] The user inputs a prompt using the terminal. For example, the user inputs text such as "I would like to read an article about the latest AI technology." The input prompt is sent from the terminal to the server.

[1691] Input: A prompt from the user (e.g., "I want to read an article about the latest AI technology")

[1692] Output: Prompt sent from terminal to server

[1693] Step 2:

[1694] The server receives the prompt text sent from the device, inputs it into a text analysis algorithm, and analyzes the content of the prompt. As a result of the analysis, the type of question, the depth of information required, the complexity of vocabulary, etc. are extracted.

[1695] Input: Prompt sent from the terminal

[1696] Output: Analysis of the prompt (type of question, depth of information required, vocabulary complexity, etc.)

[1697] Step 3:

[1698] The server selects the optimal generative AI engine and its version based on the analyzed prompts, taking into account pre-set evaluation scores and past feedback data.

[1699] Input: Parsed result of prompt statement

[1700] Output: Selected generative AI engine and its version

[1701] Step 4:

[1702] The server sends the prompt sentence to the selected AI engine to generate an appropriate response. The AI ​​engine generates a response based on the prompt sentence and sends it back to the server.

[1703] Input: Selected artificial intelligence engine, prompt

[1704] Output: The generated response

[1705] Step 5:

[1706] The server receives the generated response and sends it to the terminal, which displays the received response to the user.

[1707] Input: The generated response

[1708] Output: Sending the response to the terminal and displaying it to the user

[1709] Step 6:

[1710] The user provides feedback on the displayed response. For example, the user inputs an evaluation such as "very satisfied" into the terminal. The terminal then transmits the feedback to the server.

[1711] Input: User feedback (e.g. "Very satisfied")

[1712] Output: Sending feedback from the device to the server

[1713] Step 7:

[1714] The server updates the generative AI engine's selection algorithm based on the received feedback, allowing it to select a more appropriate engine and version for the next prompt.

[1715] Input: User feedback

[1716] Output: Updated election algorithm

[1717] In this way, the present invention realizes a system that selects the most suitable generative artificial intelligence engine in response to a user's prompts and provides customized content efficiently and effectively.

[1718] 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.

[1719] This invention relates to a system that automatically selects the optimal generative AI engine and its version for a prompt entered by a user, and further combines it with an emotion engine that recognizes the user's emotions, thereby reducing calculation costs and time and providing a more advanced response. This system is implemented through interactions between four parties: a server, a terminal, an emotion engine, and a user.

[1720] System Overview

[1721] 1. Input prompts and emotion recognition

[1722] Users can enter any question or request into the device's input field, and the emotion engine will simultaneously recognize emotions from the user's input, facial expressions, and voice.

[1723] 2. Sending prompts and emotion data

[1724] The device sends the entered prompt and recognized emotion data to the server, which is sent as JSON formatted data using an HTTP POST request.

[1725] 3. Accepting and parsing prompts

[1726] The server receives the prompt and emotion data sent from the device, analyzes the received data in an appropriate data format, and recognizes the prompt content and the user's emotion.

[1727] 4. Selection of the optimal engine

[1728] The server selects the optimal AI engine and its version based on the prompt analysis results and emotional data, referring to pre-set evaluation scores and past feedback data. Emotional data is also taken into consideration, so the engine that best suits the user's emotional state is selected.

[1729] 5. Sending prompts and generating responses

[1730] The server sends a prompt to the selected generative AI engine, which generates an appropriate response based on the received prompt and emotion data and sends it back to the server.

[1731] 6. Sending the response and displaying it to the user

[1732] The server then sends the generated response to the device, which then displays the received response to the user, including feedback based on the emotion data.

[1733] 7. Gathering feedback and updating the algorithm

[1734] Users can provide feedback on the displayed responses to their devices, which then send it to the server. The server receives the feedback and updates the selection algorithm. This feedback is collected as user satisfaction, additional comments, and emotional data.

[1735] Specific examples

[1736] For simple questions

[1737] 1. The user enters the prompt

[1738] User: As soon as the user types "What's the weather like today?", the device detects the emotion "happiness" from the user's facial expression.

[1739] 2. The device sends prompts and emotion data to the server

[1740] The terminal transmits the input prompt and the emotion data of "joy" to the server.

[1741] 3. The server analyzes the prompt and emotion data and selects the engine.

[1742] The server analyzes the prompt and emotion data, determines that it is a simple question, and selects Generative AI Engine 3.5.

[1743] 4. Send a prompt to the engine and receive a response

[1744] The server sends a prompt to engine 3.5 and receives the response "The weather is sunny today."

[1745] 5. Send the response to the device and display it to the user

[1746] The server sends the response to the terminal, which displays the response to the user.

[1747] 6. Users provide feedback

[1748] The user provides feedback such as "very satisfied," and the device sends this feedback to the server.

[1749] For complex technical questions

[1750] 1. The user enters the prompt

[1751] User: Type "Please tell me more about neural network regularization techniques" and in the process detect the emotion "interested" from the user's voice.

[1752] 2. The device sends prompts and emotion data to the server

[1753] The terminal transmits the input prompt and the emotion data of "interest" to the server.

[1754] 3. The server analyzes the prompt and emotion data and selects the engine.

[1755] The server analyzes the prompt and emotion data, determines that detailed technical content is required, and then selects a generative artificial intelligence engine 4.

[1756] 4. Send a prompt to the engine and receive a response

[1757] The server sends a prompt to Engine 4 and receives a detailed response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[1758] 5. Send the response to the device and display it to the user

[1759] The server sends the response to the terminal, which displays the response to the user.

[1760] 6. Users provide feedback

[1761] The user provides feedback such as "somewhat satisfied," and the device sends this feedback to the server.

[1762] This system allows users to receive prompt and emotional information quickly and effectively without having to worry about the optimal AI engine, and the system's accuracy is continuously improved through feedback.

[1763] The processing flow will be explained below.

[1764] Step 1:

[1765] A user types a prompt into an input field on the device. For example, the user types "Tell me about the latest AI technologies."

[1766] Step 2:

[1767] The emotion engine analyzes the user's input, facial expressions, voice, etc. to recognize the user's emotions. For example, it can detect "interest" as an emotion.

[1768] Step 3:

[1769] The device sends the input prompt and emotion data to the server as JSON format data using an HTTP POST request.

[1770] Step 4:

[1771] The server receives the prompt and emotion data sent from the device and parses the received data into an appropriate data format (e.g., JSON).

[1772] Step 5:

[1773] The server analyzes the prompt content and emotional data using text analysis and emotion recognition algorithms, including the type of question, the depth of information requested, vocabulary complexity, and the user's emotions.

[1774] Step 6:

[1775] Based on the analysis results, the server refers to the preset evaluation score, past feedback data, and emotion data to select the optimal generative AI engine and its version. For example, if the question is technical and the emotion of "interest" is detected, generative AI engine 4 will be selected.

[1776] Step 7:

[1777] The server sends the prompt and emotion data to the selected generative AI engine using an API request.

[1778] Step 8:

[1779] A generative AI engine receives the prompt and emotion data and generates an appropriate response, which is then sent back to the server.

[1780] Step 9:

[1781] The server then formats the response as needed, specifically converting the JSON data into HTML.

[1782] Step 10:

[1783] The server sends the formatted response to the device, which then sends it to the device as an HTTP response.

[1784] Step 11:

[1785] The device receives the response from the server and displays it to the user.

[1786] Step 12:

[1787] Users provide feedback on the generated responses, including satisfaction, additional comments, and sentiment data.

[1788] Step 13:

[1789] The device sends the user feedback to the server, and the feedback data is sent in JSON format again.

[1790] Step 14:

[1791] The server receives user feedback and updates the selection algorithm, which improves the accuracy of engine selection from next time onwards.

[1792] Example 2

[1793] 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."

[1794] Conventional generative AI systems require manual selection of engines and versions to generate optimal responses to input prompts, resulting in increased computational costs and time for data processing and selection. Furthermore, it is difficult to automatically generate responses based on user emotions, and systems for improving user satisfaction are lacking. Furthermore, there is an insufficient mechanism for continuous improvement by aggregating and incorporating user feedback.

[1795] 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.

[1796] In this invention, the server includes means for receiving an input prompt from a user and recognizing the user's emotional data, means for analyzing the input prompt and the emotional data and selecting an optimal generating AI engine and its version, and means for sending the prompt to the selected generating AI engine and receiving a generated response, thereby enabling the selection of an optimal engine according to the user's emotional state and the rapid generation of a response.

[1797] "User" means an individual or entity that utilizes the System to enter prompts and receive information or services.

[1798] An "input prompt" is text data such as a question or request that a user inputs to a system.

[1799] "Emotion data" refers to data that indicates the emotional state of the user as recognized from facial expressions, voice, etc.

[1800] A "generative artificial intelligence engine" is an artificial intelligence model for generating appropriate responses based on user input prompts.

[1801] "Version" refers to different releases or revisions of the Generative Artificial Intelligence Engine.

[1802] "Analysis" refers to the process of verifying input prompts, emotional data, past feedback data, etc., and processing the data to select the optimal generative AI engine and its version.

[1803] A "response" is response data generated by a generative artificial intelligence engine in response to a user's input prompt.

[1804] "Feedback" refers to reaction data such as ratings and comments provided by users in response to generated responses.

[1805] A "selection algorithm" is a computational method or set of rules for analyzing prompts and emotional data and selecting the optimal generative AI engine and its version.

[1806] The "system" is a collection of components that executes a series of processes: receiving prompts and emotional data from users, selecting the optimal generative artificial intelligence engine to generate a response, aggregating feedback, and updating the algorithm.

[1807] This invention is a system that automatically selects the optimal generative AI engine and its version for the prompt entered by the user, and further combines it with an emotion engine that recognizes the user's emotions, thereby reducing calculation costs and time and providing a more advanced response. This system is implemented through interactions between elements including the user, terminal, server, and emotion engine.

[1808] Explanation of program processing

[1809] 1. User prompt input and emotion recognition

[1810] Users enter any question or request into an input field via a device such as a smartphone or computer. At the same time, the emotion engine uses the device's camera and microphone to analyze the user's facial expressions and voice and recognize their emotions. For example, when a user types "What's the weather like today?", the device detects the user's smile and identifies it as the emotion of "joy."

[1811] 2. Sending prompts and emotion data

[1812] The device compiles the input prompt and recognized emotion data in JSON format and sends it to the server using an HTTP POST request, specifically using the fetch API or XMLHttpRequest.

[1813] 3. Accepting and parsing prompts

[1814] The server receives the prompt and emotion data sent from the device and parses it into the appropriate format. It uses the JSON.parse method to parse the data and stores the prompt and emotion data in separate variables. During this process, it also references past feedback data and evaluation scores stored in a database (e.g., MySQL or PostgreSQL).

[1815] 4. Selection of the optimal engine

[1816] The server selects the optimal AI engine and its version based on the prompt analysis results and emotion data, along with evaluation scores and feedback data. For example, Engine 3.5 is selected for simple questions, while Engine 4 is selected for detailed technical questions.

[1817] 5. Sending prompts and generating responses

[1818] The server sends a prompt to the selected generative AI engine, which then generates an appropriate response. The generative AI engine takes into account the received prompt and emotional data to create an appropriate response and sends it back to the server. For example, a response might be generated: "The weather is sunny today."

[1819] 6. Sending the response and displaying it to the user

[1820] The server sends the generated response to the device, which then displays it to the user. The device screen displays "Today's weather is sunny."

[1821] 7. Gathering feedback and updating the algorithm

[1822] The user inputs feedback on the displayed response into the terminal, including options such as "very satisfied" and "somewhat satisfied." The terminal then transmits the user's feedback to the server.

[1823] The server stores the received feedback in a database and updates the selection algorithm based on that feedback, which will result in better engine selection in the future.

[1824] Specific examples

[1825] For simple questions

[1826] 1. User: Enter "What's the weather like today?" The device detects "happiness" from the user's facial expression.

[1827] 2. Device: Sends the prompt and the "happiness" emotion data to the server.

[1828] 3. Server: Analyzes prompt and sentiment data and selects engine 3.5.

[1829] 4. Server: Sends a prompt to Engine 3.5 and receives the response "The weather is sunny today."

[1830] 5. Server: Sends the response to the terminal, which displays it to the user.

[1831] 6. User: Gives feedback that he is "very satisfied," and the device sends the feedback to the server.

[1832] For complex technical questions

[1833] 1. User: Type "Please tell me more about neural network regularization techniques." The device detects "interests" from the user's voice.

[1834] 2. Device: Sends prompt and "interest" emotion data to the server.

[1835] 3. Server: Analyzes prompt and emotion data and selects Engine 4.

[1836] 4. Server: Sends a prompt to Engine 4 and receives a response saying, "Neural network regularization techniques include dropout, L1 regularization, and L2 regularization..."

[1837] 5. Server: Sends the response to the terminal, which displays it to the user.

[1838] 6. User: Gives feedback that they are "somewhat satisfied," and the device sends that feedback to the server.

[1839] The system selects the most suitable AI engine and provides fast and effective information by simply inputting prompts and emotional data, while the accuracy of the system is continuously improved through feedback.

[1840] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1841] Divide the processing flow of the system program into processing steps

[1842] Step 1: User prompt input and emotion recognition

[1843] 1.1 Prompt Input

[1844] 1.2 Emotion recognition

[1845] Step 2: Sending prompts and emotion data

[1846] Step 3: Server accepts and parses prompt

[1847] 3.1 Receiving prompts and emotion data

[1848] 3.2 Prompt and Emotion Data Analysis

[1849] Step 4: Selecting the optimal engine

[1850] 4.1 Referencing evaluation scores and feedback data

[1851] 4.2 Selecting the optimal engine and version

[1852] Step 5: Sending prompts and generating responses

[1853] 5.1 Sending Prompts to the Selection Engine

[1854] 5.2 Engine Generates and Returns Response

[1855] Step 6: Sending the Response and Displaying it to the User

[1856] Step 7: Gather feedback and update the algorithm

[1857] 7.1 Entering and Submitting Feedback

[1858] 7.2 Receiving Feedback and Updating the Algorithm

[1859] Specific explanation of each processing step

[1860] Step 1: User prompt input and emotion recognition

[1861] 1.1 Prompt Input

[1862] The user writes a question or request into the input field of the terminal. For example, "What's the weather like today?" Input: The text data entered by the user into the terminal. Output: The prompt text entered.

[1863] 1.2 Emotion recognition

[1864] The device uses the built-in camera and microphone to collect the user's facial expressions and voice, and the emotion engine analyzes their emotions in real time. For example, if the user smiles, the emotion "joy" is detected. Input: User's facial and voice data. Output: Analyzed emotion data (e.g., "joy").

[1865] Step 2: Sending prompts and emotion data

[1866] The device collects the input prompt and emotion data and sends them to the server in JSON format. Specifically, it makes an HTTP POST request using the fetch API or XMLHttpRequest. Input: Prompt text and emotion data. Output: JSON data sent to the server.

[1867] Step 3: Server accepts and parses prompt

[1868] 3.1 Receiving prompts and emotion data

[1869] The server receives the prompt and emotion data sent from the device. Input: JSON data sent from the device. Output: Received prompt and emotion data.

[1870] 3.2 Prompt and Emotion Data Analysis

[1871] The server parses the prompt and emotion data using the JSON.parse method and stores them in separate variables. Input: Received prompt and emotion data. Output: Parsed prompt and emotion data.

[1872] Step 4: Selecting the optimal engine

[1873] 4.1 Referencing evaluation scores and feedback data

[1874] The server references past feedback data and evaluation scores from a database. For example, queries are performed using MySQL or PostgreSQL. Input: prompt text and emotion data. Output: evaluation scores and feedback data required to select the optimal engine.

[1875] 4.2 Selecting the optimal engine and version

[1876] The server selects the optimal generative AI engine and its version based on the prompt, emotion data, evaluation scores, and feedback data. Input: Evaluation scores and feedback data. Output: Selected optimal engine and its version.

[1877] Step 5: Sending prompts and generating responses

[1878] 5.1 Sending Prompts to the Selection Engine

[1879] The server sends a prompt to the selected engine. Specifically, it sends an HTTP request to the engine's API endpoint. Input: Selected engine and prompt text. Output: Sending prompt data to the generative AI engine.

[1880] 5.2 Engine Generates and Returns Response

[1881] The generative AI model analyzes the prompt and emotion data, generates an appropriate response, and sends it back to the server. For example, the response "Today's weather is sunny." Input: Prompt sentence and emotion data. Output: Generated response sentence.

[1882] Step 6: Sending the Response and Displaying it to the User

[1883] The server sends the generated response to the device. The device receives the response and displays it to the user. For example, "Today's weather is sunny" is displayed on a smartphone screen. Input: The generated response text. Output: The response text displayed to the user.

[1884] Step 7: Gather feedback and update the algorithm

[1885] 7.1 Entering and Submitting Feedback

[1886] The user inputs feedback about the displayed response into the terminal. For example, the user inputs feedback such as "very satisfied." The terminal sends the feedback to the server. Input: User feedback. Output: Feedback data sent to the server.

[1887] 7.2 Receiving Feedback and Updating the Algorithm

[1888] The server stores the received feedback in a database and updates the selection algorithm, which will result in better prompt selection in future. Input: Submitted feedback data. Output: Updated selection algorithm.

[1889] (Application example 2)

[1890] 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."

[1891] Conventional generative AI systems did not take user emotions into consideration, limiting the quality of the user experience. Furthermore, selecting the optimal AI engine was difficult and required significant time and computational costs. This often led to dissatisfaction with the quality and timeliness of information provided.

[1892] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1893] In this invention, the server includes means for analyzing the input prompt and selecting the optimal generative AI engine and its version, means for recognizing the user's emotions and acquiring emotion data, and means for transmitting the acquired emotion data to the server and selecting the optimal generative AI engine together with the prompt analysis result, thereby making it possible to provide a more appropriate response that takes the user's emotions into consideration and improve the quality of the user experience.

[1894] An "input prompt" is a question or request that a user provides to a system.

[1895] A "generative artificial intelligence engine" refers to an algorithm or system for generating responses in natural language based on input prompts.

[1896] "Emotion data" refers to information about emotions extracted from a user's facial expressions, voice, text input, etc.

[1897] "Server" refers to a central processing unit for analyzing input data and providing generated responses to users.

[1898] "Prompt analysis result" refers to information obtained by the server analyzing the input prompt.

[1899] "Optimal engine selection" refers to the process of selecting the appropriate generative AI engine and its version based on prompt analysis results and emotion data.

[1900] "Feedback" refers to the opinions and ratings that users provide in response to responses they receive.

[1901] "Selection algorithm" refers to a set of calculation procedures and rules for selecting the optimal generative artificial intelligence engine and its version.

[1902] System Configuration

[1903] The system based on the present invention allows users to communicate with emotion recognition using generative AI models. This system includes the following elements:

[1904] 1. Parsing the input prompt

[1905] The server analyzes the prompts entered by the user through the terminal. The prompts contain questions or requests from the user, and performs syntactic and semantic analysis using natural language analysis techniques.

[1906] 2. Acquiring Emotion Data

[1907] The device acquires the user's emotion data related to the input prompt. The emotion data is recognized from the user's facial expressions, voice, text input, etc. using an emotion engine (e.g., a model using OpenCV or TensorFlow).

[1908] 3. Selection of the optimal engine

[1909] The server selects the optimal generative AI engine and its version based on the prompt analysis results and emotion data. The server references past evaluation scores and feedback data and determines the optimal engine using an engine selection algorithm.

[1910] 4. Retrieving the generated response

[1911] The server sends prompts to the selected generative artificial intelligence engine and receives a generated response, which is provided in natural language format and includes an appropriate response to the user's prompt.

[1912] 5. Displaying the response to the user

[1913] The device displays the received responses to the user, and the responses are generated taking into account the user's emotional state, resulting in a satisfying interaction experience for the user.

[1914] 6. Gathering feedback and updating the algorithm

[1915] The device sends user feedback to the server, which then updates the engine selection algorithm based on the feedback to achieve better engine selection in the future.

[1916] Hardware and software used

[1917] Hardware

[1918] Devices (smartphones, smart glasses, head-mounted displays, etc.)

[1919] Server (high-performance data analysis server)

[1920] software

[1921] Emotion engine (OpenCV, TensorFlow, Keras)

[1922] HTTP request processing library (requests)

[1923] Generative AI engine (general-purpose generative AI model)

[1924] Specific examples

[1925] If a user types "What's the latest trending item?" and makes a surprised expression, the emotion engine will detect the "surprise" emotion. This prompt and emotion data will be sent to the server. The server will analyze the prompt and emotion data, select the optimal generative AI engine, and generate a response from the engine saying, "The latest trending item is a new smartwatch," which will be displayed to the user via their device.

[1926] Prompt Sentence Examples

[1927] What's the latest trending item?

[1928] Thus, the present invention is a system that enhances the user experience by recognizing the user's emotions and providing optimal responses based on them.

[1929] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1930] Step 1:

[1931] The user inputs a prompt into a device (such as a smartphone, smart glasses, or head-mounted display). The device then captures the user's input as text data. It also simultaneously records the user's facial expressions and voice, capturing them as emotion data. This allows the device to collect the prompt and emotion data as input.

[1932] Step 2:

[1933] The device sends the collected prompt sentences and emotion data to the server. This data is converted into JSON format and sent using an HTTP POST request. Specifically, the device converts the collected data into JSON and sends it to the server using the HTTP request library (requests).

[1934] Step 3:

[1935] The server analyzes the received prompt and emotion data. It parses the JSON data received as input and extracts the prompt and emotion data. It then performs syntactic and semantic analysis of the prompt using a natural language processing library (e.g., spaCy, NLTK). Meanwhile, it uses an emotion engine (e.g., TensorFlow, Keras) to classify the emotion data.

[1936] Step 4:

[1937] The server selects the optimal generative AI engine and its version based on the analysis results and emotional data. It uses past evaluation scores and feedback data to select the optimal engine using an algorithm. The algorithm uses the prompt type, the user's emotional state, and past engine performance data as inputs, and outputs the selection result.

[1938] Step 5:

[1939] The server sends a prompt to the selected generative AI engine and receives the generated response. Specifically, the server calls the selected engine's API and sends the prompt. The engine analyzes the prompt, generates an appropriate natural language response, and sends it back to the server. This response becomes the server's output.

[1940] Step 6:

[1941] The server sends the generated response to the terminal. The response data is converted to JSON format and sent to the terminal using an HTTP POST request. Specifically, the server converts the generated response to JSON and sends it to the terminal using the HTTP request library (requests).

[1942] Step 7:

[1943] The device displays the received response to the user. The user can then enter feedback such as satisfaction or additional comments. This feedback is also captured as text data and used in a later step. The device then converts the captured feedback into JSON format.

[1944] Step 8:

[1945] The device sends user feedback to the server. This feedback data is sent using an HTTP POST request. Specifically, the device converts the feedback data into JSON and sends it to the server using the HTTP request library (requests).

[1946] Step 9:

[1947] The server analyzes the received feedback and updates the engine selection algorithm. The analysis includes semantic analysis of the feedback comments and weighted averaging of the evaluation scores. Based on the analysis results, the engine selection algorithm is tuned to achieve more accurate selection in the future.

[1948] 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.

[1949] 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.

[1950] 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.

[1951] 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.

[1952] 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.

[1953] 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.

[1954] 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).

[1955] 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.

[1956] 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."

[1957] 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.

[1958] 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).

[1959] 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.

[1960] 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.

[1961] 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.

[1962] 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.

[1963] 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.

[1964] 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.

[1965] 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.

[1966] 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.

[1967] 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.

[1968] 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.

[1969] The following is further disclosed regarding the above embodiment.

[1970] (Claim 1)

[1971] A means for analyzing the input prompt and selecting an optimal generating artificial intelligence engine and its version;

[1972] means for sending the prompt to a selected generating artificial intelligence engine and receiving a generated response;

[1973] a means for transmitting the generated response to the user;

[1974] means for receiving feedback from users and updating the selection algorithm of said generative artificial intelligence engine;

[1975] A system including:

[1976] (Claim 2)

[1977] The system of claim 1, further comprising means for selecting an optimal generative artificial intelligence engine and its version based on the analyzed prompts and by referring to a preset evaluation score and feedback data.

[1978] (Claim 3)

[1979] 10. The system of claim 1, further comprising means for aggregating user feedback and updating the selection algorithm of the generative artificial intelligence engine to achieve more appropriate engine selection in the future.

[1980] "Example 1"

[1981] (Claim 1)

[1982] A means for analyzing the input prompt and selecting an optimal generating artificial intelligence engine and its version;

[1983] means for sending the prompt to a selected generating artificial intelligence engine and receiving a generated response;

[1984] means for transmitting the generated response to the terminal and displaying it to the user;

[1985] means for receiving feedback from a user and updating the selection algorithm of said generative artificial intelligence engine;

[1986] user interface means for displaying the generated response to a user;

[1987] A system including:

[1988] (Claim 2)

[1989] The system of claim 1, further comprising means for selecting an optimal generative artificial intelligence engine and its version based on the analyzed prompts and by referring to a preset evaluation score and feedback data.

[1990] (Claim 3)

[1991] 10. The system of claim 1, further comprising means for aggregating user feedback and updating the selection algorithm of the generative artificial intelligence engine to achieve more appropriate engine selection in the future.

[1992] "Application Example 1"

[1993] (Claim 1)

[1994] A means for analyzing the input prompt and selecting an optimal generating artificial intelligence engine and its version;

[1995] means for sending the prompt to a selected generating artificial intelligence engine and receiving a generated response;

[1996] a means for transmitting the generated response to the user;

[1997] means for receiving feedback from users and updating the selection algorithm of said generative artificial intelligence engine;

[1998] a means for providing customized content based on the generated response; and

[1999] A system including:

[2000] (Claim 2)

[2001] The system of claim 1, further comprising means for selecting an optimal generative artificial intelligence engine and its version based on the analyzed prompts and by referring to a preset evaluation score and feedback data.

[2002] (Claim 3)

[2003] 10. The system of claim 1, further comprising means for aggregating user feedback and updating the selection algorithm of the generative artificial intelligence engine to achieve more appropriate engine selection in the future.

[2004] "Example 2: Combining Emotion Engines"

[2005] (Claim 1)

[2006] means for receiving an input prompt from a user and recognizing emotional data of the user;

[2007] A means for analyzing the input prompt and emotion data and selecting an optimal generative artificial intelligence engine and its version;

[2008] means for sending the prompt to a selected generating artificial intelligence engine and receiving a generated response;

[2009] a means for transmitting the generated response to the user;

[2010] means for receiving feedback from users and updating the selection algorithm of said generative artificial intelligence engine;

[2011] A system including:

[2012] (Claim 2)

[2013] The system of claim 1, further comprising means for selecting an optimal generative artificial intelligence engine and its version based on the analyzed prompt and emotion data and by referring to a pre-set evaluation score and feedback data.

[2014] (Claim 3)

[2015] 10. The system of claim 1, further comprising means for aggregating user feedback and updating the selection algorithm of the generative artificial intelligence engine to achieve more appropriate engine selection in the future.

[2016] "Application example 2 when combining emotion engines"

[2017] (Claim 1)

[2018] A means for analyzing the input prompt and selecting an optimal generating artificial intelligence engine and its version;

[2019] means for sending the prompt to a selected generating artificial intelligence engine and receiving a generated response;

[2020] a means for transmitting the generated response to the user;

[2021] A means for recognizing a user's emotion and acquiring emotion data;

[2022] A means for transmitting the acquired emotion data to a server and selecting the optimal generating AI engine together with the prompt analysis results;

[2023] means for receiving feedback from users and updating the selection algorithm of said generative artificial intelligence engine;

[2024] A system including:

[2025] (Claim 2)

[2026] The system of claim 1, further comprising means for selecting an optimal generative artificial intelligence engine and its version based on the analyzed prompt and the acquired emotional data, with reference to a pre-set evaluation score and feedback data.

[2027] (Claim 3)

[2028] 10. The system of claim 1, further comprising means for aggregating user feedback and sentiment data and updating the selection algorithm of the generative artificial intelligence engine to achieve more appropriate engine selection in the future. [Explanation of symbols]

[2029] 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 analyzing the input prompt and selecting an optimal generating artificial intelligence engine and its version; means for sending the prompt to a selected generating artificial intelligence engine and receiving a generated response; a means for transmitting the generated response to the user; means for receiving feedback from users and updating the selection algorithm of said generative artificial intelligence engine; A system including:

2. The system according to claim 1, further comprising means for selecting an optimal generative artificial intelligence engine and its version based on the analyzed prompt and by referring to a preset evaluation score and feedback data.

3. The system of claim 1 further comprising means for aggregating user feedback and updating the selection algorithm of the generative artificial intelligence engine to achieve more appropriate engine selection in the future.

Images