System

Abstract

To provide a system.SOLUTION: A system comprising: means for acquiring voice data of user talking in his / her sleep; means for converting the voice data into text data; means for generating a dream scenario based on the text data using a generative AI model; means for discovering potential stress held by the user based on the generated dream scenario; and means for suggesting a way to relieve the discovered stress.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] There is no method to improve sleep quality by detecting potential stress that individuals are not aware of and suggesting ways to relieve it. [Means for solving the problem]

[0005] Sleep talking during sleep is collected and the content of dreams is inferred from that data using generative AI. The system provides a method to identify potential stress an individual may be experiencing from the inferred dream content and suggest ways to alleviate it. Furthermore, the generative AI can use deep learning to infer dream content, making it possible to use it in mental health treatment for psychiatric patients. [Brief explanation of the drawings]

[0006] [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. 2 is a sequence diagram showing a flow of processing in the data processing system according to the first embodiment of the first form example. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1. [Figure 13] FIG. 10 is a sequence diagram showing a processing flow of a data processing system in a second embodiment of the second form example. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Embodiment Example 2. [Figure 15] FIG. 10 is a sequence diagram showing the flow of processing in a data processing system according to a third embodiment of the third embodiment. [Figure 16] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Embodiment 3. [Figure 17] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Form Example 1 when an emotion engine is combined. [Figure 19] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the second embodiment of the second form example when an emotion engine is combined. [Figure 20] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Form Example 2 when an emotion engine is combined. [Figure 21] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the third embodiment of the third form example when an emotion engine is combined. [Figure 22] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Form Example 3 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0007] 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.

[0008] First, the terms used in the following description will be explained.

[0009] 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, the 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), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] 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.

[0011] 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.

[0012] 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.

[0013] 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."

[0014] [First embodiment]

[0015] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0016] 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.

[0017] 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).

[0018] 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.

[0019] 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.

[0020] 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.

[0021] 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.

[0022] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0023] 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.

[0024] 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.

[0025] 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.

[0026] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0027] "Example 1"

[0028] This embodiment of the present invention is composed of a microphone that collects sleep talking during sleep, a generative AI that analyzes the collected sleep talking and infers the content of dreams, and an advice generation module that discovers potential stress an individual is experiencing from the inferred dream content and suggests ways to alleviate it. Specifically, the microphone collects sleep talking during sleep, and the data is sent to the generative AI. The generative AI uses deep learning to infer the content of dreams from the sleep talking and sends the results to the advice generation module. The advice generation module discovers stress from the inferred dream content and suggests ways to alleviate it.

[0029] "Example 2"

[0030] In another aspect of the present invention, a generative AI uses deep learning to infer dream content. Specifically, the generative AI receives sleep talking data as input and uses a deep learning model to infer dream content. The deep learning model is trained using a large amount of sleep talking data and corresponding dream content data.

[0031] "Example 3"

[0032] In a further aspect of the present invention, the system can be used in the mental health treatment of psychiatric patients. Specifically, the system infers the content of dreams from the sleep talking of psychiatric patients, and identifies the stress the patient is experiencing from the content of those dreams. The system then suggests ways to relieve that stress, contributing to the improvement of the patient's mental health.

[0033] The processing flow of each embodiment will be described below.

[0034] "Example 1"

[0035] Step 1: A microphone collects your sleep talking while you sleep.

[0036] Step 2: The collected sleep talking data is sent to the generative AI.

[0037] Step 3: The generative AI uses deep learning to infer the content of the dream from the sleep talking.

[0038] Step 4: The estimated dream content is sent to the advice generation module.

[0039] Step 5: The advice generation module detects stress from the inferred dream content and suggests ways to relieve it.

[0040] "Example 2"

[0041] Step 1: The generative AI receives sleep talking data as input.

[0042] Step 2: The generative AI uses a deep learning model to infer the content of the dream.

[0043] Step 3: The deep learning model is trained using a large amount of sleep talking data and corresponding dream content data.

[0044] "Example 3"

[0045] Step 1: Infer the content of a psychiatric patient's dreams from their sleep talking.

[0046] Step 2: Discover the stress the patient is experiencing from the content of the dream you have guessed.

[0047] Step 3: Suggest ways to relieve the stress discovered and contribute to improving the patient's mental health.

[0048] Example 1

[0049] Next, a description will be given of Example 1 of Form 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."

[0050] Conventional sleep talking analysis systems have insufficient collection and analysis of sleep talking, making it difficult to accurately identify potential stress in users and suggest ways to alleviate it. Furthermore, the accuracy of converting voice data into text and predicting dream content is low, resulting in a lack of reliability in the advice given to users. This has led to the issue of insufficient efforts to improve users' mental health.

[0051] 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.

[0052] In this invention, the server includes means for collecting sleep talking during sleep, means for saving the collected sleep talking as audio data, means for transmitting the saved audio data to the server, means for converting the audio data into text data, means for inferring the content of dreams from the converted text data using a generative AI model, means for discovering potential stress the individual is experiencing from the inferred dream content, and means for suggesting ways to relieve the discovered stress. This makes it possible to analyze the user's sleep talking with high precision, accurately discover potential stress, and reliably suggest ways to relieve it.

[0053] The "means for collecting sleep talking during sleep" refers to a device or mechanism for collecting sleep talking uttered by a user while sleeping with high sensitivity.

[0054] The "means for saving collected sleep talking as audio data" refers to a device or mechanism for recording and saving collected sleep talking as digital audio data.

[0055] The "means for transmitting the stored voice data to the server" refers to a device or mechanism for transmitting the stored voice data to the server via a network.

[0056] "Means for converting voice data into text data" refers to a device or mechanism for converting voice data into text information using voice recognition technology.

[0057] A "means for inferring dream content using a generative AI model" is a device or mechanism that uses a deep learning algorithm to analyze and infer dream content from text data.

[0058] A "means for discovering potential stress experienced by an individual" is a device or mechanism for identifying potential stress experienced by a user based on the content of the inferred dream.

[0059] The "means for proposing a method for relieving the discovered stress" is a device or mechanism for proposing to the user a method for relieving the identified stress.

[0060] MODE FOR CARRYING OUT THE INVENTION

[0061] This invention is a system that collects sleep talking during sleep, analyzes the data to infer the content of dreams, discovers potential stress that an individual is experiencing, and suggests ways to alleviate that stress. Specific embodiments of this system are described below.

[0062] Hardware and software used

[0063] 1. Microphone

[0064] To collect sleep talking, a highly sensitive microphone is used, which is designed to clearly record the user's sleep talking.

[0065] 2. Terminal

[0066] A device is used to store the collected voice data and send it to a server. This device has the function of storing voice data in real time and sending it to a server using a secure protocol (e.g., HTTPS).

[0067] 3. Server

[0068] The server converts the received audio data into text using speech recognition software (e.g., Google® Cloud Speech-to-Text API), which is then passed to a generative AI model and used to infer the content of the dream.

[0069] 4. Generative AI Models

[0070] The generative AI model uses a deep learning algorithm to infer dream content from text data. Specifically, it analyzes the text data using natural language processing technology and generates dream scenarios.

[0071] 5. Advice Generation Module

[0072] The advice generation module identifies potential stress the user may be experiencing based on the estimated dream content and suggests ways to alleviate the stress. The suggestions include feasible and effective methods for the user. The advice generation module may include a generative AI model.

[0073] Specific examples

[0074] If a user says in their sleep, "My work isn't going well," the microphone connected to the device will collect this speech and save it as audio data. The device will then send the collected audio data to a server, where the HTTPS protocol is used to ensure data security.

[0075] The server converts the received voice data into text data using the Google Cloud Speech-to-Text API. For example, the generated text data is "My work is not going well." The server then passes the generated text data to the generative AI model. The generative AI model analyzes this text data and infers that the user is experiencing work-related stress.

[0076] The dream content predicted by the generative AI model is sent to an advice generation module via the server. The advice generation module determines that the user is experiencing work-related stress and generates advice for the user, such as "We suggest that you set aside time for relaxation every day to reduce work stress." This suggestion is sent to the device via the server and notified to the user.

[0077] Prompt Sentence Examples

[0078] "If a user says in their sleep, 'My work isn't going well,' the generative AI model will analyze this sleep talking and infer that the user is experiencing work-related stress. What advice will the advice generation module provide to the user?"

[0079] In this way, the system analyzes the user's sleep talking, discovers potential stress, and suggests ways to relieve it.

[0080] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0081] Step 1:

[0082] While the user is sleeping, a microphone connected to the device collects the user's sleep talking. The microphone is designed to be highly sensitive and to clearly record the user's sleep talking. The input is the user's sleep talking audio, and the output is the collected audio data.

[0083] Step 2:

[0084] The device saves the collected voice data. The saved voice data is temporarily stored in the device for further processing. The input is the collected voice data, and the output is the saved voice data.

[0085] Step 3:

[0086] The device sends the stored audio data to the server using a secure protocol (e.g., HTTPS) to ensure data security. The input is the stored audio data, and the output is the audio data sent to the server.

[0087] Step 4:

[0088] The server converts the received voice data into text data using speech recognition software (e.g., Google Cloud Speech-to-Text API). The input is the voice data, and the output is the converted text data. Specifically, the speech recognition algorithm analyzes the voice data and generates the corresponding text.

[0089] Step 5:

[0090] The server passes the converted text data to the generative AI model, which uses a deep learning algorithm to infer the dream content from the text data. The input is text data, and the output is the inferred dream content. Specifically, it uses natural language processing technology to analyze the text data and generate a dream scenario.

[0091] Step 6:

[0092] The dream content inferred by the generative AI model is sent via the server to the advice generation module, which then discovers the user's potential stress based on the inferred dream content. The input is the inferred dream content, and the output is the discovered stress.

[0093] Step 7:

[0094] The advice generation module proposes ways to relieve the detected stress. These proposals include feasible and effective methods for the user. The input is the detected stress, and the output is the proposed ways to relieve it. Specific actions include relaxation techniques and ways to improve lifestyle habits.

[0095] Step 8:

[0096] The proposed solution is sent to the device via the server and notified to the user. The input is the proposed solution, and the output is a notification to the user. Specifically, the device displays a notification so that the user can confirm the proposed solution.

[0097] (Application example 1)

[0098] Next, a description will be given of Application Example 1 of Embodiment 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."

[0099] In modern society, individuals experience a wide range of potential stresses, which can affect not only their mental health but also security risks. However, while conventional systems can infer the content of dreams from sleep talking and detect potential stress, they are unable to suggest appropriate security measures in response to that stress. This creates a need for systems that can simultaneously support both the user's mental health and security.

[0100] 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.

[0101] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for detecting potential stress in an individual from the inferred dream content, means for suggesting ways to relieve the detected stress, and means for recommending appropriate security measures according to the stress level, thereby supporting the user's mental health and reducing security risks.

[0102] "Somnitalking" refers to speech that is involuntarily uttered by an individual while they are asleep.

[0103] "Means for collecting" refers to a device or method for acquiring audio data.

[0104] "Generative AI" is a system that uses artificial intelligence technology to generate new information from data.

[0105] A "means for inferring the content of dreams" is a method or device for analyzing collected data to infer the content of dreams.

[0106] A "means for detecting potential stress" is a method or device for identifying the stress an individual is experiencing from the inferred content of dreams.

[0107] A "means for suggesting ways to relieve stress" is a method or device that suggests appropriate ways to deal with discovered stress.

[0108] The "means for recommending security measures" is a method or device that suggests appropriate security measures to a user depending on the user's stress level.

[0109] "Deep learning" is a technology that uses multi-layered neural networks to analyze data and learn patterns and features.

[0110] A "mentally ill person" is an individual who has a mental illness.

[0111] "Mental health care" is a medical procedure that evaluates mental health status and provides treatment and support.

[0112] The following system configuration will be described as an embodiment of the present invention.

[0113] System Configuration

[0114] The system is centered around the "Sleep Security Assistant" application installed on the user's smartphone. The system includes the following main components:

[0115] 1. A method for collecting sleep talking

[0116] The smartphone microphone is used to collect the sleep talking of the user. The audio data is converted into text data using a speech recognition library (e.g., speech_recognition library).

[0117] 2. A way to guess the content of dreams

[0118] The collected text data of sleep talking is sent to a generative AI model (e.g., the rinna / japanese-gpt2-medium model using the transformers library) to infer the content of dreams. This generative AI model uses deep learning to analyze the content of dreams from the text data.

[0119] 3. Ways to detect underlying stress

[0120] The inferred dream content is used to discover potential stress levels of individuals, assess stress levels, and understand the user's mental health.

[0121] 4. Suggesting ways to relieve stress

[0122] The system suggests appropriate ways to relieve stress, such as deep breathing to relax or recommending activities that will help relieve stress.

[0123] 5. Means of recommending security measures

[0124] The system suggests appropriate security measures to users based on their stress level. For example, if stress increases, the system suggests changing passwords and recommends setting up two-factor authentication.

[0125] Processing flow

[0126] 1. Audio collection and recognition

[0127] The sleep talking uttered by the user while sleeping is collected using the smartphone's microphone and converted into text data using a voice recognition library.

[0128] 2. Guessing the content of the dream

[0129] The collected text data is input into a generative AI model to infer the content of dreams. The generative AI model generates dream content using the following prompt sentences:

[0130] Prompt: "User says in their sleep: 'Work isn't going well.' Infer the content of their dream from this sleep rambling."

[0131] 3. Identifying and assessing stress

[0132] The content of the dream is inferred to identify the individual's potential stress and assess its level.

[0133] 4. Generating Advice

[0134] The system suggests appropriate ways to relieve stress, such as "Take deep breaths to relax."

[0135] 5. Proposal of security measures

[0136] The system suggests appropriate security measures to users depending on their stress level. For example, if stress increases, the system will recommend specific measures such as "change your password" or "enable two-factor authentication."

[0137] In this way, it is possible to support the mental health of users and reduce security risks.

[0138] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0139] Step 1:

[0140] Sleep talking uttered by a user while sleeping is collected.

[0141] Input: Audio data from the smartphone microphone.

[0142] Specific operation: The device (smartphone) uses a microphone to collect the user's sleep talking in real time. The collected voice data is converted into text data using a speech recognition library (e.g., speech_recognition library).

[0143] Output: Sleep talking as text data.

[0144] Step 2:

[0145] The collected text data of sleep talking is input into a generative AI model to infer the content of the dream.

[0146] Input: Text data obtained in step 1.

[0147] Specific operation: The server inputs text data into a generative AI model (e.g., the rinna / japanese-gpt2-medium model using the transformers library) to infer the content of the dream. The generative AI model generates the content of the dream using the following prompt sentence:

[0148] Prompt: "User says in their sleep: 'Work isn't going well.' Infer the content of their dream from this sleep rambling."

[0149] Output: Inferred dream content.

[0150] Step 3:

[0151] The content of the dream is inferred to identify the individual's potential stress and assess its level.

[0152] Input: Dream content obtained in step 2.

[0153] How it works: The server analyzes the inferred dream content to discover the individual's potential stress. It uses a pre-trained stress assessment model to assess the stress level.

[0154] Output: Stress level assessment results.

[0155] Step 4:

[0156] We will suggest appropriate ways to relieve any stress that is discovered.

[0157] Input: Stress level assessment results obtained in step 3.

[0158] What it does: The server suggests ways to relieve stress based on your stress level, such as taking deep breaths to relax or recommending activities that will help you relieve stress.

[0159] Output: Suggestions for stress relief.

[0160] Step 5:

[0161] Appropriate security measures are suggested to users based on their stress level.

[0162] Input: Stress level assessment results obtained in step 3.

[0163] Specific behavior: When the stress level increases, the server suggests appropriate security measures to the user, such as suggesting a password change or setting up two-factor authentication.

[0164] Output: Suggested security measures.

[0165] Example 2

[0166] Next, a description will be given of Example 2 of Form 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."

[0167] In conventional mental health treatment, it has been difficult to accurately grasp the potential stress that an individual is experiencing. In addition, there has been a lack of methods for analyzing the content of dreams, making it impossible to understand an individual's psychological state in detail. This has made it difficult to suggest appropriate stress relief methods.

[0168] 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.

[0169] In this invention, the server includes a means for collecting voice data during sleep, a means for converting the collected voice data into text data using voice recognition software, and a means for inputting the converted text data into a deep learning model to infer the content of dreams. This makes it possible to accurately infer the content of dreams from an individual's sleep talking, discover potential stress, and suggest appropriate stress relief methods.

[0170] "Audio data" refers to data that is a digital recording of the sounds made by the user while they are sleeping.

[0171] "Speech recognition software" is a program for converting voice data into text data.

[0172] "Text data" is text information converted by voice recognition software.

[0173] A "deep learning model" is an artificial intelligence model that is trained using large amounts of data to extract specific patterns and features from input data.

[0174] "Dream content" refers to the specific scenario or situation of the dream the user is having while sleeping.

[0175] "Latent stress" is stress that the user is not aware of but that has a psychological or physiological effect on the user.

[0176] "Stress relief methods" are specific measures or suggestions for reducing or eliminating the stress that has been identified.

[0177] "Mental health care" is a medical procedure that assesses mental health and provides treatment and support.

[0178] MODE FOR CARRYING OUT THE INVENTION

[0179] The present invention is a system for analyzing sleep talking uttered by a user while sleeping and predicting the content of the dream. Specific embodiments of this system will be described below.

[0180] 1. Program Generation

[0181] The server generates a generative AI program that uses deep learning to predict the content of dreams. Specifically, it builds a model using deep learning frameworks such as TENSORFLOW (registered trademark) and PyTorch.

[0182] 2. Program processing explanation

[0183] The server receives the user's sleep talking data as input. This sleep talking data is converted into text data using speech recognition software (e.g., Google Speech-to-Text API). The server then inputs this text data into a deep learning model. This deep learning model has been trained using a large amount of sleep talking data and corresponding dream content data.

[0184] The deep learning model analyzes the input sleep talking data and infers the content of the dream. The inferred results are fed back to the user. Specifically, the server sends the inferred dream content in text format to the user's device.

[0185] 3. Specific Examples

[0186] For example, if a user says in their sleep, "I'm swimming in the ocean," the server converts this sleep-talking data into text using speech recognition software, and inputs the text data "I'm swimming in the ocean" into the deep learning model. The deep learning model analyzes this text data and infers that "the user is dreaming about swimming in the ocean." This inference is sent to the user's device in text format.

[0187] 4. Examples of prompts

[0188] An example of a prompt might be:

[0189] "If a user says in their sleep, 'I'm swimming in the ocean,' how would a generative AI infer the content of their dream?"

[0190] By inputting this prompt into the generative AI model, it is possible to provide a detailed explanation of the system's processing.

[0191] In this way, the system analyzes the user's sleep talking data and infers the content of their dreams, which allows for a detailed understanding of an individual's psychological state, the detection of potential stress, and the suggestion of appropriate stress relief methods.

[0192] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0193] Step 1:

[0194] The user uses a device (e.g., a smartphone or dedicated recording device) to record their sleep talking. The device collects the user's sleep talking as audio data. This audio data is then sent to a server.

[0195] Input: User's sleep talking (audio data)

[0196] Output: Audio data sent to the server

[0197] Step 2:

[0198] The server converts the received voice data into text data using voice recognition software (e.g., Google Speech-to-Text API). Specifically, the server sends the voice data to the API and obtains the returned text data.

[0199] Input: Audio data

[0200] Output: Text data

[0201] Step 3:

[0202] The server inputs the acquired text data into a deep learning model, which is built using frameworks such as TensorFlow and PyTorch and trained using a large amount of sleep talking data and corresponding dream content data.

[0203] Input: Text data

[0204] Output: Dream content guess result

[0205] Step 4:

[0206] The server uses a deep learning model to analyze the text data and predict the content of the dream. Specifically, the model analyzes patterns in the text data and compares them with past data to predict the most likely dream content.

[0207] Input: Text data

[0208] Output: Inferred dream content

[0209] Step 5:

[0210] The server sends the predicted dream content in text format to the user's device, where the user can check the results.

[0211] Input: Inferred dream content

[0212] Output: The text message sent to the user's device.

[0213] For example, if a user says in their sleep, "I'm swimming in the ocean," the following happens:

[0214] 1. Step 1: The user records their sleep talking using a recording app on their smartphone. The device then sends the recorded audio data to the server.

[0215] 2. Step 2: The server sends the received voice data to the Google Speech-to-Text API, which converts the voice data into text data such as "swimming in the ocean" and returns it to the server.

[0216] 3. Step 3: The server inputs the text data "swimming in the ocean" into the deep learning model. The model analyzes this text data.

[0217] 4. Step 4: The deep learning model analyzes the text data and infers that "the user is dreaming about swimming in the ocean."

[0218] 5. Step 5: The server sends the guess result to the user's device as a text message saying, "You were dreaming about swimming in the ocean." The user can check the notification on their device to find out the content of their dream.

[0219] (Application example 2)

[0220] Next, a description will be given of Application Example 2 of Form 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."

[0221] Conventional dream analysis systems focus on inferring the content of dreams, but lack the ability to link the results of these inferences to the user's interests and provide specific content. This means that users are unable to not only understand the content of their dreams, but also to enjoy useful information and entertainment based on that content.

[0222] 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.

[0223] In this invention, the server includes a means for collecting sleep talking during sleep, a means for inferring dream content from the collected sleep talking using a generative AI, a means for delivering customized content relevant to the individual based on the inferred dream content, and a means for analyzing the user's interests and concerns based on the delivered content. This allows the user to not only understand the content of their dreams, but also enjoy customized content based on that content.

[0224] "Means for collecting sleep talking" refers to devices or software that record sounds made by users while they sleep and store the data.

[0225] "Means for inferring dream content from collected sleep talking using generative AI" refers to algorithms and systems that analyze recorded sleep talking data and use deep learning models to infer dream content.

[0226] "Means for delivering customized content relevant to individuals based on the content of inferred dreams" refers to a system or service for providing content such as music, video, or articles that are appropriate for users based on the content of their inferred dreams.

[0227] "Means for analyzing user interests and concerns based on delivered content" refers to algorithms or systems for analyzing users' reactions and behavior to received content and identifying users' interests and concerns.

[0228] The system for implementing this invention is mainly composed of three elements: a server, a terminal, and a user. The function of each element will be explained below.

[0229] server

[0230] The server has the following main functions:

[0231] 1. Collection of sleep talking data: The server receives and stores the sleep talking data sent from the device. The sleep talking data is a recording of the user's speech while sleeping.

[0232] 2. Dream content prediction: The server inputs the stored sleep talking data into a generative AI model and uses deep learning to predict the dream content. This generative AI model is trained using a large amount of sleep talking data and corresponding dream content data.

[0233] 3. Delivery of customized content: Based on the content of the predicted dream, the server selects customized content (music, video, articles, etc.) relevant to the user and delivers it to the device.

[0234] 4. Analysis of user interests: The server analyzes the user's reactions and behavior to the delivered content to identify the user's interests.

[0235] Terminal

[0236] The terminal has the following main functions:

[0237] 1. Recording sleep talking data: The device is equipped with a microphone to record the user's sleep talking. The recorded sleep talking data is sent to the server.

[0238] 2. Receiving and displaying content: The terminal receives the customized content delivered from the server and displays it to the user.

[0239] User

[0240] Users interact with the system as follows:

[0241] 1. Providing sleep talking data: The user uses the device to record sleep talking while sleeping.

[0242] 2. Content Usage: The user uses the customized content delivered to the device, for example, listening to music or watching videos based on the content of the predicted dream.

[0243] Hardware and software used

[0244] Hardware: Devices such as smartphones and tablets, and servers

[0245] Software: TensorFlow (running deep learning models), SoundDevice (collecting audio data), Librosa (preprocessing audio data)

[0246] Specific examples

[0247] For example, if a user talks in their sleep with the keyword "adventure," the server analyzes the sleep-talking data and infers that the dream content is adventure-related.The server then delivers customized content, such as an adventure movie trailer or an introductory article for an adventure novel, to the device.The user can then enjoy this content on their device.

[0248] Prompt Sentence Examples

[0249] "Analyze last night's sleep talking data and infer the content of your dream. Based on the inferred content of your dream, suggest related content."

[0250] The above is an embodiment of the present invention.

[0251] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0252] Step 1:

[0253] Recording of sleep talking data

[0254] The device records the user's sleep speech. Specifically, it uses a built-in microphone to collect audio data at regular intervals. The input is the user's sleep talking, and the output is the recorded audio data.

[0255] Step 2:

[0256] Sending sleep talking data

[0257] The device sends the recorded sleep talking data to a server. Specifically, the device uploads the audio data to the server via an internet connection. The input is the recorded audio data, and the output is the audio data sent to the server.

[0258] Step 3:

[0259] Saving sleep talking data

[0260] The server receives the sleep talking data sent from the device and stores it in a database. The input is the voice data sent from the device, and the output is the stored voice data.

[0261] Step 4:

[0262] Audio data preprocessing

[0263] The server preprocesses the stored sleep talking data. Specifically, it converts the audio data into MFCC (Mel-Frequency Cepstrum Coefficients) using the Librosa library. The input is the stored audio data, and the output is the preprocessed audio data.

[0264] Step 5:

[0265] Guessing the content of the dream

[0266] The server inputs the preprocessed audio data into a generative AI model to infer the dream content. Specifically, it uses TensorFlow to run a deep learning model to infer the dream content. The input is the preprocessed audio data, and the output is the inferred dream content.

[0267] Step 6:

[0268] Customized content selection

[0269] The server selects customized content relevant to the user based on the inferred dream content. Specifically, it searches a database related to the inferred dream content for appropriate content (music, video, articles, etc.). The input is the inferred dream content, and the output is the selected content.

[0270] Step 7:

[0271] Content Delivery

[0272] The server delivers the selected customized content to the terminal, specifically by transmitting content data to the terminal via an Internet connection, with the input being the selected content and the output being the content delivered to the terminal.

[0273] Step 8:

[0274] View content

[0275] The terminal receives the customized content delivered from the server and displays it to the user. Specifically, the content is played back using the terminal's display and speakers. The input is the content delivered from the server, and the output is the content displayed to the user.

[0276] Step 9:

[0277] Collecting user responses

[0278] The device records the user's actions and reactions to the delivered content. Specifically, it collects data such as the user's taps, clicks, and playback time. The input is the user's actions and reactions, and the output is the collected reaction data.

[0279] Step 10:

[0280] Analysis of user interests and concerns

[0281] The server analyzes the user's reaction data sent from the device and identifies the user's interests. Specifically, the collected reaction data is input into an analysis algorithm to infer the user's interests. The input is the collected reaction data, and the output is the inferred user's interests.

[0282] Example 3

[0283] Next, a description will be given of a third embodiment of the third embodiment. 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."

[0284] Traditional mental health treatments often rely on patients' subjective reports, making it difficult to accurately identify potential stressors. Furthermore, there are limited means of discovering stressors that patients are not aware of, making it difficult to propose effective stress relief methods. This can potentially delay the improvement of mental health in psychiatric patients.

[0285] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0286] In this invention, the server includes means for collecting voice data during sleep, means for transmitting the collected voice data to the server, means for converting the voice data into text data, means for inferring dream content from the text data using a generative AI model, means for discovering potential stress the individual is experiencing from the inferred dream content, and means for suggesting ways to relieve the discovered stress. This makes it possible to infer the content of dreams from a patient's sleep talking, identify potential stressors, and suggest appropriate ways to relieve stress.

[0287] "Sound data during sleep" refers to data recorded from sounds made by a patient while sleeping.

[0288] "Means of collection" refers to devices or software for recording audio data, including, for example, smartphones and dedicated recording devices.

[0289] "Means for sending to a server" refers to the functions and protocols for uploading collected voice data to a server via a network.

[0290] "Means for converting voice data into text data" refers to a process for converting voice data into text information using voice recognition technology, and includes, for example, voice recognition software.

[0291] A "generative AI model" refers to an artificial intelligence model that uses deep learning to perform a specific task, including, for example, a natural language processing model.

[0292] "Method of inferring dream content" refers to the process of using a generative AI model to analyze and infer dream content from text data.

[0293] "Methods for detecting potential stress" refers to the process of analyzing the content of inferred dreams to identify potential stressors for an individual.

[0294] "Means for suggesting ways to relieve stress" refers to the process of suggesting appropriate ways to relieve stress based on the stress factors identified.

[0295] This invention is a system that can be used in the mental health treatment of psychiatric patients, and it analyzes the patient's sleep talking to infer the content of their dreams, identify potential stress factors, and propose appropriate stress relief methods. A specific embodiment of this system will be described below.

[0296] 1. System Configuration

[0297] The system is mainly composed of the following hardware and software.

[0298] Device: A device used to record the patient's sleep talking. This includes smartphones and dedicated recording devices.

[0299] Server: A computer system that receives and analyzes audio data.

[0300] Speech recognition software: Software for converting voice data into text data. Examples include the Google Cloud Speech-to-Text API.

[0301] Generative AI model: An artificial intelligence model for inferring dream content from text data. Examples include OpenAI's GPT-4.

[0302] 2. System Operation

[0303] The operation of the system is as follows.

[0304] Audio data collection

[0305] The device records the patient's sleep talking while they sleep. For example, by activating a recording app on a smartphone and placing it next to the pillow at night, high-quality recordings of sleep talking can be made.

[0306] Sending audio data

[0307] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using Wi-Fi or mobile data communication. When the upload is complete, the device notifies the user.

[0308] Analysis of audio data

[0309] The server uses speech recognition software to convert the received voice data into text data, which is then used in the next step.

[0310] Guessing the content of the dream

[0311] The server uses a generative AI model to infer dream content from text data, specifically by inputting the text data as prompts and generating dream content.

[0312] The discovery of stress

[0313] The server analyzes the content of the predicted dream and identifies potential stressors the patient may be experiencing. For example, if the dream contains content about being chased, it will determine this as a sign of stress.

[0314] Suggestions for stress relief

[0315] The server then proposes appropriate stress relief methods based on the identified stress factors. The proposals are sent to the user's device. For example, it suggests relaxation techniques or counseling.

[0316] 3. Specific Examples

[0317] Users use their smartphones to record their sleep talking at night. The recorded audio data is sent to a server and converted into text using speech recognition software. A generative AI model analyzes the text data and infers the content of the dream. For example, if the dream contains content about being chased, the server will interpret this as a sign of stress and suggest relaxation techniques or counseling.

[0318] 4. Examples of prompts

[0319] "You must analyze the patient's sleep talking and infer the content of their dream. Then, you must identify the stressors that the patient is experiencing based on the content of the dream and suggest appropriate ways to relieve stress."

[0320] In this way, a system is realized that contributes to improving the mental health of psychiatric patients through cooperation between the server, the terminal, and the user. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0321] Step 1:

[0322] Audio data collection

[0323] The device records the patient's sleep talking. Specifically, it uses a smartphone or a dedicated recording device and automatically starts recording at night. The input is the patient's sleep talking, and the output is the recorded audio file. When recording is finished, the audio file is temporarily saved.

[0324] Step 2:

[0325] Sending audio data

[0326] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using Wi-Fi or mobile data communication. The input is the recorded audio file, and the output is the audio data sent to the server. When the upload is complete, the device notifies the user.

[0327] Step 3:

[0328] Analysis of audio data

[0329] The server converts the received voice data into text data using voice recognition software. Specifically, it uses Google Cloud Speech-to-Text API or similar. The input is the voice data sent to the server, and the output is the converted text data. The conversion results are stored in a database.

[0330] Step 4:

[0331] Guessing the content of the dream

[0332] The server uses a generative AI model to infer dream content from text data. Specifically, the text data is input as a prompt, and the generative AI model (e.g., OpenAI's GPT-4) is used to generate dream content. The input is the converted text data, and the output is the inferred dream content. The generated dream content is stored in a database.

[0333] Step 5:

[0334] The discovery of stress

[0335] The server analyzes the estimated dream content and identifies potential stressors for the patient. Specifically, it applies psychological algorithms to analyze the frequency of specific keywords and phrases. The input is the estimated dream content, and the output is the identified stressors.

[0336] Step 6:

[0337] Suggestions for stress relief

[0338] The server proposes appropriate stress relief methods based on the identified stress factors. Specifically, it generates suggestions for relaxation techniques and counseling. The input is the identified stress factors, and the output is the proposed stress relief methods. The suggestions are sent to the user's device as push notifications.

[0339] (Application example 3)

[0340] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[0341] In modern society, managing the mental health of psychiatric patients is an important issue. However, conventional methods make it difficult to accurately grasp a patient's underlying stress and propose appropriate ways to relieve it. In particular, technology for inferring dream content from sleep talking and detecting stress is still in its infancy, and a system for realizing effective mental health treatment is needed.

[0342] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0343] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for detecting potential stress in an individual from the inferred dream content, means for suggesting ways to relieve the detected stress, means for notifying the user of the suggested stress relief methods, and means for recording the sleep talking, inferring dream content, detecting stress, and suggesting ways to relieve it on a smartphone. This makes it possible to effectively manage the mental health of psychiatric patients, and to detect potential stress early and suggest appropriate ways to relieve it.

[0344] A "means for collecting sleep talking" is a device or software that records the sounds a user makes while sleeping and saves them as data.

[0345] "Generative AI" is a system that uses artificial intelligence technology to generate new information from input data, and in particular, uses deep learning to infer the content of dreams.

[0346] A "means for inferring the content of dreams" is an algorithm or system that analyzes collected sleep talking data and infers the content of the dream the user is having.

[0347] A "means for detecting potential stress in an individual" is a method or system for identifying potential stress factors that a user may be experiencing from the content of their inferred dreams.

[0348] The "means for suggesting ways to relieve stress" is a system that suggests ways for users to relieve stress that they can implement in response to discovered stress factors.

[0349] A "means for notifying the user" is a communication means or interface for informing the user of the proposed stress relief method.

[0350] The "means to be executed on the smartphone" refers to software and hardware for executing the above-mentioned means as an application on the smartphone.

[0351] The following describes an embodiment of the present invention. The system is implemented as an application that runs on a user's smartphone. The main functions of the system and how they are implemented are described below.

[0352] 1. Collection of sleep talking

[0353] The app uses the smartphone's microphone to collect sleep talking while the user is asleep. The app automatically starts recording when the user talks in their sleep and saves the audio data.

[0354] 2. Guessing the content of the dream

[0355] The collected sleep talking data is input into a generative AI model (e.g., GPT-3 (registered trademark)). The generative AI model uses deep learning to infer the content of dreams from sleep talking. The following prompt sentences are used in this process:

[0356] Example prompt sentence:

[0357] "Please guess the content of the dream from the following sleep talking: A scary dog ​​is chasing me."

[0358] 3. Discovering stress

[0359] We again use a generative AI model to discover potential stressors from the inferred dream content. In this process, we use the following prompt:

[0360] Example prompt sentence:

[0361] "Identify the stress in the following dream and suggest ways to relieve it: The user dreams of being chased by a dog."

[0362] 4. Suggestions for ways to relieve stress

[0363] For each stress factor detected, the generative AI model will suggest an appropriate stress relief method, which will then be notified to the user.

[0364] 5. Notice to Users

[0365] The proposed stress relief methods are notified to users using the smartphone's notification function, allowing users to implement the stress relief methods at the appropriate time.

[0366] Hardware and software used

[0367] Hardware: Smartphone with microphone

[0368] Software: Python, speech_recognition library, transformers library (generative AI model)

[0369] Specific examples

[0370] For example, if a user says in their sleep, "A scary dog ​​is chasing me," the system will record the sleep talking and use a generative AI model to infer that "the user is dreaming about being chased by a dog." It will then identify the cause of stress as "fear" and suggest ways to relieve it, such as "deep breathing techniques to relax," and notify the user via their smartphone's notification function.

[0371] In this way, the system can effectively manage the user's mental health, detect potential stress early, and suggest appropriate ways to relieve it.

[0372] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0373] Step 1:

[0374] Sleep talking uttered by a user while sleeping is collected.

[0375] Input: User's voice data

[0376] Output: Audio file of recorded sleep talking

[0377] What it does: It uses the smartphone's microphone to automatically start recording when the user talks in their sleep and saves the audio data.

[0378] Step 2:

[0379] The collected sleep talking data is converted into text.

[0380] Input: Audio file of recorded sleep talking

[0381] Output: Sleep talking data in text format

[0382] What it does: Converts audio data to text using the speech_recognition library.

[0383] Step 3:

[0384] Text-based sleep talking data is input into a generative AI model to infer the content of the dream.

[0385] Input: Text-formatted sleep talking data

[0386] Output: Inferred dream content

[0387] Specific operation: The following prompt sentence is input into a generative AI model (e.g., GPT-3) to generate the content of the dream.

[0388] "Please guess the content of the dream from the following sleep talking: [Sleep talking data in text format]"

[0389] Step 4:

[0390] The predicted dream content is fed into a generative AI model to discover potential stressors.

[0391] Input: Inferred dream content

[0392] Output: Stressors discovered

[0393] Specific operation: The following prompt sentence is input into the generative AI model to identify stress factors.

[0394] "Please identify the stress in the following dream and suggest ways to relieve it: [guessed dream content]"

[0395] Step 5:

[0396] For each stress factor identified, a generative AI model is used to suggest ways to relieve stress.

[0397] Input: Discovered stressors

[0398] Output: Suggested stress relief methods

[0399] Specific operation: Stress factors are input into the generative AI model, and appropriate stress relief methods are generated.

[0400] Step 6:

[0401] Inform the user of suggested stress relief methods.

[0402] Input: Suggested stress relief method

[0403] Output: User notification

[0404] What it does: Uses smartphone notifications to notify users of suggested stress relief methods.

[0405] Through these steps, the system can effectively manage the user's mental health, detect potential stress early, and suggest appropriate ways to relieve it.

[0406] 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.

[0407] "Example 1"

[0408] In one embodiment of the present invention, the emotion engine recognizes the user's emotions and adjusts stress relief methods based on the user's emotional state. Specifically, the emotion engine analyzes the user's emotions from their voice, facial expressions, and behavior, and provides feedback to the generative AI. The generative AI uses this information to more accurately predict the content of the user's dreams and suggest stress relief methods tailored to each individual user.

[0409] "Example 2"

[0410] In another embodiment, the emotion engine monitors the user's emotional state in real time and updates stress relief methods in response to changes in the emotional state. Specifically, when the user expresses negative emotions such as anger or sadness, the emotion engine feeds that information back to the generative AI, which then uses that information to update the stress relief methods. For example, if the user is feeling angry, the generative AI suggests stress relief methods such as relaxation or meditation.

[0411] "Example 3"

[0412] In another embodiment, the emotion engine and generative AI work together to propose stress relief methods that take into account both the user's emotional state and the content of their dreams. Specifically, the emotion engine analyzes the user's emotional state and feeds the results back to the generative AI. The generative AI combines this information with the content of their dreams inferred from their sleep talking to more accurately understand the user's stress and propose ways to relieve it.

[0413] The processing flow of each embodiment will be described below.

[0414] "Example 1"

[0415] Step 1: The emotion engine analyzes the user's emotions from their voice, facial expressions, behavior, etc.

[0416] Step 2: The emotion engine feeds back the analysis results to the generative AI.

[0417] Step 3: The generative AI uses this information to infer the content of the user's dream.

[0418] Step 4: The generative AI suggests stress relief methods that are appropriate for each individual user.

[0419] "Example 2"

[0420] Step 1: The emotion engine monitors the user's emotional state in real time.

[0421] Step 2: If the user expresses negative emotions, the emotion engine feeds that information back to the generative AI.

[0422] Step 3: The generative AI uses the information to update stress relief strategies.

[0423] "Example 3"

[0424] Step 1: The emotion engine analyzes the user's emotional state and feeds the results back to the generative AI.

[0425] Step 2: The generative AI combines this information with the dream content inferred from sleep talking to more accurately understand the user's stress.

[0426] Step 3: Generative AI proposes a solution.

[0427] Example 1

[0428] Next, a description will be given of Example 1 of Form 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."

[0429] In modern society, many people suffer from stress, which has a negative impact on the quality of sleep and mental health. However, there is no system that can analyze sleep states in detail and propose stress relief methods suited to individual users. In particular, there is no established method yet for analyzing sleep talking during sleep to infer the content of dreams and then using that information to discover and relieve stress. Therefore, there is a need for a system that can analyze sleep talking during sleep and effectively relieve users' stress.

[0430] 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.

[0431] In this invention, the server includes means for collecting sleep talking during sleep, means for transmitting the collected sleep talking to the server, means for the server to transfer the sleep talking data to the generative AI, means for the generative AI to infer the content of dreams from the sleep talking using deep learning, means for transmitting the inferred dream content to the advice generation module, means for the advice generation module to discover the individual's latent stress from the inferred dream content, means for suggesting ways to relieve the discovered stress, and means for an emotion engine to analyze the user's emotions and feed the results back to the generative AI. This makes it possible to analyze the user's sleep talking during sleep, discover stress from the content of dreams, and suggest ways to relieve it.

[0432] The "means for collecting sleep talking during sleep" refers to a device or function for collecting with high sensitivity the sounds made by the user while sleeping.

[0433] The "means for transmitting collected sleep talking to a server" refers to a device or function for transferring collected voice data to a server using the Internet or other communication means.

[0434] "Means by which the server transfers sleep talking data to the generative AI" refers to devices or functions that allow the server to pass on the voice data it receives to the generative AI.

[0435] "Means for generative AI to infer dream content from sleep talking using deep learning" refers to devices or functions that allow generative AI to analyze audio data and use deep learning algorithms to infer the content of the dream the user is having.

[0436] "Means for sending the inferred dream content to the advice generation module" refers to a device or function for transferring the dream content inferred by the generative AI to the advice generation module.

[0437] "Means for discovering the potential stress an individual is experiencing from the dream content inferred by the advice generation module" refers to a device or function that analyzes the dream content received by the advice generation module from the generative AI and identifies the potential stress the user is experiencing.

[0438] The "means for proposing a method for relieving the discovered stress" refers to a device or function that allows the advice generation module to suggest to the user an appropriate method for relieving the identified stress.

[0439] "Means for the emotion engine to analyze the user's emotions and provide the results as feedback to the generative AI" refers to devices or functions that allow the emotion engine to analyze emotions from the user's voice, facial expressions, actions, etc., and provide the analysis results to the generative AI.

[0440] MODE FOR CARRYING OUT THE INVENTION

[0441] This invention is a system that analyzes the sleep talking of a user while they sleep, identifies stress from the content of their dreams, and suggests ways to relieve it. This system is composed of the following main components:

[0442] 1. Recording sleep talking using a microphone

[0443] A highly sensitive microphone connected to the device picks up the sleep talking uttered by the user while sleeping. This microphone has a noise-canceling function, which eliminates surrounding noise and clearly records the sleep talking. For example, if the user says "help me" in their sleep, the audio data will be collected by the microphone.

[0444] 2. Sending collected data to the server

[0445] The device sends the collected sleep talking audio data to a server. This transmission is done in real time, and the data is encrypted and transferred securely. For example, when the user's sleep talking data is sent from the device to the server, the SSL / TLS protocol is used to ensure data security.

[0446] 3. The server transfers data to the generative AI

[0447] The server then forwards the received sleep-talking audio data to a generative AI. This generative AI is built using deep learning frameworks such as TensorFlow and PyTorch. For example, the server sends the sleep-talking data to the generative AI's API endpoint.

[0448] 4. Generative AI for dream content prediction

[0449] The generative AI analyzes the received sleep talking audio data and infers the content of the dream the user is having. This analysis uses voice recognition technology and natural language processing technology. For example, if the sleep talking contains the phrase "I'm being chased," the generative AI infers that the user is having a stressful dream.

[0450] 5. Sending the inference results to the advice generation module

[0451] The content of the dream inferred by the generative AI is sent to the advice generation module via the server. The server converts this data into an appropriate format and passes it to the advice generation module. For example, if the generative AI infers that "the user is dreaming of being chased," the server sends the result to the advice generation module.

[0452] 6. Proposing stress relief methods using the advice generation module

[0453] The advice generation module detects potential stress the user may be experiencing from the estimated dream content and suggests ways to alleviate it. The suggestions are based on the user's past data and common stress relief methods. For example, specific advice such as "Try yoga to relax" is provided.

[0454] 7. Emotional analysis and feedback of users using an emotion engine

[0455] The emotion engine analyzes the user's emotions from their voice, facial expressions, and behavior. The results of this analysis are fed back to the generative AI. The generative AI uses this information to more accurately predict the content of the user's dreams and suggest stress relief methods that are tailored to each individual user. For example, if a user feels stressed during the day, that information is fed back from the emotion engine to the generative AI, allowing it to provide more appropriate advice.

[0456] Examples and prompts

[0457] Specific examples

[0458] If a user says in their sleep, "Help me, someone is chasing me," a microphone connected to the device will collect this sleep talking. The collected data is sent from the device to a server, which then forwards it to the generative AI. From this sleep talking, the generative AI infers that the user is dreaming of being chased. The inference result is sent via the server to an advice generation module, which then provides advice such as, "Try yoga to relax."

[0459] Prompt Sentence Examples

[0460] "Collect the sleep talking that the user utters while sleeping, analyze that data, and infer the content of their dreams. Discover the potential stress that the user is experiencing from the inferred dream content, and generate advice suggesting ways to relieve that stress."

[0461] This system supports the user's mental health by analyzing the user's sleep talking, identifying stress from the content of their dreams, and suggesting ways to relieve it.

[0462] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0463] Step 1:

[0464] A highly sensitive microphone connected to the device collects the sleep talking uttered by the user while sleeping. The microphone is equipped with a noise-canceling function, eliminating surrounding noise and recording the sleep talking clearly. The input is the user's sleep talking audio, and the output is the collected audio data. For example, if the user says "help me" in their sleep, that audio data will be collected by the microphone.

[0465] Step 2:

[0466] The device sends the collected sleep talking audio data to a server. This transmission is done in real time, and the data is encrypted and transferred securely. The input is the collected audio data, and the output is the audio data sent to the server. For example, when the user's sleep talking data is transmitted from the device to the server, the SSL / TLS protocol is used to ensure data security.

[0467] Step 3:

[0468] The server forwards the received sleep talking audio data to a generative AI. This generative AI is built using deep learning frameworks such as TensorFlow and PyTorch. The input is the audio data sent to the server, and the output is the audio data forwarded to the generative AI. For example, the server sends the sleep talking data to the API endpoint of the generative AI.

[0469] Step 4:

[0470] The generative AI analyzes the received sleep talking audio data and infers the content of the dream the user is having. This analysis uses voice recognition technology and natural language processing technology. The input is the audio data transferred to the generative AI, and the output is the inferred dream content. For example, if the sleep talking includes the phrase "I'm being chased," the generative AI will infer that the user is having a stressful dream.

[0471] Step 5:

[0472] The dream content inferred by the generative AI is sent to the advice generation module via the server. The server converts this data into an appropriate format and passes it to the advice generation module. The input is the inferred dream content, and the output is the data sent to the advice generation module. For example, the generative AI infers that "the user is dreaming of being chased," and the server sends this result to the advice generation module.

[0473] Step 6:

[0474] The advice generation module detects potential stress the user may be experiencing from the estimated dream content and suggests ways to relieve it. These suggestions are based on the user's past data and common stress relief methods. The input is the data sent to the advice generation module, and the output is the suggested stress relief method. For example, specific advice such as "Try yoga to relax" is provided.

[0475] Step 7:

[0476] The emotion engine analyzes emotions from the user's voice, facial expressions, behavior, etc. The results of this analysis are fed back to the generative AI. The generative AI uses this information to more accurately predict the content of the user's dreams and propose stress relief methods that are suited to each individual user. The input is the user's emotional data, and the output is the analysis results fed back to the generative AI. For example, if a user feels stressed during the day, that information is fed back from the emotion engine to the generative AI, which then provides more appropriate advice.

[0477] (Application example 1)

[0478] Next, a description will be given of Application Example 1 of Embodiment 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."

[0479] In modern society, many people are suffering from latent stress, which has a negative impact on health. However, it is difficult to identify the causes of stress and propose appropriate ways to relieve it. Furthermore, there is no system that can analyze a user's emotional state in real time and adjust stress relief methods based on that analysis. This makes it difficult to effectively manage users' mental health.

[0480] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means. In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for discovering potential stress that an individual has from the inferred dream content, means for suggesting methods for relieving the discovered stress, means for analyzing the user's emotional state, and means for adjusting the stress relief method based on the analyzed emotional state. This makes it possible to identify the user's potential stress and suggest an appropriate stress relief method according to the individual's emotional state.

[0481] The "means for collecting sleep talking during sleep" refers to a device or system for recording sounds made by a user while sleeping.

[0482] "Generative AI" is an artificial intelligence technology that analyzes collected data and generates specific information.

[0483] "Means for inferring the content of dreams" refers to technology for analyzing and inferring the content of a user's dreams based on collected sleep talking data.

[0484] The "means for detecting potential stress" is a technology for identifying the stress a user may be experiencing from the content of their inferred dreams.

[0485] The "means for suggesting a method for relieving stress" is a technique for suggesting to the user an appropriate method for relieving the discovered stress.

[0486] "Means for analyzing the user's emotional state" refers to technology for analyzing emotions from the user's voice, facial expressions, behavior, etc.

[0487] The "means for adjusting a stress relief method based on emotional state" is a technique for appropriately adjusting the proposed stress relief method in accordance with the analyzed emotional state.

[0488] The following describes an embodiment of the present invention. First, the system is composed of the following main components.

[0489] 1. How to collect sleep talking during sleep:

[0490] The sleep talking that the user makes while sleeping is collected using the smartphone microphone, and the collected audio data is stored in local storage.

[0491] 2. Generative AI:

[0492] The collected sleep talking data is analyzed using deep learning frameworks such as TensorFlow. The generative AI includes a model for inferring dream content from sleep talking.

[0493] 3. How to guess the content of your dream:

[0494] The generative AI analyzes the sleep talking data and infers the content of the dream the user is having. The inferred dream content is used in the next step.

[0495] 4. Ways to spot underlying stress:

[0496] The content of the dream is inferred, and potential stress that the user may be experiencing is identified. The content of the dream is analyzed using a natural language processing library (such as NLTK) to identify the cause of stress.

[0497] 5. Suggested ways to relieve stress:

[0498] The system suggests appropriate ways to relieve the stress detected to the user. The suggestions are sent to the user using a notification service (such as Twilio).

[0499] 6. Means of analyzing the user's emotional state:

[0500] The system uses the smartphone camera to analyze the user's facial expressions in real time. It uses image processing libraries such as OpenCV to collect facial expression data and analyzes it with an emotion engine.

[0501] 7. How to adjust your stress management strategies based on your emotional state:

[0502] Based on the analyzed emotional state, the proposed stress relief method is adjusted appropriately, providing the most appropriate relief method for the user's current emotional state.

[0503] Specific examples

[0504] For example, if a user says in their sleep, "Work isn't going well," the generative AI will infer "work-related stress." If the emotion engine detects "anxiety" from the user's facial expression, the advice generation module will suggest "breathing techniques to relax."

[0505] Prompt Sentence Examples

[0506] Analyze the user's sleep talking data and infer the content of their dreams. Next, identify potential stress from the inferred dream content and suggest ways to relieve it. Take the user's emotional state into consideration.

[0507] In this way, a system can be realized that utilizes data from users while they are sleeping, identifies their individual stress levels, and suggests appropriate ways to relieve them.

[0508] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0509] Step 1:

[0510] The user's device (smartphone) uses a microphone to collect sleep talking while sleeping. The collected audio data is stored in local storage. The input is the user's sleep talking audio, and the output is the stored audio data.

[0511] Step 2:

[0512] The device sends the saved sleep-talking audio data to the server. The server inputs the received audio data into the generative AI. The input is the sleep-talking audio data, and the output is the input data for the generative AI.

[0513] Step 3:

[0514] The server's generative AI uses TensorFlow to analyze sleep-talking audio data and infer the content of dreams. The input is the sleep-talking audio data, and the output is the inferred dream content. Specifically, the audio data is converted into text and analyzed using a deep learning model.

[0515] Step 4:

[0516] The server detects potential stress based on the inferred dream content. It uses a natural language processing library (such as NLTK) to analyze the dream content and identify the causes of stress. The input is the inferred dream content, and the output is the identified causes of stress.

[0517] Step 5:

[0518] The server proposes appropriate solutions to the detected stress. The proposals are sent to the user using a notification service (such as Twilio). The input is the identified cause of stress, and the output is the proposed stress relief solution. The specific operation is to generate a stress relief solution and create a notification message.

[0519] Step 6:

[0520] The user's device uses a camera to analyze the user's facial expressions in real time. Using image processing libraries such as OpenCV, the facial expression data is collected and analyzed by the emotion engine. The input is the user's facial expression data, and the output is the analyzed emotional state.

[0521] Step 7:

[0522] The server appropriately adjusts the proposed stress relief method based on the analyzed emotional state. The input is the analyzed emotional state and the proposed stress relief method, and the output is the adjusted stress relief method. Specifically, the server optimizes the stress relief method according to the emotional state.

[0523] In this way, a system can be realized that utilizes data from users while they are sleeping, identifies their individual stress levels, and suggests appropriate ways to relieve them.

[0524] Example 2

[0525] Next, a description will be given of Example 2 of Form 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."

[0526] Conventional systems have difficulty accurately predicting the content of a user's dreams while they are sleeping, and are unable to monitor the user's emotional state in real time or suggest appropriate stress relief methods, making it difficult to effectively support the user's mental health.

[0527] 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.

[0528] In this invention, the server includes means for collecting voice data during sleep, means for converting the collected voice data into text data, means for inputting the converted text data into a deep learning model to infer the content of dreams, means for detecting potential stress the individual is experiencing from the inferred dream content, means for suggesting methods for relieving the discovered stress, means for monitoring the user's emotional state in real time, and means for updating the stress relief methods in response to changes in the emotional state. This makes it possible to accurately infer the content of the user's dreams, monitor the emotional state in real time, and suggest appropriate stress relief methods.

[0529] "Sound data during sleep" is data that records sounds made by the user while sleeping.

[0530] "Means for converting into text data" refers to technology or devices for converting voice data into text information.

[0531] A "deep learning model" is a type of artificial intelligence that is trained using large amounts of data to learn complex patterns and relationships.

[0532] A "means for predicting the content of dreams" is a technology or device for predicting the content of a user's dream from input data.

[0533] "Means for detecting latent stress" are technologies and devices for detecting stress that the user is not aware of.

[0534] "Means for suggesting ways to relieve stress" refers to techniques or devices that suggest specific ways to reduce stress to the user.

[0535] "Means for monitoring emotional state in real time" refers to technology or devices for observing and recording changes in a user's emotions in real time.

[0536] The "means for updating stress relief methods in response to changes in the emotional state" refers to a technique or device for appropriately changing the stress relief methods suggested based on changes in the user's emotions.

[0537] This invention is a system that collects voice data while a user is sleeping, analyzes the data to infer the content of dreams, and monitors the user's emotional state in real time to suggest appropriate stress relief methods.

[0538] First, the device records the user's speech while sleeping through a microphone. This recorded speech data is temporarily stored on the device. Next, the device uses speech recognition software such as the Google Speech-to-Text API to convert the recorded speech data into text data. This converted text data is then sent to a server.

[0539] The server inputs the received text data into a deep learning model. This deep learning model is built using frameworks such as TensorFlow and PyTorch and is trained using large amounts of audio data and corresponding dream content data. The generative AI model analyzes the input text data and infers the content of the user's dream. The inferred results are stored on the server.

[0540] The server then sends the inference results to the device, which then provides feedback to the user. For example, if a user says in their sleep, "I'm swimming in the ocean," the generative AI model infers that "the user is dreaming about swimming in the ocean," and the device notifies the user of the result.

[0541] Examples:

[0542] User sleep talking data: "Swimming in the ocean"

[0543] Generative AI model predicts: "The user is dreaming about swimming in the ocean."

[0544] Additionally, the device monitors the user's facial expressions and vocal tone in real time using emotion recognition software such as Affectiva or the Microsoft® Azure® Emotion API. If the user displays negative emotions such as anger or sadness, the device transmits this information to the server.

[0545] The server inputs the received emotional data into the emotion engine. The emotion engine detects changes in the user's emotional state and feeds that information back to the generative AI model. The generative AI model updates stress relief methods according to the user's emotional state. For example, if the user is feeling angry, the generative AI model will suggest stress relief methods such as relaxation or meditation.

[0546] Examples:

[0547] User sentiment data: "I'm really frustrated today"

[0548] The emotion engine determines: "The user is angry."

[0549] Generative AI model suggestion: "Try relaxation or meditation."

[0550] In this way, the system can accurately predict the content of a user's dreams, monitor their emotional state in real time, and suggest appropriate stress relief methods.

[0551] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0552] Step 1:

[0553] The device records the sounds made by the user while sleeping through a microphone. The recorded sound data is temporarily stored in the device. The input is the user's voice, and the output is the recorded sound data. Specifically, the device's microphone captures the sound and saves it as a recording file.

[0554] Step 2:

[0555] The device converts the recorded audio data into text data using the Google Speech-to-Text API. The input is the recorded audio data, and the output is the converted text data. Specifically, the device sends the audio file to the API and receives the returned text data.

[0556] Step 3:

[0557] The terminal transmits the converted text data to the server. The input is the text data, and the output is the transmission of the data to the server. Specifically, the terminal uploads the text data to the server via the network.

[0558] Step 4:

[0559] The server inputs the received text data into the deep learning model. The input is text data, and the output is the predicted dream content. Specifically, the server passes the text data to the deep learning model, which then analyzes it.

[0560] Step 5:

[0561] The generative AI model analyzes input text data and infers the content of dreams. The input is text data, and the output is the inferred dream content. Specifically, the model analyzes text data based on the data it has been trained on and infers the content of dreams.

[0562] Step 6:

[0563] The server sends the guess results to the terminal. The input is the guess result of the dream content, and the output is the transmission of data to the terminal. In concrete terms, the server sends the guess results to the terminal via the network.

[0564] Step 7:

[0565] The device then feeds back the results of its guesses to the user. The input is the guessed content of the dream, and the output is a notification to the user. Specifically, the device communicates the results to the user through a screen display or audio notification.

[0566] Step 8:

[0567] The device monitors the user's facial expressions and vocal tone in real time. The input is the user's facial expressions and vocal tone, and the output is emotional data. Specifically, the device collects data using a camera and microphone and analyzes it with emotion recognition software.

[0568] Step 9:

[0569] The device sends the collected emotion data to the server. The input is emotion data, and the output is data transmission to the server. Specifically, the device uploads the emotion data to the server via the network.

[0570] Step 10:

[0571] The server inputs the received emotion data into the emotion engine. The input is emotion data, and the output is the analysis result of the emotional state. Specifically, the server passes the emotion data to the emotion engine for analysis.

[0572] Step 11:

[0573] The emotion engine detects changes in the user's emotional state and feeds that information back to the generative AI model. The input is emotional data, and the output is information about changes in emotional state. Specifically, the emotion engine analyzes the data and detects changes.

[0574] Step 12:

[0575] The generative AI model receives feedback from the emotion engine and updates stress relief methods according to the user's emotional state. The input is information on changes in emotional state, and the output is updated stress relief methods. Specifically, the model recalculates stress relief methods based on the feedback.

[0576] Step 13:

[0577] The server sends the updated stress relief method to the terminal. The input is the updated stress relief method, and the output is data transmission to the terminal. In specific operations, the server sends the updated information to the terminal through the network.

[0578] Step 14:

[0579] The device proposes updated stress relief methods to the user. The input is the updated stress relief methods, and the output is the proposal to the user. As a specific operation, the device notifies the user of the proposal through a screen display or a voice notification.

[0580] (Application example 2)

[0581] Next, a description will be given of Application Example 2 of Form 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."

[0582] In modern society, stress and mental strain are increasing, causing many people to suffer from mental health problems. In particular, there is a need for a system that can effectively detect potential stress and emotional changes that occur during sleep and suggest appropriate stress relief methods. However, current systems lack the functionality to monitor the user's emotional state in real time and dynamically update stress relief methods, preventing effective mental health care.

[0583] 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.

[0584] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for discovering potential stress that an individual has from the inferred dream content, means for suggesting ways to relieve the discovered stress, means for monitoring the user's emotional state in real time, and means for updating the stress relief methods in response to changes in the emotional state. This makes it possible to grasp the user's emotional state in real time and dynamically suggest appropriate stress relief methods.

[0585] The "means for collecting sleep talking" refers to a device or software for recording sounds made by a user while they are sleeping.

[0586] "Generative AI" is an artificial intelligence technology that generates new information based on collected data, and in particular uses deep learning.

[0587] The "means for inferring the content of dreams" refers to algorithms or models that analyze collected sleep talking data and infer the content of the dreams the user is having.

[0588] The "means for discovering potential stress in an individual" is a method or system for identifying a user's potential stress factors from the content of their inferred dreams.

[0589] The "means for suggesting ways to relieve stress" is a function for providing the user with an appropriate way to relieve stress based on the discovered stress factors.

[0590] "Means for monitoring the user's emotional state in real time" refers to devices or software that monitor the user's biometric information, such as heart rate and facial expressions, in real time and analyze the user's emotional state.

[0591] The "means for updating stress relief methods according to changes in emotional state" is a function for detecting changes in the user's emotional state and dynamically changing the stress relief methods suggested in response to those changes.

[0592] The following system configuration will be described as an embodiment of the present invention.

[0593] System Configuration

[0594] Hardware

[0595] Smartphone: Collects sleep talking data and monitors the user's emotional state.

[0596] Smart glasses: Monitor the user's emotional state and suggest ways to relieve stress.

[0597] Server: Analyzes collected data and runs generative AI models.

[0598] software

[0599] TensorFlow: For building and inferencing deep learning models.

[0600] SoundDevice: Collects audio data.

[0601] Librosa: Preprocessing audio data.

[0602] Processing flow

[0603] 1. Collection of sleep talking data

[0604] The sleep talking uttered by the user while sleeping is collected using the smartphone microphone.

[0605] The collected voice data is sent to a server.

[0606] 2. Guessing the content of the dream

[0607] The server preprocesses the collected sleep talking data using Librosa and extracts MFCC features.

[0608] The extracted features are input into a TensorFlow deep learning model to infer the content of the dream.

[0609] 3. Identifying underlying stress

[0610] The content of the dream is inferred to identify potential stressors the user may be experiencing.

[0611] The identified stress factors are stored on a server.

[0612] 4. Suggestions for ways to relieve stress

[0613] Based on the identified stress factors, the server uses a generative AI model to suggest appropriate stress relief methods.

[0614] The proposed stress relief methods are communicated to the user via their smartphone or smart glasses.

[0615] 5. Monitoring your emotional state

[0616] Using sensors on smartphones and smart glasses, the system monitors the user's heart rate and facial expressions in real time.

[0617] The collected biometric information is transmitted to a server.

[0618] 6. Update your stress relief methods

[0619] The server analyzes changes in the user's emotional state and dynamically updates stress relief methods as needed.

[0620] The updated stress relief method is again notified to the user.

[0621] Specific examples

[0622] For example, a smartphone can collect sleep talking while a user sleeps and notify the user the next morning, saying, "Last night's dream was about adventure." If a user feels stressed at work, smart glasses can suggest, "Try relaxation or meditation."

[0623] Prompt Sentence Examples

[0624] "Guess what I dreamed about last night."

[0625] "Please suggest ways for users to relieve stress if they are feeling angry."

[0626] In this way, the embodiments for carrying out the invention have been specifically shown.

[0627] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0628] Step 1:

[0629] Collection of sleep talking data

[0630] Input: Audio made by the user while sleeping

[0631] How it works: The app uses the smartphone's microphone to record the user's sleep talking in real time. The recorded audio data is temporarily stored on the smartphone.

[0632] Output: Recorded sleep talking data

[0633] Step 2:

[0634] Sending sleep talking data

[0635] Input: Recorded sleep talking

[0636] How it works: The smartphone sends the collected sleep talking data to a server using an internet connection.

[0637] Output: Sleep talking data sent to the server

[0638] Step 3:

[0639] Guessing the content of the dream

[0640] Input: Sleep talking data sent to the server

[0641] How it works: The server preprocesses the sleep talking data using Librosa and extracts MFCC features. The extracted features are then input into a TensorFlow deep learning model to infer the content of the dream.

[0642] Output: Inferred dream content

[0643] Step 4:

[0644] Identifying underlying stress

[0645] Input: Inferred dream content

[0646] How it works: The server analyzes the content of the predicted dream and identifies potential stressors that the user may be experiencing. The identified stressors are then stored in a database.

[0647] Output: Identified potential stressors

[0648] Step 5:

[0649] Suggestions for stress relief

[0650] Input: Identified potential stressors

[0651] How it works: The server uses a generative AI model to generate appropriate stress relief methods based on the identified stress factors. The generated stress relief methods are then sent to a smartphone or smart glasses.

[0652] Output: Suggested stress relief methods

[0653] Step 6:

[0654] Emotional state monitoring

[0655] Input: User's biometric information such as heart rate and facial expression

[0656] How it works: Using sensors in smartphones and smart glasses, the system monitors the user's heart rate and facial expressions in real time, and transmits the collected biometric information to a server.

[0657] Output: Biometric information sent to the server

[0658] Step 7:

[0659] Updates on stress relief methods

[0660] Input: Biometric information sent to the server

[0661] How it works: The server analyzes changes in the user's emotional state and dynamically updates the stress relief methods as needed. The updated stress relief methods are then sent to the smartphone or smart glasses.

[0662] Output: Updated stress relief methods

[0663] In this way, the specific operations performed at each processing step and their inputs and outputs are clearly indicated.

[0664] Example 3

[0665] Next, a description will be given of a third embodiment of the third embodiment. 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."

[0666] Conventional mental health treatment systems often rely on subjective reports from patients, making it difficult to accurately identify stress and suggest appropriate ways to relieve it. Furthermore, if patients are not aware of their own stress, appropriate treatment may be delayed. This has led to the problem of insufficient improvement in the mental health of psychiatric patients.

[0667] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0668] In this invention, the server includes means for collecting voice data during sleep, means for converting the collected voice data into text data, means for analyzing the converted text data to infer the content of dreams, means for detecting potential stress in the individual from the inferred dream content, means for suggesting ways to relieve the discovered stress, and means for notifying the individual of the suggested stress relief methods. This makes it possible to analyze the content of dreams from the patient's sleep talking, identify potential stress, and suggest appropriate stress relief methods.

[0669] "Sound data during sleep" refers to sounds made by an individual while sleeping, and is data that includes sounds made unconsciously, such as talking in one's sleep.

[0670] "Means of collection" refers to the devices and software used to capture and store audio data.

[0671] "Means for converting into text data" refers to speech recognition technology for converting voice data into text information.

[0672] "Means of analyzing and inferring the content of dreams" refers to natural language processing technology that analyzes converted text data and infers the content of dreams from it.

[0673] "Means for detecting potential stress" refers to algorithms and analytical techniques for identifying potential stress an individual may be experiencing from the content of their inferred dreams.

[0674] "Means for suggesting stress relief methods" refers to a generative AI model that generates and suggests appropriate stress relief methods for identified stress.

[0675] "Means for notifying" refers to a communication means or display device for informing an individual of the proposed stress relief method.

[0676] This invention is a system that can be used in mental health treatment for psychiatric patients, and identifies the patient's stress by collecting and analyzing voice data during sleep and proposing appropriate stress relief methods.

[0677] Hardware and software used

[0678] server

[0679] The server uses the following hardware and software:

[0680] Speech Recognition Software: We use speech recognition technology to convert voice data into text data, specifically, we use Google Cloud Speech-to-Text.

[0681] Natural language processing engine: We use natural language processing technology, specifically SpaCy, to analyze the converted text data and infer the content of dreams.

[0682] Stress analysis algorithm: To identify stress from the inferred dream content, stress analysis technology is used. Specifically, IBM Watson (registered trademark) is used.

[0683] Generative AI model: We use a generative AI model, specifically OpenAI GPT-4, to generate appropriate stress relief methods based on the identified stress.

[0684] Terminal

[0685] The terminal uses the following hardware and software:

[0686] Microphone: Used to record the patient's sleep talking.

[0687] Communication module: Used to send recorded audio data to the server.

[0688] Display device: Used to display the stress relief methods received from the server to the patient.

[0689] User

[0690] The user performs the following operations:

[0691] Record your sleep talking: Record your sleep talking through your device.

[0692] Check and execute stress relief methods: Check and execute the stress relief methods sent from the server.

[0693] Specific examples

[0694] Audio data collection and analysis

[0695] 1. The device uses a microphone to record the patient's sleep talking. For example, if a patient says at night, "Today I had a dream that I was being swept away by a big wave," the device will record that voice.

[0696] 2. The device sends the recorded audio data to the server in real time.

[0697] 3. The server converts the voice data into text data using Google Cloud Speech-to-Text. In this case, the voice data is converted into text data such as "Today I had a dream that I was swept away by a big wave."

[0698] 4. The server uses SpaCy to process the converted text data using natural language processing. For example, the phrase "swallowed by a big wave" is extracted.

[0699] 5. The server uses an NLP engine to infer the content of the dream. In this case, the content of the dream is inferred to be "being swept away by a big wave."

[0700] Identifying stress and suggesting ways to relieve it

[0701] 1. The server inputs the predicted dream content into IBM Watson's stress analysis algorithm. For example, a dream about being swept away by a big wave is input.

[0702] 2. The server uses the analysis results to identify the stress the patient is experiencing. In this case, it identifies "stress caused by anxiety or fear."

[0703] 3. The server generates stress relief methods using OpenAI GPT-4 based on the identified stress. For example, it generates "deep breathing techniques to relax."

[0704] 4. The server sends the generated stress relief method to the terminal.

[0705] 5. The terminal displays the stress relief method received from the server to the patient. For example, "deep breathing techniques to relax" is displayed.

[0706] 6. The user checks the displayed stress relief method and performs it, for example, by following the deep breathing procedure to relax.

[0707] Prompt Sentence Examples

[0708] "Analyze the patient's sleep talking, identify the stress from the content of the dream, and suggest an appropriate method for relieving stress. The content of the sleep talking was, 'Today I had a dream that I was swept away by a big wave.' The user's emotional state is 'anxious.'" The flow of the identification process in Example 3 will be described with reference to Figure 21.

[0709] Step 1:

[0710] Audio data collection

[0711] The device uses a microphone to record audio when the patient talks in their sleep.

[0712] Input: Patient's sleep talking (audio data)

[0713] Output: Recorded audio data

[0714] Specific operation: If the patient says at night, "Today I had a dream that I was being swept away by a big wave," the device will record their voice.

[0715] Step 2:

[0716] Sending audio data

[0717] The terminal transmits the recorded voice data to the server in real time.

[0718] Input: Recorded audio data

[0719] Output: Audio data sent to the server

[0720] Specific operation: The device sends the recorded audio data to a server via the Internet.

[0721] Step 3:

[0722] Converting audio data to text

[0723] The server converts the audio data into text data using Google Cloud Speech-to-Text.

[0724] Input: Audio data sent to the server

[0725] Output: Converted text data

[0726] Specific operation: The server converts the audio data "Today I had a dream that I was swept away by a big wave" into text data.

[0727] Step 4:

[0728] Text data analysis

[0729] The server performs natural language processing on the converted text data using SpaCy.

[0730] Input: Converted text data

[0731] Output: Parsed dream content

[0732] Specific operation: The server analyzes the text data "Today I had a dream that I was swallowed by a big wave" and extracts the content of the dream, "being swallowed by a big wave."

[0733] Step 5:

[0734] Identifying stress

[0735] The server inputs the estimated dream content into IBM Watson's stress analysis algorithm.

[0736] Input: Parsed dream content

[0737] Output: Identified stress

[0738] Specific operation: The server inputs the content of the dream about being "swept away by a big wave" into a stress analysis algorithm and identifies "stress caused by anxiety and fear."

[0739] Step 6:

[0740] Creating ways to relieve stress

[0741] Based on the identified stress, the server generates a stress relief method using OpenAI GPT-4.

[0742] Input: Identified stress

[0743] Output: Generated stress relief methods

[0744] Specific operation: The server generates "deep breathing techniques to relax" in response to "stress caused by anxiety or fear."

[0745] Step 7:

[0746] Stress relief notifications

[0747] The server transmits the generated stress relief method to the terminal.

[0748] Input: Generated stress relief methods

[0749] Output: Stress relief method sent to device

[0750] Specific operation: The server sends the generated "deep breathing method for relaxation" to the device.

[0751] Step 8:

[0752] Displaying ways to relieve stress

[0753] The terminal displays the stress relief method received from the server to the patient.

[0754] Input: Stress relief method sent from the server

[0755] Output: Stress relief strategies displayed to the patient

[0756] Specific operation: The device displays the received "deep breathing techniques for relaxation" to the patient.

[0757] Step 9:

[0758] Implementing stress relief methods

[0759] The user checks the displayed stress relief method and implements it.

[0760] Input: Stress relief methods displayed on the device

[0761] Output: Stress relief method performed

[0762] Specific action: The user follows the displayed deep breathing instructions to relax.

[0763] (Application example 3)

[0764] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[0765] In conventional mental health treatment, it is difficult to accurately grasp a patient's emotional state and the causes of stress, making it difficult to propose effective stress relief methods. Furthermore, technology for analyzing sleep talking and dream content has not yet fully developed, limiting the means to discover a patient's underlying stress. Furthermore, there has been no system in brick-and-mortar counseling or mental health clinics that can comprehensively analyze a patient's emotional state and dream content and propose appropriate stress relief methods.

[0766] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0767] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for discovering potential stress a person has from the inferred dream content, means for suggesting ways to relieve the discovered stress, means for analyzing the person's emotional state using an emotion engine, means for suggesting ways to relieve stress by combining the analyzed emotional state and dream content, means for recording and analyzing sleep talking using a smart device, and means for use in brick-and-mortar counseling rooms and mental health clinics. This makes it possible to comprehensively analyze a patient's emotional state and dream content and suggest more accurate ways to relieve stress.

[0768] A "means for collecting sleep talking during sleep" is a device or method for recording an individual's sleep talking during sleep and collecting it as data.

[0769] "Generative AI" is an artificial intelligence technology that generates specific information based on collected data, and in particular uses deep learning to infer the content of dreams.

[0770] A "means for inferring the content of dreams" is a method or device for analyzing collected sleep talking data and inferring the specific content of the dreams an individual had from that content.

[0771] A "means for detecting potential stress" is a method or device for identifying potential stress an individual may be experiencing from the content of inferred dreams.

[0772] "Means for suggesting ways to relieve stress" refers to a method or device for suggesting appropriate ways to relieve stress that has been discovered.

[0773] The "emotion engine" is a technology for analyzing an individual's emotional state, and in particular, for quantifying and classifying emotional changes and states.

[0774] The "means for analyzing emotional state" refers to a method or device for analyzing an individual's emotional state using an emotion engine and obtaining the results.

[0775] A "smart device" is a portable electronic device that can connect to the Internet and is used specifically to record and analyze sleep talking.

[0776] "Brick and mortar store" means a store that exists in a physical location, and specifically includes counseling rooms and mental health clinics.

[0777] The embodiments of the present invention will be described in detail below.

[0778] System Overview

[0779] This system collects sleep talking during sleep, uses generative AI to infer the content of dreams from the collected sleep talking, discovers potential stress in individuals from the inferred dream content, and suggests ways to relieve the discovered stress.Furthermore, it uses an emotion engine to analyze an individual's emotional state, and proposes stress relief methods by combining the analyzed emotional state with the content of dreams.

[0780] Hardware and software used

[0781] Smart devices: Use portable electronic devices that can connect to the internet, such as smartphones and head-mounted displays, to record and analyze sleep talking.

[0782] Generative AI: Using artificial intelligence techniques that use deep learning to infer dream content.

[0783] Emotion Engine: Uses technology to analyze an individual's emotional state.

[0784] Server: Uses a central processing unit to process the collected data and suggest ways to relieve stress.

[0785] Data processing and calculation

[0786] 1. Sleep-talking collection: The user's sleep-talking is recorded using the microphone on the smart device. The recorded audio data is sent to the server.

[0787] 2. Dream content inference: The server inputs the collected sleep talking data into a generative AI to infer the dream content. The generative AI uses a deep learning model to infer the dream content from the sleep talking text data.

[0788] 3. Emotional state analysis: The server uses the emotion engine to analyze the user's emotional state. The emotion engine quantifies and classifies the emotional state based on the content of the sleep talking and other input data.

[0789] 4. Suggesting ways to relieve stress: The server combines the inferred dream content with the analyzed emotional state and uses generative AI to suggest ways to relieve stress.

[0790] Specific examples

[0791] For example, if a user says in their sleep, "I had a scary dream," the emotion engine will interpret this as "fear." Based on this information, the generative AI will suggest ways to relieve stress, such as "take deep breaths to relax" or "listen to relaxing music."

[0792] Prompt Sentence Examples

[0793] User's sleep talking: "I had a scary dream"

[0794] Emotional state: "Fear"

[0795] Prompt for generative AI model: "The user feels they had a scary dream. What stress relief would you suggest to the user who is feeling scared?"

[0796] In this way, a system can be realized that can be used in brick-and-mortar counseling rooms and mental health clinics.

[0797] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0798] Step 1:

[0799] The user uses a smart device (smartphone or head-mounted display) to record their sleep talking while they are sleeping.

[0800] Input: User's sleep talking (audio data)

[0801] Output: Audio data of recorded sleep talking

[0802] Specific operation: Using the microphone on the smart device, the user's sleep talking is recorded and saved as audio data.

[0803] Step 2:

[0804] The device sends the recorded audio data of the sleep talking to a server.

[0805] Input: Recorded audio data of sleep talking

[0806] Output: Audio data of sleep talking sent to the server

[0807] Specific operation: The smart device uploads the recorded audio data to a server via the Internet.

[0808] Step 3:

[0809] The server uses voice recognition technology to convert the sleep talking audio data into text data.

[0810] Input: Sleep talking audio data

[0811] Output: Text data of sleep talking

[0812] Specific operation: The server uses speech recognition software (e.g., Google's speech recognition API) to convert the voice data into text data.

[0813] Step 4:

[0814] The server uses generative AI to infer the content of dreams from the text data of sleep talking.

[0815] Input: Text data of sleep talking

[0816] Output: Inferred dream content

[0817] Specific operation: The server inputs the text data of sleep talking into a generative AI (deep learning model) and infers the content of the dream.

[0818] Step 5:

[0819] The server uses an emotion engine to analyze the user's emotional state.

[0820] Input: Text data of sleep talking

[0821] Output: Parsed emotional state

[0822] Specific operation: The server inputs the text data of sleep talking into the emotion engine, which then quantifies and classifies the emotional state.

[0823] Step 6:

[0824] The server combines the inferred dream content with the analyzed emotional state and uses generative AI to suggest ways to relieve stress.

[0825] Input: Inferred dream content, analyzed emotional state

[0826] Output: Suggested stress relief method

[0827] Specific operation: The server inputs the dream content and emotional state into the generative AI, which then generates an appropriate stress relief method.

[0828] Step 7:

[0829] The server sends the suggested stress relief methods to the user's smart device.

[0830] Input: Suggested stress relief method

[0831] Output: Stress relief methods displayed on the user's smart device

[0832] Specific operation: The server sends the suggested stress relief methods to the user's smart device via the Internet and displays them.

[0833] In this way, users can receive stress relief methods based on their sleep talking and emotional state.

[0834] 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.

[0835] 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> ) 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.

[0836] Another example of generative AI is Gemini (registered trademark) (Internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0837] 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.

[0838] [Second embodiment]

[0839] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0840] 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.

[0841] 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).

[0842] 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.

[0843] 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.

[0844] 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).

[0845] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0846] 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.

[0847] 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.

[0848] 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.

[0849] 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.

[0850] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0851] "Example 1"

[0852] This embodiment of the present invention is composed of a microphone that collects sleep talking during sleep, a generative AI that analyzes the collected sleep talking and infers the content of dreams, and an advice generation module that discovers potential stress an individual is experiencing from the inferred dream content and suggests ways to alleviate it. Specifically, the microphone collects sleep talking during sleep, and the data is sent to the generative AI. The generative AI uses deep learning to infer the content of dreams from the sleep talking and sends the results to the advice generation module. The advice generation module discovers stress from the inferred dream content and suggests ways to alleviate it.

[0853] "Example 2"

[0854] In another aspect of the present invention, a generative AI uses deep learning to infer dream content. Specifically, the generative AI receives sleep talking data as input and uses a deep learning model to infer dream content. The deep learning model is trained using a large amount of sleep talking data and corresponding dream content data.

[0855] "Example 3"

[0856] In a further aspect of the present invention, the system can be used in the mental health treatment of psychiatric patients. Specifically, the system infers the content of dreams from the sleep talking of psychiatric patients, and identifies the stress the patient is experiencing from the content of those dreams. The system then suggests ways to relieve that stress, contributing to the improvement of the patient's mental health.

[0857] The processing flow of each embodiment will be described below.

[0858] "Example 1"

[0859] Step 1: A microphone collects your sleep talking while you sleep.

[0860] Step 2: The collected sleep talking data is sent to the generative AI.

[0861] Step 3: The generative AI uses deep learning to infer the content of the dream from the sleep talking.

[0862] Step 4: The estimated dream content is sent to the advice generation module.

[0863] Step 5: The advice generation module detects stress from the inferred dream content and suggests ways to relieve it.

[0864] "Example 2"

[0865] Step 1: The generative AI receives sleep talking data as input.

[0866] Step 2: The generative AI uses a deep learning model to infer the content of the dream.

[0867] Step 3: The deep learning model is trained using a large amount of sleep talking data and corresponding dream content data.

[0868] "Example 3"

[0869] Step 1: Infer the content of a psychiatric patient's dreams from their sleep talking.

[0870] Step 2: Discover the stress the patient is experiencing from the content of the dream you have guessed.

[0871] Step 3: Suggest ways to relieve the stress discovered and contribute to improving the patient's mental health.

[0872] Example 1

[0873] Next, a description will be given of Example 1 of Form 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."

[0874] Conventional sleep talking analysis systems have insufficient collection and analysis of sleep talking, making it difficult to accurately identify potential stress in users and suggest ways to alleviate it. Furthermore, the accuracy of converting voice data into text and predicting dream content is low, resulting in a lack of reliability in the advice given to users. This has led to the issue of insufficient efforts to improve users' mental health.

[0875] 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.

[0876] In this invention, the server includes means for collecting sleep talking during sleep, means for saving the collected sleep talking as audio data, means for transmitting the saved audio data to the server, means for converting the audio data into text data, means for inferring the content of dreams from the converted text data using a generative AI model, means for discovering potential stress the individual is experiencing from the inferred dream content, and means for suggesting ways to relieve the discovered stress. This makes it possible to analyze the user's sleep talking with high precision, accurately discover potential stress, and reliably suggest ways to relieve it.

[0877] The "means for collecting sleep talking during sleep" refers to a device or mechanism for collecting sleep talking uttered by a user while sleeping with high sensitivity.

[0878] The "means for saving collected sleep talking as audio data" refers to a device or mechanism for recording and saving collected sleep talking as digital audio data.

[0879] The "means for transmitting the stored voice data to the server" refers to a device or mechanism for transmitting the stored voice data to the server via a network.

[0880] "Means for converting voice data into text data" refers to a device or mechanism for converting voice data into text information using voice recognition technology.

[0881] A "means for inferring dream content using a generative AI model" is a device or mechanism that uses a deep learning algorithm to analyze and infer dream content from text data.

[0882] A "means for discovering potential stress experienced by an individual" is a device or mechanism for identifying potential stress experienced by a user based on the content of the inferred dream.

[0883] The "means for proposing a method for relieving the discovered stress" is a device or mechanism for proposing to the user a method for relieving the identified stress.

[0884] MODE FOR CARRYING OUT THE INVENTION

[0885] This invention is a system that collects sleep talking during sleep, analyzes the data to infer the content of dreams, discovers potential stress that an individual is experiencing, and suggests ways to alleviate that stress. Specific embodiments of this system are described below.

[0886] Hardware and software used

[0887] 1. Microphone

[0888] To collect sleep talking, a highly sensitive microphone is used, which is designed to clearly record the user's sleep talking.

[0889] 2. Terminal

[0890] A device is used to store the collected voice data and send it to a server. This device has the function of storing voice data in real time and sending it to a server using a secure protocol (e.g., HTTPS).

[0891] 3. Server

[0892] The server converts the received audio data into text using speech recognition software (e.g., Google Cloud Speech-to-Text API), which is then passed to a generative AI model and used to infer the content of the dream.

[0893] 4. Generative AI Models

[0894] The generative AI model uses a deep learning algorithm to infer dream content from text data. Specifically, it analyzes the text data using natural language processing technology and generates dream scenarios.

[0895] 5. Advice Generation Module

[0896] The advice generation module detects potential stresses the user may be experiencing based on the content of the inferred dream and suggests ways to alleviate them, including feasible and effective methods for the user.

[0897] Specific examples

[0898] If a user says in their sleep, "My work isn't going well," the microphone connected to the device will collect this speech and save it as audio data. The device will then send the collected audio data to a server, where the HTTPS protocol is used to ensure data security.

[0899] The server converts the received voice data into text data using the Google Cloud Speech-to-Text API. For example, the generated text data is "My work is not going well." The server then passes the generated text data to the generative AI model. The generative AI model analyzes this text data and infers that the user is experiencing work-related stress.

[0900] The dream content predicted by the generative AI model is sent to an advice generation module via the server. The advice generation module determines that the user is experiencing work-related stress and generates advice for the user, such as "We suggest that you set aside time for relaxation every day to reduce work stress." This suggestion is sent to the device via the server and notified to the user.

[0901] Prompt Sentence Examples

[0902] "If a user says in their sleep, 'My work isn't going well,' the generative AI model will analyze this sleep talking and infer that the user is experiencing work-related stress. What advice will the advice generation module provide to the user?"

[0903] In this way, the system analyzes the user's sleep talking, discovers potential stress, and suggests ways to relieve it.

[0904] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0905] Step 1:

[0906] While the user is sleeping, a microphone connected to the device collects the user's sleep talking. The microphone is designed to be highly sensitive and to clearly record the user's sleep talking. The input is the user's sleep talking audio, and the output is the collected audio data.

[0907] Step 2:

[0908] The device saves the collected voice data. The saved voice data is temporarily stored in the device for further processing. The input is the collected voice data, and the output is the saved voice data.

[0909] Step 3:

[0910] The device sends the stored audio data to the server using a secure protocol (e.g., HTTPS) to ensure data security. The input is the stored audio data, and the output is the audio data sent to the server.

[0911] Step 4:

[0912] The server converts the received voice data into text data using speech recognition software (e.g., Google Cloud Speech-to-Text API). The input is the voice data, and the output is the converted text data. Specifically, the speech recognition algorithm analyzes the voice data and generates the corresponding text.

[0913] Step 5:

[0914] The server passes the converted text data to the generative AI model, which uses a deep learning algorithm to infer the dream content from the text data. The input is text data, and the output is the inferred dream content. Specifically, it uses natural language processing technology to analyze the text data and generate a dream scenario.

[0915] Step 6:

[0916] The dream content inferred by the generative AI model is sent via the server to the advice generation module, which then discovers the user's potential stress based on the inferred dream content. The input is the inferred dream content, and the output is the discovered stress.

[0917] Step 7:

[0918] The advice generation module proposes ways to relieve the detected stress. These proposals include feasible and effective methods for the user. The input is the detected stress, and the output is the proposed ways to relieve it. Specific actions include relaxation techniques and ways to improve lifestyle habits.

[0919] Step 8:

[0920] The proposed solution is sent to the device via the server and notified to the user. The input is the proposed solution, and the output is a notification to the user. Specifically, the device displays a notification so that the user can confirm the proposed solution.

[0921] (Application example 1)

[0922] Next, a description will be given of Application Example 1 of Form 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."

[0923] In modern society, individuals experience a wide range of potential stresses, which can affect not only their mental health but also security risks. However, while conventional systems can infer the content of dreams from sleep talking and detect potential stress, they are unable to suggest appropriate security measures in response to that stress. This creates a need for systems that can simultaneously support both the user's mental health and security.

[0924] 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.

[0925] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for detecting potential stress in an individual from the inferred dream content, means for suggesting ways to relieve the detected stress, and means for recommending appropriate security measures according to the stress level, thereby supporting the user's mental health and reducing security risks.

[0926] "Somnitalking" refers to speech that is involuntarily uttered by an individual while they are asleep.

[0927] "Means for collecting" refers to a device or method for acquiring audio data.

[0928] "Generative AI" is a system that uses artificial intelligence technology to generate new information from data.

[0929] A "means for inferring the content of dreams" is a method or device for analyzing collected data to infer the content of dreams.

[0930] A "means for detecting potential stress" is a method or device for identifying the stress an individual is experiencing from the inferred content of dreams.

[0931] A "means for suggesting ways to relieve stress" is a method or device that suggests appropriate ways to deal with discovered stress.

[0932] The "means for recommending security measures" is a method or device that suggests appropriate security measures to a user depending on the user's stress level.

[0933] "Deep learning" is a technology that uses multi-layered neural networks to analyze data and learn patterns and features.

[0934] A "mentally ill person" is an individual who has a mental illness.

[0935] "Mental health care" is a medical procedure that evaluates mental health status and provides treatment and support.

[0936] The following system configuration will be described as an embodiment of the present invention.

[0937] System Configuration

[0938] The system is centered around the "Sleep Security Assistant" application installed on the user's smartphone. The system includes the following main components:

[0939] 1. A method for collecting sleep talking

[0940] The smartphone microphone is used to collect the sleep talking of the user. The audio data is converted into text data using a speech recognition library (e.g., speech_recognition library).

[0941] 2. A way to guess the content of dreams

[0942] The collected text data of sleep talking is sent to a generative AI model (e.g., the rinna / japanese-gpt2-medium model using the transformers library) to infer the content of dreams. This generative AI model uses deep learning to analyze the content of dreams from the text data.

[0943] 3. Ways to detect underlying stress

[0944] The inferred dream content is used to discover potential stress levels of individuals, assess stress levels, and understand the user's mental health.

[0945] 4. Suggesting ways to relieve stress

[0946] The system suggests appropriate ways to relieve stress, such as deep breathing to relax or recommending activities that will help relieve stress.

[0947] 5. Means of recommending security measures

[0948] The system suggests appropriate security measures to users based on their stress level. For example, if stress increases, the system suggests changing passwords and recommends setting up two-factor authentication.

[0949] Processing flow

[0950] 1. Audio collection and recognition

[0951] The sleep talking uttered by the user while sleeping is collected using the smartphone's microphone and converted into text data using a voice recognition library.

[0952] 2. Guessing the content of the dream

[0953] The collected text data is input into a generative AI model to infer the content of dreams. The generative AI model generates dream content using the following prompt sentences:

[0954] Prompt: "User says in their sleep: 'Work isn't going well.' Infer the content of their dream from this sleep rambling."

[0955] 3. Identifying and assessing stress

[0956] The content of the dream is inferred to identify the individual's potential stress and assess its level.

[0957] 4. Generating Advice

[0958] The system will suggest appropriate ways to relieve the stress detected, such as "Take a deep breath to relax."

[0959] 5. Proposal of security measures

[0960] The system suggests appropriate security measures to users depending on their stress level. For example, if stress increases, the system will recommend specific measures such as "change your password" or "enable two-factor authentication."

[0961] In this way, it is possible to support the mental health of users and reduce security risks.

[0962] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0963] Step 1:

[0964] Sleep talking uttered by a user while sleeping is collected.

[0965] Input: Audio data from the smartphone microphone.

[0966] Specific operation: The device (smartphone) uses a microphone to collect the user's sleep talking in real time. The collected voice data is converted into text data using a speech recognition library (e.g., speech_recognition library).

[0967] Output: Sleep talking as text data.

[0968] Step 2:

[0969] The collected text data of sleep talking is input into a generative AI model to infer the content of the dream.

[0970] Input: Text data obtained in step 1.

[0971] Specific operation: The server inputs text data into a generative AI model (e.g., the rinna / japanese-gpt2-medium model using the transformers library) to infer the content of the dream. The generative AI model generates the content of the dream using the following prompt sentence:

[0972] Prompt: "User says in their sleep: 'Work isn't going well.' Infer the content of their dream from this sleep rambling."

[0973] Output: Inferred dream content.

[0974] Step 3:

[0975] The content of the dream is inferred to identify the individual's potential stress and assess its level.

[0976] Input: Dream content obtained in step 2.

[0977] How it works: The server analyzes the inferred dream content to discover the individual's potential stress. It uses a pre-trained stress assessment model to assess the stress level.

[0978] Output: Stress level assessment results.

[0979] Step 4:

[0980] We will suggest appropriate ways to relieve any stress that is discovered.

[0981] Input: Stress level assessment results obtained in step 3.

[0982] What it does: The server suggests ways to relieve stress based on your stress level, such as taking deep breaths to relax or recommending activities that will help you relieve stress.

[0983] Output: Suggestions for stress relief.

[0984] Step 5:

[0985] Appropriate security measures are suggested to users based on their stress level.

[0986] Input: Stress level assessment results obtained in step 3.

[0987] Specific behavior: When the stress level increases, the server suggests appropriate security measures to the user, such as suggesting a password change or setting up two-factor authentication.

[0988] Output: Suggested security measures.

[0989] Example 2

[0990] Next, a description will be given of Example 2 of Form 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."

[0991] In conventional mental health treatment, it has been difficult to accurately grasp the potential stress that an individual is experiencing. In addition, there has been a lack of methods for analyzing the content of dreams, making it impossible to understand an individual's psychological state in detail. This has made it difficult to suggest appropriate stress relief methods.

[0992] 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.

[0993] In this invention, the server includes a means for collecting voice data during sleep, a means for converting the collected voice data into text data using voice recognition software, and a means for inputting the converted text data into a deep learning model to infer the content of dreams. This makes it possible to accurately infer the content of dreams from an individual's sleep talking, discover potential stress, and suggest appropriate stress relief methods.

[0994] "Audio data" refers to data that is a digital recording of the sounds made by the user while they are sleeping.

[0995] "Speech recognition software" is a program for converting voice data into text data.

[0996] "Text data" is text information converted by voice recognition software.

[0997] A "deep learning model" is an artificial intelligence model that is trained using large amounts of data to extract specific patterns and features from input data.

[0998] "Dream content" refers to the specific scenario or situation of the dream the user is having while sleeping.

[0999] "Latent stress" is stress that the user is not aware of but that has a psychological or physiological effect on the user.

[1000] "Stress relief methods" are specific measures or suggestions for reducing or eliminating the stress that has been identified.

[1001] "Mental health care" is a medical procedure that assesses mental health and provides treatment and support.

[1002] MODE FOR CARRYING OUT THE INVENTION

[1003] The present invention is a system for analyzing sleep talking uttered by a user while sleeping and predicting the content of the dream. A specific embodiment of this system will be described below.

[1004] 1. Program Generation

[1005] The server generates a generative AI program that uses deep learning to predict the content of dreams. Specifically, it builds a model using deep learning frameworks such as TensorFlow and PyTorch.

[1006] 2. Program processing explanation

[1007] The server receives the user's sleep talking data as input. This sleep talking data is converted into text data using speech recognition software (e.g., Google Speech-to-Text API). The server then inputs this text data into a deep learning model. This deep learning model has been trained using a large amount of sleep talking data and corresponding dream content data.

[1008] The deep learning model analyzes the input sleep talking data and infers the content of the dream. The inferred results are fed back to the user. Specifically, the server sends the inferred dream content in text format to the user's device.

[1009] 3. Specific Examples

[1010] For example, if a user says in their sleep, "I'm swimming in the ocean," the server converts this sleep-talking data into text using speech recognition software, and inputs the text data "I'm swimming in the ocean" into the deep learning model. The deep learning model analyzes this text data and infers that "the user is dreaming about swimming in the ocean." This inference is sent to the user's device in text format.

[1011] 4. Examples of prompts

[1012] An example of a prompt might be:

[1013] "If a user says in their sleep, 'I'm swimming in the ocean,' how would a generative AI infer the content of their dream?"

[1014] By inputting this prompt into the generative AI model, it is possible to provide a detailed explanation of the system's processing.

[1015] In this way, the system analyzes the user's sleep talking data and infers the content of their dreams, which allows for a detailed understanding of an individual's psychological state, the detection of potential stress, and the suggestion of appropriate stress relief methods.

[1016] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1017] Step 1:

[1018] The user uses a device (e.g., a smartphone or dedicated recording device) to record their sleep talking. The device collects the user's sleep talking as audio data. This audio data is then sent to a server.

[1019] Input: User's sleep talking (audio data)

[1020] Output: Audio data sent to the server

[1021] Step 2:

[1022] The server converts the received voice data into text data using voice recognition software (e.g., Google Speech-to-Text API). Specifically, the server sends the voice data to the API and obtains the returned text data.

[1023] Input: Audio data

[1024] Output: Text data

[1025] Step 3:

[1026] The server inputs the acquired text data into a deep learning model, which is built using frameworks such as TensorFlow and PyTorch and trained using a large amount of sleep talking data and corresponding dream content data.

[1027] Input: Text data

[1028] Output: Dream content guess result

[1029] Step 4:

[1030] The server uses a deep learning model to analyze the text data and predict the content of the dream. Specifically, the model analyzes patterns in the text data and compares them with past data to predict the most likely dream content.

[1031] Input: Text data

[1032] Output: Inferred dream content

[1033] Step 5:

[1034] The server sends the predicted dream content in text format to the user's device, where the user can check the results.

[1035] Input: Inferred dream content

[1036] Output: The text message sent to the user's device.

[1037] For example, if a user says in their sleep, "I'm swimming in the ocean," the following happens:

[1038] 1. Step 1: The user records their sleep talking using a recording app on their smartphone. The device then sends the recorded audio data to the server.

[1039] 2. Step 2: The server sends the received voice data to the Google Speech-to-Text API, which converts the voice data into text data such as "swimming in the ocean" and returns it to the server.

[1040] 3. Step 3: The server inputs the text data "swimming in the ocean" into the deep learning model. The model analyzes this text data.

[1041] 4. Step 4: The deep learning model analyzes the text data and infers that "the user is dreaming about swimming in the ocean."

[1042] 5. Step 5: The server sends the guess result to the user's device as a text message saying, "You were dreaming about swimming in the ocean." The user can check the notification on their device to find out the content of their dream.

[1043] (Application example 2)

[1044] Next, a description will be given of Application Example 2 of Form 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."

[1045] Conventional dream analysis systems focus on inferring the content of dreams, but lack the ability to link the results of these inferences to the user's interests and provide specific content. This means that users are unable to not only understand the content of their dreams, but also to enjoy useful information and entertainment based on that content.

[1046] 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.

[1047] In this invention, the server includes a means for collecting sleep talking during sleep, a means for inferring dream content from the collected sleep talking using a generative AI, a means for delivering customized content relevant to the individual based on the inferred dream content, and a means for analyzing the user's interests and concerns based on the delivered content. This allows the user to not only understand the content of their dreams, but also enjoy customized content based on that content.

[1048] A "means for collecting sleep talking" is a device or software that records the sounds a user makes while sleeping and saves the data.

[1049] "Means for inferring dream content from collected sleep talking using generative AI" refers to algorithms and systems that analyze recorded sleep talking data and use deep learning models to infer dream content.

[1050] "Means for delivering customized content relevant to individuals based on the content of inferred dreams" refers to a system or service for providing content such as music, video, or articles that are appropriate for users based on the content of their inferred dreams.

[1051] "Means for analyzing user interests and concerns based on delivered content" refers to algorithms or systems for analyzing users' reactions and behavior to received content and identifying users' interests and concerns.

[1052] The system for implementing this invention mainly consists of three elements: a server, a terminal, and a user. The function of each element will be explained below.

[1053] server

[1054] The server has the following main functions:

[1055] 1. Collection of sleep talking data: The server receives and stores the sleep talking data sent from the device. The sleep talking data is a recording of the user's speech while sleeping.

[1056] 2. Dream content prediction: The server inputs the stored sleep talking data into a generative AI model and uses deep learning to predict the dream content. This generative AI model is trained using a large amount of sleep talking data and corresponding dream content data.

[1057] 3. Delivery of customized content: Based on the content of the predicted dream, the server selects customized content (music, video, articles, etc.) relevant to the user and delivers it to the device.

[1058] 4. Analysis of user interests: The server analyzes the user's reactions and behavior to the delivered content to identify the user's interests.

[1059] Terminal

[1060] The terminal has the following main functions:

[1061] 1. Recording sleep talking data: The device is equipped with a microphone to record the user's sleep talking. The recorded sleep talking data is sent to the server.

[1062] 2. Receiving and displaying content: The terminal receives the customized content delivered from the server and displays it to the user.

[1063] User

[1064] Users interact with the system as follows:

[1065] 1. Providing sleep talking data: The user uses the device to record sleep talking while sleeping.

[1066] 2. Content Usage: The user uses the customized content delivered to the device, for example, listening to music or watching videos based on the content of the predicted dream.

[1067] Hardware and software used

[1068] Hardware: Devices such as smartphones and tablets, and servers

[1069] Software: TensorFlow (running deep learning models), SoundDevice (collecting audio data), Librosa (preprocessing audio data)

[1070] Specific examples

[1071] For example, if a user talks in their sleep with the keyword "adventure," the server analyzes the sleep-talking data and infers that the dream content is adventure-related.The server then delivers customized content, such as an adventure movie trailer or an introductory article for an adventure novel, to the device.The user can then enjoy this content on their device.

[1072] Prompt Sentence Examples

[1073] "Analyze last night's sleep talking data and infer the content of your dream. Based on the inferred content of your dream, suggest related content."

[1074] The above is an embodiment of the present invention.

[1075] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1076] Step 1:

[1077] Recording of sleep talking data

[1078] The device records the user's sleep speech. Specifically, it uses a built-in microphone to collect audio data at regular intervals. The input is the user's sleep talking, and the output is the recorded audio data.

[1079] Step 2:

[1080] Sending sleep talking data

[1081] The device sends the recorded sleep talking data to a server. Specifically, the device uploads the audio data to the server via an internet connection. The input is the recorded audio data, and the output is the audio data sent to the server.

[1082] Step 3:

[1083] Saving sleep talking data

[1084] The server receives the sleep talking data sent from the device and stores it in a database. The input is the voice data sent from the device, and the output is the stored voice data.

[1085] Step 4:

[1086] Audio data preprocessing

[1087] The server preprocesses the stored sleep talking data. Specifically, it converts the audio data into MFCC (Mel-Frequency Cepstrum Coefficients) using the Librosa library. The input is the stored audio data, and the output is the preprocessed audio data.

[1088] Step 5:

[1089] Guessing the content of the dream

[1090] The server inputs the preprocessed audio data into a generative AI model to infer the content of the dream. Specifically, it uses TensorFlow to run a deep learning model to infer the content of the dream. The input is the preprocessed audio data, and the output is the inferred dream content.

[1091] Step 6:

[1092] Customized content selection

[1093] The server selects customized content relevant to the user based on the inferred dream content. Specifically, it searches a database related to the inferred dream content for appropriate content (music, video, articles, etc.). The input is the inferred dream content, and the output is the selected content.

[1094] Step 7:

[1095] Content Delivery

[1096] The server delivers the selected customized content to the terminal, specifically by transmitting content data to the terminal via an Internet connection, with the input being the selected content and the output being the content delivered to the terminal.

[1097] Step 8:

[1098] View content

[1099] The terminal receives the customized content delivered from the server and displays it to the user. Specifically, the content is played back using the terminal's display and speakers. The input is the content delivered from the server, and the output is the content displayed to the user.

[1100] Step 9:

[1101] Collecting user responses

[1102] The device records the user's actions and reactions to the delivered content. Specifically, it collects data such as the user's taps, clicks, and playback time. The input is the user's actions and reactions, and the output is the collected reaction data.

[1103] Step 10:

[1104] Analysis of user interests and concerns

[1105] The server analyzes the user's reaction data sent from the device and identifies the user's interests. Specifically, the collected reaction data is input into an analysis algorithm to infer the user's interests. The input is the collected reaction data, and the output is the inferred user's interests.

[1106] Example 3

[1107] Next, a description will be given of Example 3 of Form Example 3. 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."

[1108] Traditional mental health treatments often rely on patients' subjective reports, making it difficult to accurately identify potential stressors. Furthermore, there are limited means of discovering stressors that patients are not aware of, making it difficult to propose effective stress relief methods. This can potentially delay the improvement of mental health in psychiatric patients.

[1109] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1110] In this invention, the server includes means for collecting voice data during sleep, means for transmitting the collected voice data to the server, means for converting the voice data into text data, means for inferring dream content from the text data using a generative AI model, means for discovering potential stress the individual is experiencing from the inferred dream content, and means for suggesting ways to relieve the discovered stress. This makes it possible to infer the content of dreams from a patient's sleep talking, identify potential stressors, and suggest appropriate ways to relieve stress.

[1111] "Sound data during sleep" refers to data recorded from sounds made by a patient while sleeping.

[1112] "Means of collection" refers to devices or software for recording audio data, including, for example, smartphones and dedicated recording devices.

[1113] "Means for sending to a server" refers to the functions and protocols for uploading collected voice data to a server via a network.

[1114] "Means for converting voice data into text data" refers to a process for converting voice data into text information using voice recognition technology, and includes, for example, voice recognition software.

[1115] A "generative AI model" refers to an artificial intelligence model that uses deep learning to perform a specific task, including, for example, a natural language processing model.

[1116] "Method of inferring dream content" refers to the process of using a generative AI model to analyze and infer dream content from text data.

[1117] "Methods for detecting potential stress" refers to the process of analyzing the content of inferred dreams to identify potential stressors for an individual.

[1118] "Means for suggesting ways to relieve stress" refers to the process of suggesting appropriate ways to relieve stress based on the stress factors identified.

[1119] This invention is a system that can be used in the mental health treatment of psychiatric patients, and it analyzes the patient's sleep talking to infer the content of their dreams, identify potential stress factors, and propose appropriate stress relief methods. A specific embodiment of this system will be described below.

[1120] 1. System Configuration

[1121] The system is mainly composed of the following hardware and software.

[1122] Device: A device used to record the patient's sleep talking. This includes smartphones and dedicated recording devices.

[1123] Server: A computer system that receives and analyzes audio data.

[1124] Speech recognition software: Software for converting voice data into text data. Examples include the Google Cloud Speech-to-Text API.

[1125] Generative AI models: Artificial intelligence models for inferring dream content from text data. Examples include OpenAI's GPT-4.

[1126] 2. System Operation

[1127] The operation of the system is as follows.

[1128] Audio data collection

[1129] The device records the patient's sleep talking while they sleep. For example, by activating a recording app on a smartphone and placing it next to the pillow at night, high-quality recordings of sleep talking can be made.

[1130] Sending audio data

[1131] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using Wi-Fi or mobile data communication. When the upload is complete, the device notifies the user.

[1132] Analysis of audio data

[1133] The server uses speech recognition software to convert the received voice data into text data, which is then used in the next step.

[1134] Guessing the content of the dream

[1135] The server uses a generative AI model to infer dream content from text data, specifically by inputting the text data as prompts and generating dream content.

[1136] The discovery of stress

[1137] The server analyzes the content of the predicted dream and identifies potential stressors the patient may be experiencing. For example, if the dream contains content about being chased, it will determine this as a sign of stress.

[1138] Suggestions for stress relief

[1139] The server then proposes appropriate stress relief methods based on the identified stress factors. The proposals are sent to the user's device. For example, it suggests relaxation techniques or counseling.

[1140] 3. Specific Examples

[1141] Users use their smartphones to record their sleep talking at night. The recorded audio data is sent to a server and converted into text using speech recognition software. A generative AI model analyzes the text data and infers the content of the dream. For example, if the dream contains content about being chased, the server will interpret this as a sign of stress and suggest relaxation techniques or counseling.

[1142] 4. Examples of prompts

[1143] "You must analyze the patient's sleep talking and infer the content of their dream. Then, you must identify the stressors that the patient is experiencing based on the content of the dream and suggest appropriate ways to relieve stress."

[1144] In this way, a system is realized that contributes to improving the mental health of psychiatric patients through cooperation between the server, the terminal, and the user. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1145] Step 1:

[1146] Audio data collection

[1147] The device records the patient's sleep talking. Specifically, it uses a smartphone or a dedicated recording device and automatically starts recording at night. The input is the patient's sleep talking, and the output is the recorded audio file. When recording is finished, the audio file is temporarily saved.

[1148] Step 2:

[1149] Sending audio data

[1150] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using Wi-Fi or mobile data communication. The input is the recorded audio file, and the output is the audio data sent to the server. When the upload is complete, the device notifies the user.

[1151] Step 3:

[1152] Analysis of audio data

[1153] The server converts the received voice data into text data using voice recognition software. Specifically, it uses Google Cloud Speech-to-Text API or similar. The input is the voice data sent to the server, and the output is the converted text data. The conversion results are stored in a database.

[1154] Step 4:

[1155] Guessing the content of the dream

[1156] The server uses a generative AI model to infer dream content from text data. Specifically, the text data is input as a prompt, and the generative AI model (e.g., OpenAI's GPT-4) is used to generate dream content. The input is the converted text data, and the output is the inferred dream content. The generated dream content is stored in a database.

[1157] Step 5:

[1158] The discovery of stress

[1159] The server analyzes the estimated dream content and identifies potential stressors for the patient. Specifically, it applies psychological algorithms to analyze the frequency of specific keywords and phrases. The input is the estimated dream content, and the output is the identified stressors.

[1160] Step 6:

[1161] Suggestions for stress relief

[1162] The server proposes appropriate stress relief methods based on the identified stress factors. Specifically, it generates suggestions for relaxation techniques and counseling. The input is the identified stress factors, and the output is the proposed stress relief methods. The suggestions are sent to the user's device as push notifications.

[1163] (Application example 3)

[1164] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1165] In modern society, managing the mental health of psychiatric patients is an important issue. However, conventional methods make it difficult to accurately grasp a patient's underlying stress and propose appropriate ways to relieve it. In particular, technology for inferring dream content from sleep talking and detecting stress is still in its infancy, and a system for realizing effective mental health treatment is needed.

[1166] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1167] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for detecting potential stress in an individual from the inferred dream content, means for suggesting ways to relieve the detected stress, means for notifying the user of the suggested stress relief methods, and means for recording the sleep talking, inferring dream content, detecting stress, and suggesting ways to relieve it on a smartphone. This makes it possible to effectively manage the mental health of psychiatric patients, and to detect potential stress early and suggest appropriate ways to relieve it.

[1168] A "means for collecting sleep talking" is a device or software that records the sounds a user makes while sleeping and saves them as data.

[1169] "Generative AI" is a system that uses artificial intelligence technology to generate new information from input data, and in particular, uses deep learning to infer the content of dreams.

[1170] A "means for inferring the content of dreams" is an algorithm or system that analyzes collected sleep talking data and infers the content of the dream the user is having.

[1171] A "means for detecting potential stress in an individual" is a method or system for identifying potential stress factors that a user may be experiencing from the content of their inferred dreams.

[1172] The "means for suggesting ways to relieve stress" is a system that suggests ways for users to relieve stress that they can implement in response to discovered stress factors.

[1173] A "means for notifying the user" is a communication means or interface for informing the user of the proposed stress relief method.

[1174] The "means to be executed on the smartphone" refers to software and hardware for executing the above-mentioned means as an application on the smartphone.

[1175] The following describes an embodiment of the present invention. The system is implemented as an application that runs on a user's smartphone. The main functions of the system and how they are implemented are described below.

[1176] 1. Collection of sleep talking

[1177] The app uses the smartphone's microphone to collect sleep talking while the user is asleep. The app automatically starts recording when the user talks in their sleep and saves the audio data.

[1178] 2. Guessing the content of the dream

[1179] The collected sleep talking data is input into a generative AI model (e.g., GPT-3). The generative AI model uses deep learning to infer the content of dreams from sleep talking. The following prompt sentences are used in this process:

[1180] Example prompt sentence:

[1181] "Please guess the content of the dream from the following sleep talking: A scary dog ​​is chasing me."

[1182] 3. Discovering stress

[1183] We again use a generative AI model to discover potential stressors from the inferred dream content. In this process, we use the following prompt:

[1184] Example prompt sentence:

[1185] "Identify the stress in the following dream and suggest ways to relieve it: The user dreams of being chased by a dog."

[1186] 4. Suggestions for ways to relieve stress

[1187] For each stress factor detected, the generative AI model will suggest an appropriate stress relief method, which will then be notified to the user.

[1188] 5. Notice to Users

[1189] The proposed stress relief methods are notified to users using the smartphone's notification function, allowing users to implement the stress relief methods at the appropriate time.

[1190] Hardware and software used

[1191] Hardware: Smartphone with microphone

[1192] Software: Python, speech_recognition library, transformers library (generative AI model)

[1193] Specific examples

[1194] For example, if a user says in their sleep, "A scary dog ​​is chasing me," the system will record the sleep talking and use a generative AI model to infer that "the user is dreaming about being chased by a dog." It will then identify the cause of stress as "fear" and suggest ways to relieve it, such as "deep breathing techniques to relax," and notify the user via their smartphone's notification function.

[1195] In this way, the system can effectively manage the user's mental health, detect potential stress early, and suggest appropriate ways to relieve it.

[1196] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1197] Step 1:

[1198] Sleep talking uttered by a user while sleeping is collected.

[1199] Input: User's voice data

[1200] Output: Audio file of recorded sleep talking

[1201] What it does: It uses the smartphone's microphone to automatically start recording when the user talks in their sleep and saves the audio data.

[1202] Step 2:

[1203] The collected sleep talking data is converted into text.

[1204] Input: Audio file of recorded sleep talking

[1205] Output: Sleep talking data in text format

[1206] What it does: Converts audio data to text using the speech_recognition library.

[1207] Step 3:

[1208] Text-based sleep talking data is input into a generative AI model to infer the content of the dream.

[1209] Input: Text-formatted sleep talking data

[1210] Output: Inferred dream content

[1211] Specific operation: The following prompt sentence is input into a generative AI model (e.g., GPT-3) to generate the content of the dream.

[1212] "Please guess the content of the dream from the following sleep talking: [Sleep talking data in text format]"

[1213] Step 4:

[1214] The predicted dream content is fed into a generative AI model to discover potential stressors.

[1215] Input: Inferred dream content

[1216] Output: Stressors discovered

[1217] Specific operation: The following prompt sentence is input into the generative AI model to identify stress factors.

[1218] "Please identify the stress in the following dream and suggest ways to relieve it: [guessed dream content]"

[1219] Step 5:

[1220] For each stress factor identified, a generative AI model is used to suggest ways to relieve stress.

[1221] Input: Discovered stressors

[1222] Output: Suggested stress relief methods

[1223] Specific operation: Stress factors are input into the generative AI model, and appropriate stress relief methods are generated.

[1224] Step 6:

[1225] Inform the user of suggested stress relief methods.

[1226] Input: Suggested stress relief method

[1227] Output: User notification

[1228] What it does: Uses smartphone notifications to notify users of suggested stress relief methods.

[1229] Through these steps, the system can effectively manage the user's mental health, detect potential stress early, and suggest appropriate ways to relieve it.

[1230] 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.

[1231] "Example 1"

[1232] In one embodiment of the present invention, the emotion engine recognizes the user's emotions and adjusts stress relief methods based on the user's emotional state. Specifically, the emotion engine analyzes the user's emotions from their voice, facial expressions, and behavior, and provides feedback to the generative AI. The generative AI uses this information to more accurately predict the content of the user's dreams and suggest stress relief methods tailored to each individual user.

[1233] "Example 2"

[1234] In another embodiment, the emotion engine monitors the user's emotional state in real time and updates stress relief methods in response to changes in the emotional state. Specifically, when the user expresses negative emotions such as anger or sadness, the emotion engine feeds that information back to the generative AI, which then uses that information to update the stress relief methods. For example, if the user is feeling angry, the generative AI suggests stress relief methods such as relaxation or meditation.

[1235] "Example 3"

[1236] In another embodiment, the emotion engine and generative AI work together to propose stress relief methods that take into account both the user's emotional state and the content of their dreams. Specifically, the emotion engine analyzes the user's emotional state and feeds the results back to the generative AI. The generative AI combines this information with the content of their dreams inferred from their sleep talking to more accurately understand the user's stress and propose ways to relieve it.

[1237] The processing flow of each embodiment will be described below.

[1238] "Example 1"

[1239] Step 1: The emotion engine analyzes the user's emotions from their voice, facial expressions, behavior, etc.

[1240] Step 2: The emotion engine feeds back the analysis results to the generative AI.

[1241] Step 3: The generative AI uses this information to infer the content of the user's dream.

[1242] Step 4: The generative AI suggests stress relief methods that are appropriate for each individual user.

[1243] "Example 2"

[1244] Step 1: The emotion engine monitors the user's emotional state in real time.

[1245] Step 2: If the user expresses negative emotions, the emotion engine feeds that information back to the generative AI.

[1246] Step 3: The generative AI uses the information to update stress relief strategies.

[1247] "Example 3"

[1248] Step 1: The emotion engine analyzes the user's emotional state and feeds the results back to the generative AI.

[1249] Step 2: The generative AI combines this information with the dream content inferred from sleep talking to more accurately understand the user's stress.

[1250] Step 3: Generative AI proposes a solution.

[1251] Example 1

[1252] Next, a description will be given of Example 1 of Form 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."

[1253] In modern society, many people suffer from stress, which has a negative impact on the quality of sleep and mental health. However, there is no system that can analyze sleep states in detail and propose stress relief methods suited to individual users. In particular, there is no established method yet for analyzing sleep talking during sleep to infer the content of dreams and then using that information to discover and relieve stress. Therefore, there is a need for a system that can analyze sleep talking during sleep and effectively relieve users' stress.

[1254] 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.

[1255] In this invention, the server includes means for collecting sleep talking during sleep, means for transmitting the collected sleep talking to the server, means for the server to transfer the sleep talking data to the generative AI, means for the generative AI to infer the content of dreams from the sleep talking using deep learning, means for transmitting the inferred dream content to the advice generation module, means for the advice generation module to discover the individual's latent stress from the inferred dream content, means for suggesting ways to relieve the discovered stress, and means for an emotion engine to analyze the user's emotions and feed the results back to the generative AI. This makes it possible to analyze the user's sleep talking during sleep, discover stress from the content of dreams, and suggest ways to relieve it.

[1256] The "means for collecting sleep talking during sleep" refers to a device or function for collecting with high sensitivity the sounds made by the user while sleeping.

[1257] The "means for transmitting collected sleep talking to a server" refers to a device or function for transferring collected voice data to a server using the Internet or other communication means.

[1258] "Means by which the server transfers sleep talking data to the generative AI" refers to devices or functions that allow the server to pass on the voice data it receives to the generative AI.

[1259] "Means for generative AI to infer dream content from sleep talking using deep learning" refers to devices or functions that allow generative AI to analyze audio data and use deep learning algorithms to infer the content of the dream the user is having.

[1260] "Means for sending the inferred dream content to the advice generation module" refers to a device or function for transferring the dream content inferred by the generative AI to the advice generation module.

[1261] "Means for discovering the potential stress an individual is experiencing from the dream content inferred by the advice generation module" refers to a device or function that analyzes the dream content received by the advice generation module from the generative AI and identifies the potential stress the user is experiencing.

[1262] The "means for proposing a method for relieving the discovered stress" refers to a device or function that allows the advice generation module to suggest to the user an appropriate method for relieving the identified stress.

[1263] "Means for the emotion engine to analyze the user's emotions and provide the results as feedback to the generative AI" refers to devices or functions that allow the emotion engine to analyze emotions from the user's voice, facial expressions, actions, etc., and provide the analysis results to the generative AI.

[1264] MODE FOR CARRYING OUT THE INVENTION

[1265] This invention is a system that analyzes the sleep talking of a user while they sleep, identifies stress from the content of their dreams, and suggests ways to relieve it. This system is composed of the following main components:

[1266] 1. Recording sleep talking using a microphone

[1267] A highly sensitive microphone connected to the device picks up the sleep talking uttered by the user while sleeping. This microphone has a noise-canceling function, which eliminates surrounding noise and clearly records the sleep talking. For example, if the user says "help me" in their sleep, the audio data will be collected by the microphone.

[1268] 2. Sending collected data to the server

[1269] The device sends the collected sleep talking audio data to a server. This transmission is done in real time, and the data is encrypted and transferred securely. For example, when the user's sleep talking data is sent from the device to the server, the SSL / TLS protocol is used to ensure data security.

[1270] 3. The server transfers data to the generative AI

[1271] The server then forwards the received sleep-talking audio data to a generative AI. This generative AI is built using deep learning frameworks such as TensorFlow and PyTorch. For example, the server sends the sleep-talking data to the generative AI's API endpoint.

[1272] 4. Generative AI for dream content prediction

[1273] The generative AI analyzes the received sleep talking audio data and infers the content of the dream the user is having. This analysis uses voice recognition technology and natural language processing technology. For example, if the sleep talking contains the phrase "I'm being chased," the generative AI infers that the user is having a stressful dream.

[1274] 5. Sending the inference results to the advice generation module

[1275] The content of the dream inferred by the generative AI is sent to the advice generation module via the server. The server converts this data into an appropriate format and passes it to the advice generation module. For example, if the generative AI infers that "the user is dreaming of being chased," the server sends the result to the advice generation module.

[1276] 6. Proposing stress relief methods using the advice generation module

[1277] The advice generation module detects potential stress the user may be experiencing from the estimated dream content and suggests ways to alleviate it. The suggestions are based on the user's past data and common stress relief methods. For example, specific advice such as "Try yoga to relax" is provided.

[1278] 7. Emotional analysis and feedback of users using an emotion engine

[1279] The emotion engine analyzes the user's emotions from their voice, facial expressions, and behavior. The results of this analysis are fed back to the generative AI. The generative AI uses this information to more accurately predict the content of the user's dreams and suggest stress relief methods that are tailored to each individual user. For example, if a user feels stressed during the day, that information is fed back from the emotion engine to the generative AI, allowing it to provide more appropriate advice.

[1280] Examples and prompts

[1281] Specific examples

[1282] If a user says in their sleep, "Help me, someone is chasing me," a microphone connected to the device will collect this sleep talking. The collected data is sent from the device to a server, which then forwards it to the generative AI. From this sleep talking, the generative AI infers that the user is dreaming of being chased. The inference result is sent via the server to an advice generation module, which then provides advice such as, "Try yoga to relax."

[1283] Prompt Sentence Examples

[1284] "Collect the sleep talking that the user utters while sleeping, analyze that data, and infer the content of their dreams. Discover the potential stress that the user is experiencing from the inferred dream content, and generate advice suggesting ways to relieve that stress."

[1285] This system supports the user's mental health by analyzing the user's sleep talking, identifying stress from the content of their dreams, and suggesting ways to relieve it.

[1286] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1287] Step 1:

[1288] A highly sensitive microphone connected to the device collects the sleep talking uttered by the user while sleeping. The microphone is equipped with a noise-canceling function, eliminating surrounding noise and recording the sleep talking clearly. The input is the user's sleep talking audio, and the output is the collected audio data. For example, if the user says "help me" in their sleep, that audio data will be collected by the microphone.

[1289] Step 2:

[1290] The device sends the collected sleep talking audio data to a server. This transmission is done in real time, and the data is encrypted and transferred securely. The input is the collected audio data, and the output is the audio data sent to the server. For example, when the user's sleep talking data is transmitted from the device to the server, the SSL / TLS protocol is used to ensure data security.

[1291] Step 3:

[1292] The server forwards the received sleep talking audio data to a generative AI. This generative AI is built using deep learning frameworks such as TensorFlow and PyTorch. The input is the audio data sent to the server, and the output is the audio data forwarded to the generative AI. For example, the server sends the sleep talking data to the API endpoint of the generative AI.

[1293] Step 4:

[1294] The generative AI analyzes the received sleep talking audio data and infers the content of the dream the user is having. This analysis uses voice recognition technology and natural language processing technology. The input is the audio data transferred to the generative AI, and the output is the inferred dream content. For example, if the sleep talking includes the phrase "I'm being chased," the generative AI will infer that the user is having a stressful dream.

[1295] Step 5:

[1296] The dream content inferred by the generative AI is sent to the advice generation module via the server. The server converts this data into an appropriate format and passes it to the advice generation module. The input is the inferred dream content, and the output is the data sent to the advice generation module. For example, the generative AI infers that "the user is dreaming of being chased," and the server sends this result to the advice generation module.

[1297] Step 6:

[1298] The advice generation module detects potential stress the user may be experiencing from the estimated dream content and suggests ways to relieve it. These suggestions are based on the user's past data and common stress relief methods. The input is the data sent to the advice generation module, and the output is the suggested stress relief method. For example, specific advice such as "Try yoga to relax" is provided.

[1299] Step 7:

[1300] The emotion engine analyzes emotions from the user's voice, facial expressions, behavior, etc. The results of this analysis are fed back to the generative AI. The generative AI uses this information to more accurately predict the content of the user's dreams and propose stress relief methods that are suited to each individual user. The input is the user's emotional data, and the output is the analysis results fed back to the generative AI. For example, if a user feels stressed during the day, that information is fed back from the emotion engine to the generative AI, which then provides more appropriate advice.

[1301] (Application example 1)

[1302] Next, a description will be given of Application Example 1 of Form 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."

[1303] In modern society, many people are suffering from latent stress, which has a negative impact on health. However, it is difficult to identify the causes of stress and propose appropriate ways to relieve it. Furthermore, there is no system that can analyze a user's emotional state in real time and adjust stress relief methods based on that analysis. This makes it difficult to effectively manage users' mental health.

[1304] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means. In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for discovering potential stress that an individual has from the inferred dream content, means for suggesting methods for relieving the discovered stress, means for analyzing the user's emotional state, and means for adjusting the stress relief method based on the analyzed emotional state. This makes it possible to identify the user's potential stress and suggest an appropriate stress relief method according to the individual's emotional state.

[1305] The "means for collecting sleep talking during sleep" refers to a device or system for recording sounds made by a user while sleeping.

[1306] "Generative AI" is an artificial intelligence technology that analyzes collected data and generates specific information.

[1307] "Means for inferring the content of dreams" refers to technology for analyzing and inferring the content of a user's dreams based on collected sleep talking data.

[1308] The "means for detecting potential stress" is a technology for identifying the stress a user may be experiencing from the content of their inferred dreams.

[1309] The "means for suggesting a method for relieving stress" is a technique for suggesting to the user an appropriate method for relieving the discovered stress.

[1310] "Means for analyzing the user's emotional state" refers to technology for analyzing emotions from the user's voice, facial expressions, behavior, etc.

[1311] The "means for adjusting a stress relief method based on emotional state" is a technique for appropriately adjusting the proposed stress relief method in accordance with the analyzed emotional state.

[1312] The following describes an embodiment of the present invention. First, the system is composed of the following main components.

[1313] 1. How to collect sleep talking during sleep:

[1314] The sleep talking that the user makes while sleeping is collected using the smartphone microphone, and the collected audio data is stored in local storage.

[1315] 2. Generative AI:

[1316] The collected sleep talking data is analyzed using deep learning frameworks such as TensorFlow. The generative AI includes a model for inferring dream content from sleep talking.

[1317] 3. How to guess the content of your dream:

[1318] The generative AI analyzes the sleep talking data and infers the content of the dream the user is having. The inferred dream content is used in the next step.

[1319] 4. Ways to spot underlying stress:

[1320] The content of the dream is inferred, and potential stress that the user may be experiencing is identified. The content of the dream is analyzed using a natural language processing library (such as NLTK) to identify the cause of stress.

[1321] 5. Suggested ways to relieve stress:

[1322] The system suggests appropriate ways to relieve the stress detected to the user. The suggestions are sent to the user using a notification service (such as Twilio).

[1323] 6. Means of analyzing the user's emotional state:

[1324] The system uses the smartphone camera to analyze the user's facial expressions in real time. It uses image processing libraries such as OpenCV to collect facial expression data and analyzes it with an emotion engine.

[1325] 7. How to adjust your stress management strategies based on your emotional state:

[1326] Based on the analyzed emotional state, the proposed stress relief method is adjusted appropriately, providing the most appropriate relief method for the user's current emotional state.

[1327] Specific examples

[1328] For example, if a user says in their sleep, "Work isn't going well," the generative AI will infer "work-related stress." If the emotion engine detects "anxiety" from the user's facial expression, the advice generation module will suggest "breathing techniques to relax."

[1329] Prompt Sentence Examples

[1330] Analyze the user's sleep talking data and infer the content of their dreams. Next, identify potential stress from the inferred dream content and suggest ways to relieve it. Take the user's emotional state into consideration.

[1331] In this way, a system can be realized that utilizes data from users while they are sleeping, identifies their individual stress levels, and suggests appropriate ways to relieve them.

[1332] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1333] Step 1:

[1334] The user's device (smartphone) uses a microphone to collect sleep talking while sleeping. The collected audio data is stored in local storage. The input is the user's sleep talking audio, and the output is the stored audio data.

[1335] Step 2:

[1336] The device sends the saved sleep-talking audio data to the server. The server inputs the received audio data into the generative AI. The input is the sleep-talking audio data, and the output is the input data for the generative AI.

[1337] Step 3:

[1338] The server's generative AI uses TensorFlow to analyze sleep-talking audio data and infer the content of dreams. The input is the sleep-talking audio data, and the output is the inferred dream content. Specifically, the audio data is converted into text and analyzed using a deep learning model.

[1339] Step 4:

[1340] The server detects potential stress based on the inferred dream content. It uses a natural language processing library (such as NLTK) to analyze the dream content and identify the causes of stress. The input is the inferred dream content, and the output is the identified causes of stress.

[1341] Step 5:

[1342] The server proposes appropriate solutions to the detected stress. The proposals are sent to the user using a notification service (such as Twilio). The input is the identified cause of stress, and the output is the proposed stress relief solution. The specific operation is to generate a stress relief solution and create a notification message.

[1343] Step 6:

[1344] The user's device uses a camera to analyze the user's facial expressions in real time. Using image processing libraries such as OpenCV, the facial expression data is collected and analyzed by the emotion engine. The input is the user's facial expression data, and the output is the analyzed emotional state.

[1345] Step 7:

[1346] The server appropriately adjusts the proposed stress relief method based on the analyzed emotional state. The input is the analyzed emotional state and the proposed stress relief method, and the output is the adjusted stress relief method. Specifically, the server optimizes the stress relief method according to the emotional state.

[1347] In this way, a system can be realized that utilizes data from users while they are sleeping, identifies their individual stress levels, and suggests appropriate ways to relieve them.

[1348] Example 2

[1349] Next, a description will be given of Example 2 of Form 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."

[1350] Conventional systems have difficulty accurately predicting the content of a user's dreams while they are sleeping, and are unable to monitor the user's emotional state in real time or suggest appropriate stress relief methods, making it difficult to effectively support the user's mental health.

[1351] 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.

[1352] In this invention, the server includes means for collecting voice data during sleep, means for converting the collected voice data into text data, means for inputting the converted text data into a deep learning model to infer the content of dreams, means for detecting potential stress the individual is experiencing from the inferred dream content, means for suggesting methods for relieving the discovered stress, means for monitoring the user's emotional state in real time, and means for updating the stress relief methods in response to changes in the emotional state. This makes it possible to accurately infer the content of the user's dreams, monitor the emotional state in real time, and suggest appropriate stress relief methods.

[1353] "Sound data during sleep" is data that records sounds made by the user while sleeping.

[1354] "Means for converting into text data" refers to technology or devices for converting voice data into text information.

[1355] A "deep learning model" is a type of artificial intelligence that is trained using large amounts of data to learn complex patterns and relationships.

[1356] A "means for predicting the content of dreams" is a technology or device for predicting the content of a user's dream from input data.

[1357] "Means for detecting latent stress" are technologies and devices for detecting stress that the user is not aware of.

[1358] "Means for suggesting ways to relieve stress" refers to techniques or devices that suggest specific ways to reduce stress to the user.

[1359] "Means for monitoring emotional state in real time" refers to technology or devices for observing and recording changes in a user's emotions in real time.

[1360] The "means for updating stress relief methods in response to changes in the emotional state" refers to a technique or device for appropriately changing the stress relief methods suggested based on changes in the user's emotions.

[1361] This invention is a system that collects voice data while a user is sleeping, analyzes the data to infer the content of dreams, and monitors the user's emotional state in real time to suggest appropriate stress relief methods.

[1362] First, the device records the user's speech while sleeping through a microphone. This recorded speech data is temporarily stored on the device. Next, the device uses speech recognition software such as the Google Speech-to-Text API to convert the recorded speech data into text data. This converted text data is then sent to a server.

[1363] The server inputs the received text data into a deep learning model. This deep learning model is built using frameworks such as TensorFlow and PyTorch and is trained using large amounts of audio data and corresponding dream content data. The generative AI model analyzes the input text data and infers the content of the user's dream. The inferred results are stored on the server.

[1364] The server then sends the inference results to the device, which then provides feedback to the user. For example, if a user says in their sleep, "I'm swimming in the ocean," the generative AI model infers that "the user is dreaming about swimming in the ocean," and the device notifies the user of the result.

[1365] Examples:

[1366] User sleep talking data: "Swimming in the ocean"

[1367] Generative AI model predicts: "The user is dreaming about swimming in the ocean."

[1368] Additionally, the device monitors the user's facial expressions and vocal tone in real time using emotion recognition software such as Affectiva and Microsoft Azure Emotion API. If the user displays negative emotions such as anger or sadness, the device sends this information to the server.

[1369] The server inputs the received emotional data into the emotion engine. The emotion engine detects changes in the user's emotional state and feeds that information back to the generative AI model. The generative AI model updates stress relief methods according to the user's emotional state. For example, if the user is feeling angry, the generative AI model will suggest stress relief methods such as relaxation or meditation.

[1370] Examples:

[1371] User sentiment data: "I'm really frustrated today"

[1372] The emotion engine determines: "The user is angry."

[1373] Generative AI model suggestion: "Try relaxation or meditation."

[1374] In this way, the system can accurately predict the content of a user's dreams, monitor their emotional state in real time, and suggest appropriate stress relief methods.

[1375] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1376] Step 1:

[1377] The device records the sounds made by the user while sleeping through a microphone. The recorded sound data is temporarily stored in the device. The input is the user's voice, and the output is the recorded sound data. Specifically, the device's microphone captures the sound and saves it as a recording file.

[1378] Step 2:

[1379] The device converts the recorded audio data into text data using the Google Speech-to-Text API. The input is the recorded audio data, and the output is the converted text data. Specifically, the device sends the audio file to the API and receives the returned text data.

[1380] Step 3:

[1381] The terminal transmits the converted text data to the server. The input is the text data, and the output is the transmission of the data to the server. In concrete terms, the terminal uploads the text data to the server via a network.

[1382] Step 4:

[1383] The server inputs the received text data into the deep learning model. The input is text data, and the output is the predicted dream content. Specifically, the server passes the text data to the deep learning model, which then analyzes it.

[1384] Step 5:

[1385] The generative AI model analyzes input text data and infers the content of dreams. The input is text data, and the output is the inferred dream content. Specifically, the model analyzes text data based on the data it has been trained on and infers the content of dreams.

[1386] Step 6:

[1387] The server sends the guess results to the terminal. The input is the guess result of the dream content, and the output is the transmission of data to the terminal. In concrete terms, the server sends the guess results to the terminal via the network.

[1388] Step 7:

[1389] The device then feeds back the results of its guesses to the user. The input is the guessed content of the dream, and the output is a notification to the user. Specifically, the device communicates the results to the user through a screen display or audio notification.

[1390] Step 8:

[1391] The device monitors the user's facial expressions and vocal tone in real time. The input is the user's facial expressions and vocal tone, and the output is emotional data. Specifically, the device collects data using a camera and microphone and analyzes it with emotion recognition software.

[1392] Step 9:

[1393] The device sends the collected emotion data to the server. The input is emotion data, and the output is data transmission to the server. Specifically, the device uploads the emotion data to the server via the network.

[1394] Step 10:

[1395] The server inputs the received emotion data into the emotion engine. The input is emotion data, and the output is the analysis result of the emotional state. Specifically, the server passes the emotion data to the emotion engine for analysis.

[1396] Step 11:

[1397] The emotion engine detects changes in the user's emotional state and feeds that information back to the generative AI model. The input is emotional data, and the output is information about changes in emotional state. Specifically, the emotion engine analyzes the data and detects changes.

[1398] Step 12:

[1399] The generative AI model receives feedback from the emotion engine and updates stress relief methods according to the user's emotional state. The input is information on changes in emotional state, and the output is updated stress relief methods. Specifically, the model recalculates stress relief methods based on the feedback.

[1400] Step 13:

[1401] The server sends the updated stress relief method to the terminal. The input is the updated stress relief method, and the output is data transmission to the terminal. In specific operations, the server sends the updated information to the terminal through the network.

[1402] Step 14:

[1403] The device proposes updated stress relief methods to the user. The input is the updated stress relief methods, and the output is the proposal to the user. As a specific operation, the device notifies the user of the proposal through a screen display or a voice notification.

[1404] (Application example 2)

[1405] Next, a description will be given of Application Example 2 of Form 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."

[1406] In modern society, stress and mental strain are increasing, causing many people to suffer from mental health problems. In particular, there is a need for a system that can effectively detect potential stress and emotional changes that occur during sleep and suggest appropriate stress relief methods. However, current systems lack the functionality to monitor the user's emotional state in real time and dynamically update stress relief methods, preventing effective mental health care.

[1407] 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.

[1408] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for discovering potential stress that an individual has from the inferred dream content, means for suggesting ways to relieve the discovered stress, means for monitoring the user's emotional state in real time, and means for updating the stress relief methods in response to changes in the emotional state. This makes it possible to grasp the user's emotional state in real time and dynamically suggest appropriate stress relief methods.

[1409] The "means for collecting sleep talking" refers to a device or software for recording sounds made by a user while they are sleeping.

[1410] "Generative AI" is an artificial intelligence technology that generates new information based on collected data, and in particular uses deep learning.

[1411] The "means for inferring the content of dreams" refers to algorithms or models that analyze collected sleep talking data and infer the content of the dreams the user is having.

[1412] The "means for discovering potential stress in an individual" is a method or system for identifying a user's potential stress factors from the content of their inferred dreams.

[1413] The "means for suggesting ways to relieve stress" is a function for providing the user with an appropriate way to relieve stress based on the discovered stress factors.

[1414] "Means for monitoring the user's emotional state in real time" refers to devices or software that monitor the user's biometric information, such as heart rate and facial expression, in real time and analyze the user's emotional state.

[1415] The "means for updating stress relief methods according to changes in emotional state" is a function for detecting changes in the user's emotional state and dynamically changing the stress relief methods suggested in response to those changes.

[1416] The following system configuration will be described as an embodiment of the present invention.

[1417] System Configuration

[1418] Hardware

[1419] Smartphone: Collects sleep talking data and monitors the user's emotional state.

[1420] Smart glasses: Monitor the user's emotional state and suggest ways to relieve stress.

[1421] Server: Analyzes collected data and runs generative AI models.

[1422] software

[1423] TensorFlow: For building and inferencing deep learning models.

[1424] SoundDevice: Collects audio data.

[1425] Librosa: Preprocessing audio data.

[1426] Processing flow

[1427] 1. Collection of sleep talking data

[1428] The sleep talking uttered by the user while sleeping is collected using the smartphone microphone.

[1429] The collected voice data is sent to a server.

[1430] 2. Guessing the content of the dream

[1431] The server preprocesses the collected sleep talking data using Librosa and extracts MFCC features.

[1432] The extracted features are input into a TensorFlow deep learning model to infer the content of the dream.

[1433] 3. Identifying underlying stress

[1434] The content of the dream is inferred to identify potential stressors the user may be experiencing.

[1435] The identified stress factors are stored on a server.

[1436] 4. Suggestions for ways to relieve stress

[1437] Based on the identified stress factors, the server uses a generative AI model to suggest appropriate stress relief methods.

[1438] The proposed stress relief methods are communicated to the user via their smartphone or smart glasses.

[1439] 5. Monitoring your emotional state

[1440] Using sensors on smartphones and smart glasses, the system monitors the user's heart rate and facial expressions in real time.

[1441] The collected biometric information is transmitted to a server.

[1442] 6. Update your stress relief methods

[1443] The server analyzes changes in the user's emotional state and dynamically updates stress relief methods as needed.

[1444] The updated stress relief method is again notified to the user.

[1445] Specific examples

[1446] For example, a smartphone can collect sleep talking while a user sleeps and notify the user the next morning, saying, "Last night's dream was about adventure." If a user feels stressed at work, smart glasses can suggest, "Try relaxation or meditation."

[1447] Prompt Sentence Examples

[1448] "Guess what I dreamed about last night."

[1449] "Please suggest ways for users to relieve stress if they are feeling angry."

[1450] In this way, the embodiments for carrying out the invention have been specifically shown.

[1451] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1452] Step 1:

[1453] Collection of sleep talking data

[1454] Input: Audio made by the user while sleeping

[1455] How it works: The app uses the smartphone's microphone to record the user's sleep talking in real time. The recorded audio data is temporarily stored on the smartphone.

[1456] Output: Recorded sleep talking data

[1457] Step 2:

[1458] Sending sleep talking data

[1459] Input: Recorded sleep talking

[1460] How it works: The smartphone sends the collected sleep talking data to a server using an internet connection.

[1461] Output: Sleep talking data sent to the server

[1462] Step 3:

[1463] Guessing the content of the dream

[1464] Input: Sleep talking data sent to the server

[1465] How it works: The server preprocesses the sleep talking data using Librosa and extracts MFCC features. The extracted features are then input into a TensorFlow deep learning model to infer the content of the dream.

[1466] Output: Inferred dream content

[1467] Step 4:

[1468] Identifying underlying stress

[1469] Input: Inferred dream content

[1470] How it works: The server analyzes the content of the predicted dream and identifies potential stressors that the user may be experiencing. The identified stressors are then stored in a database.

[1471] Output: Identified potential stressors

[1472] Step 5:

[1473] Suggestions for stress relief

[1474] Input: Identified potential stressors

[1475] How it works: The server uses a generative AI model to generate appropriate stress relief methods based on the identified stress factors. The generated stress relief methods are then sent to a smartphone or smart glasses.

[1476] Output: Suggested stress relief methods

[1477] Step 6:

[1478] Emotional state monitoring

[1479] Input: User's biometric information such as heart rate and facial expression

[1480] How it works: Using sensors in smartphones and smart glasses, the system monitors the user's heart rate and facial expressions in real time, and transmits the collected biometric information to a server.

[1481] Output: Biometric information sent to the server

[1482] Step 7:

[1483] Updates on stress relief methods

[1484] Input: Biometric information sent to the server

[1485] How it works: The server analyzes changes in the user's emotional state and dynamically updates the stress relief methods as needed. The updated stress relief methods are then sent to the smartphone or smart glasses.

[1486] Output: Updated stress relief methods

[1487] In this way, the specific operations performed at each processing step and their inputs and outputs are clearly indicated.

[1488] Example 3

[1489] Next, a description will be given of Example 3 of Form Example 3. 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."

[1490] Conventional mental health treatment systems often rely on subjective reports from patients, making it difficult to accurately identify stress and suggest appropriate ways to relieve it. Furthermore, if patients are not aware of their own stress, appropriate treatment may be delayed. This has led to the problem of insufficient improvement in the mental health of psychiatric patients.

[1491] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1492] In this invention, the server includes means for collecting voice data during sleep, means for converting the collected voice data into text data, means for analyzing the converted text data to infer the content of dreams, means for detecting potential stress in the individual from the inferred dream content, means for suggesting ways to relieve the discovered stress, and means for notifying the individual of the suggested stress relief methods. This makes it possible to analyze the content of dreams from the patient's sleep talking, identify potential stress, and suggest appropriate stress relief methods.

[1493] "Sound data during sleep" refers to sounds made by an individual while sleeping, and is data that includes sounds made unconsciously, such as talking in one's sleep.

[1494] "Means of collection" refers to the devices and software used to capture and store audio data.

[1495] "Means for converting into text data" refers to speech recognition technology for converting voice data into text information.

[1496] "Means of analyzing and inferring the content of dreams" refers to natural language processing technology that analyzes converted text data and infers the content of dreams from it.

[1497] "Means for detecting potential stress" refers to algorithms and analytical techniques for identifying potential stress an individual may be experiencing from the content of their inferred dreams.

[1498] "Means for suggesting stress relief methods" refers to a generative AI model that generates and suggests appropriate stress relief methods for identified stress.

[1499] "Means for notifying" refers to a communication means or display device for informing an individual of the proposed stress relief method.

[1500] This invention is a system that can be used in mental health treatment for psychiatric patients, and identifies the patient's stress by collecting and analyzing voice data during sleep and proposing appropriate stress relief methods.

[1501] Hardware and software used

[1502] server

[1503] The server uses the following hardware and software:

[1504] Speech Recognition Software: We use speech recognition technology to convert voice data into text data, specifically, we use Google Cloud Speech-to-Text.

[1505] Natural language processing engine: We use natural language processing technology, specifically SpaCy, to analyze the converted text data and infer the content of dreams.

[1506] Stress analysis algorithm: To identify stress from the inferred dream content, stress analysis technology is used, specifically IBM Watson.

[1507] Generative AI model: We use a generative AI model, specifically OpenAI GPT-4, to generate appropriate stress relief methods based on the identified stress.

[1508] Terminal

[1509] The terminal uses the following hardware and software:

[1510] Microphone: Used to record the patient's sleep talking.

[1511] Communication module: Used to send recorded audio data to the server.

[1512] Display device: Used to display the stress relief methods received from the server to the patient.

[1513] User

[1514] The user performs the following operations:

[1515] Record your sleep talking: Record your sleep talking through your device.

[1516] Check and execute stress relief methods: Check and execute the stress relief methods sent from the server.

[1517] Specific examples

[1518] Audio data collection and analysis

[1519] 1. The device uses a microphone to record the patient's sleep talking. For example, if a patient says at night, "Today I had a dream that I was being swept away by a big wave," the device will record that voice.

[1520] 2. The device sends the recorded audio data to the server in real time.

[1521] 3. The server converts the voice data into text data using Google Cloud Speech-to-Text. In this case, the voice data is converted into text data such as "Today I had a dream that I was swept away by a big wave."

[1522] 4. The server uses SpaCy to process the converted text data using natural language processing. For example, the phrase "swallowed by a big wave" is extracted.

[1523] 5. The server uses an NLP engine to infer the content of the dream. In this case, the content of the dream is inferred to be "being swept away by a big wave."

[1524] Identifying stress and suggesting ways to relieve it

[1525] 1. The server inputs the predicted dream content into IBM Watson's stress analysis algorithm. For example, a dream about being swept away by a big wave is input.

[1526] 2. The server uses the analysis results to identify the stress the patient is experiencing. In this case, it identifies "stress caused by anxiety or fear."

[1527] 3. The server generates stress relief methods using OpenAI GPT-4 based on the identified stress. For example, it generates "deep breathing techniques to relax."

[1528] 4. The server sends the generated stress relief method to the terminal.

[1529] 5. The terminal displays the stress relief method received from the server to the patient. For example, "deep breathing techniques to relax" is displayed.

[1530] 6. The user checks the displayed stress relief method and performs it, for example, by following the deep breathing procedure to relax.

[1531] Prompt Sentence Examples

[1532] "Analyze the patient's sleep talking, identify the stress from the content of the dream, and suggest an appropriate method for relieving stress. The content of the sleep talking was, 'Today I had a dream that I was swept away by a big wave.' The user's emotional state is 'anxious.'" The flow of the identification process in Example 3 will be described with reference to Figure 21.

[1533] Step 1:

[1534] Audio data collection

[1535] The device uses a microphone to record audio when the patient talks in their sleep.

[1536] Input: Patient's sleep talking (audio data)

[1537] Output: Recorded audio data

[1538] Specific operation: If the patient says at night, "Today I had a dream that I was being swept away by a big wave," the device will record their voice.

[1539] Step 2:

[1540] Sending audio data

[1541] The terminal transmits the recorded voice data to the server in real time.

[1542] Input: Recorded audio data

[1543] Output: Audio data sent to the server

[1544] Specific operation: The device sends the recorded audio data to a server via the Internet.

[1545] Step 3:

[1546] Converting audio data to text

[1547] The server converts the audio data into text data using Google Cloud Speech-to-Text.

[1548] Input: Audio data sent to the server

[1549] Output: Converted text data

[1550] Specific operation: The server converts the audio data "Today I had a dream that I was swept away by a big wave" into text data.

[1551] Step 4:

[1552] Text data analysis

[1553] The server performs natural language processing on the converted text data using SpaCy.

[1554] Input: Converted text data

[1555] Output: Parsed dream content

[1556] Specific operation: The server analyzes the text data "Today I had a dream that I was swallowed by a big wave" and extracts the content of the dream, "being swallowed by a big wave."

[1557] Step 5:

[1558] Identifying stress

[1559] The server inputs the estimated dream content into IBM Watson's stress analysis algorithm.

[1560] Input: Parsed dream content

[1561] Output: Identified stress

[1562] Specific operation: The server inputs the content of the dream about being "swept away by a big wave" into a stress analysis algorithm and identifies "stress caused by anxiety and fear."

[1563] Step 6:

[1564] Creating ways to relieve stress

[1565] Based on the identified stress, the server generates a stress relief method using OpenAI GPT-4.

[1566] Input: Identified stress

[1567] Output: Generated stress relief methods

[1568] Specific operation: The server generates "deep breathing techniques to relax" in response to "stress caused by anxiety or fear."

[1569] Step 7:

[1570] Stress relief notifications

[1571] The server transmits the generated stress relief method to the terminal.

[1572] Input: Generated stress relief methods

[1573] Output: Stress relief method sent to device

[1574] Specific operation: The server sends the generated "deep breathing method for relaxation" to the device.

[1575] Step 8:

[1576] Displaying ways to relieve stress

[1577] The terminal displays the stress relief method received from the server to the patient.

[1578] Input: Stress relief method sent from the server

[1579] Output: Stress relief strategies displayed to the patient

[1580] Specific operation: The device displays the received "deep breathing techniques for relaxation" to the patient.

[1581] Step 9:

[1582] Implementing stress relief methods

[1583] The user checks the displayed stress relief method and implements it.

[1584] Input: Stress relief methods displayed on the device

[1585] Output: Stress relief method performed

[1586] Specific action: The user follows the displayed deep breathing instructions to relax.

[1587] (Application example 3)

[1588] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1589] In conventional mental health treatment, it is difficult to accurately grasp a patient's emotional state and the causes of stress, making it difficult to propose effective stress relief methods. Furthermore, technology for analyzing sleep talking and dream content has not yet fully developed, limiting the means to discover a patient's underlying stress. Furthermore, there has been no system in brick-and-mortar counseling or mental health clinics that can comprehensively analyze a patient's emotional state and dream content and propose appropriate stress relief methods.

[1590] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1591] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for discovering potential stress a person has from the inferred dream content, means for suggesting ways to relieve the discovered stress, means for analyzing the person's emotional state using an emotion engine, means for suggesting ways to relieve stress by combining the analyzed emotional state and dream content, means for recording and analyzing sleep talking using a smart device, and means for use in brick-and-mortar counseling rooms and mental health clinics. This makes it possible to comprehensively analyze a patient's emotional state and dream content and suggest more accurate ways to relieve stress.

[1592] A "means for collecting sleep talking during sleep" is a device or method for recording an individual's sleep talking during sleep and collecting it as data.

[1593] "Generative AI" is an artificial intelligence technology that generates specific information based on collected data, and in particular uses deep learning to infer the content of dreams.

[1594] A "means for inferring the content of dreams" is a method or device for analyzing collected sleep talking data and inferring the specific content of the dreams an individual had from that content.

[1595] A "means for detecting potential stress" is a method or device for identifying potential stress an individual may be experiencing from the content of inferred dreams.

[1596] "Means for suggesting ways to relieve stress" refers to a method or device for suggesting appropriate ways to relieve stress that has been discovered.

[1597] The "emotion engine" is a technology for analyzing an individual's emotional state, and in particular, for quantifying and classifying changes and states of emotions.

[1598] The "means for analyzing emotional state" refers to a method or device for analyzing an individual's emotional state using an emotion engine and obtaining the results.

[1599] A "smart device" is a portable electronic device that can connect to the Internet and is used specifically to record and analyze sleep talking.

[1600] "Brick and mortar store" means a store that exists in a physical location, and specifically includes counseling rooms and mental health clinics.

[1601] The embodiments of the present invention will be described in detail below.

[1602] System Overview

[1603] This system collects sleep talking during sleep, uses generative AI to infer the content of dreams from the collected sleep talking, discovers potential stress in individuals from the inferred dream content, and suggests ways to relieve the discovered stress.Furthermore, it uses an emotion engine to analyze an individual's emotional state, and proposes stress relief methods by combining the analyzed emotional state with the content of dreams.

[1604] Hardware and software used

[1605] Smart devices: Use portable electronic devices that can connect to the internet, such as smartphones and head-mounted displays, to record and analyze sleep talking.

[1606] Generative AI: Using artificial intelligence techniques that use deep learning to infer dream content.

[1607] Emotion Engine: Uses technology to analyze an individual's emotional state.

[1608] Server: Uses a central processing unit to process the collected data and suggest ways to relieve stress.

[1609] Data processing and calculation

[1610] 1. Sleep-talking collection: The user's sleep-talking is recorded using the microphone on the smart device. The recorded audio data is sent to the server.

[1611] 2. Dream content inference: The server inputs the collected sleep talking data into a generative AI to infer the dream content. The generative AI uses a deep learning model to infer the dream content from the sleep talking text data.

[1612] 3. Emotional state analysis: The server uses the emotion engine to analyze the user's emotional state. The emotion engine quantifies and classifies the emotional state based on the content of the sleep talking and other input data.

[1613] 4. Suggesting ways to relieve stress: The server combines the inferred dream content with the analyzed emotional state and uses generative AI to suggest ways to relieve stress.

[1614] Specific examples

[1615] For example, if a user says in their sleep, "I had a scary dream," the emotion engine will interpret this as "fear." Based on this information, the generative AI will suggest ways to relieve stress, such as "take deep breaths to relax" or "listen to relaxing music."

[1616] Prompt Sentence Examples

[1617] User's sleep talking: "I had a scary dream"

[1618] Emotional state: "Fear"

[1619] Prompt for generative AI model: "The user feels they had a scary dream. What stress relief would you suggest to the user who is feeling scared?"

[1620] In this way, a system can be realized that can be used in brick-and-mortar counseling rooms and mental health clinics.

[1621] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1622] Step 1:

[1623] The user uses a smart device (smartphone or head-mounted display) to record their sleep talking while they are sleeping.

[1624] Input: User's sleep talking (audio data)

[1625] Output: Audio data of recorded sleep talking

[1626] Specific operation: Using the microphone on the smart device, the user's sleep talking is recorded and saved as audio data.

[1627] Step 2:

[1628] The device sends the recorded audio data of the sleep talking to a server.

[1629] Input: Recorded audio data of sleep talking

[1630] Output: Audio data of sleep talking sent to the server

[1631] Specific operation: The smart device uploads the recorded audio data to a server via the Internet.

[1632] Step 3:

[1633] The server uses voice recognition technology to convert the sleep talking audio data into text data.

[1634] Input: Sleep talking audio data

[1635] Output: Text data of sleep talking

[1636] Specific operation: The server uses speech recognition software (e.g., Google's speech recognition API) to convert the voice data into text data.

[1637] Step 4:

[1638] The server uses generative AI to infer the content of dreams from the text data of sleep talking.

[1639] Input: Text data of sleep talking

[1640] Output: Inferred dream content

[1641] Specific operation: The server inputs the text data of sleep talking into a generative AI (deep learning model) and infers the content of the dream.

[1642] Step 5:

[1643] The server uses an emotion engine to analyze the user's emotional state.

[1644] Input: Text data of sleep talking

[1645] Output: Parsed emotional state

[1646] Specific operation: The server inputs the text data of sleep talking into the emotion engine, which then quantifies and classifies the emotional state.

[1647] Step 6:

[1648] The server combines the inferred dream content with the analyzed emotional state and uses generative AI to suggest ways to relieve stress.

[1649] Input: Inferred dream content, analyzed emotional state

[1650] Output: Suggested stress relief method

[1651] Specific operation: The server inputs the dream content and emotional state into the generative AI, which then generates an appropriate stress relief method.

[1652] Step 7:

[1653] The server sends the suggested stress relief methods to the user's smart device.

[1654] Input: Suggested stress relief method

[1655] Output: Stress relief methods displayed on the user's smart device

[1656] Specific operation: The server sends the suggested stress relief methods to the user's smart device via the Internet and displays them.

[1657] In this way, users can receive stress relief methods based on their sleep talking and emotional state.

[1658] 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.

[1659] 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> ) and other generative AIs. Data generation models 58

[1660] , 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 the data format of voice data, text data, etc. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1661] Another example of generative AI is Gemini (internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1662] 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.

[1663] [Third embodiment]

[1664] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1665] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1666] 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).

[1667] 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.

[1668] 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.

[1669] 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).

[1670] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1671] 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.

[1672] 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.

[1673] 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.

[1674] 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.

[1675] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1676] "Example 1"

[1677] This embodiment of the present invention is composed of a microphone that collects sleep talking during sleep, a generative AI that analyzes the collected sleep talking and infers the content of dreams, and an advice generation module that discovers potential stress an individual is experiencing from the inferred dream content and suggests ways to alleviate it. Specifically, the microphone collects sleep talking during sleep, and the data is sent to the generative AI. The generative AI uses deep learning to infer the content of dreams from the sleep talking and sends the results to the advice generation module. The advice generation module discovers stress from the inferred dream content and suggests ways to alleviate it.

[1678] "Example 2"

[1679] In another aspect of the present invention, a generative AI uses deep learning to infer dream content. Specifically, the generative AI receives sleep talking data as input and uses a deep learning model to infer dream content. The deep learning model is trained using a large amount of sleep talking data and corresponding dream content data.

[1680] "Example 3"

[1681] In a further aspect of the present invention, the system can be used in the mental health treatment of psychiatric patients. Specifically, the system infers the content of dreams from the sleep talking of psychiatric patients, and identifies the stress the patient is experiencing from the content of those dreams. The system then suggests ways to relieve that stress, contributing to the improvement of the patient's mental health.

[1682] The processing flow of each embodiment will be described below.

[1683] "Example 1"

[1684] Step 1: A microphone collects your sleep talking while you sleep.

[1685] Step 2: The collected sleep talking data is sent to the generative AI.

[1686] Step 3: The generative AI uses deep learning to infer the content of the dream from the sleep talking.

[1687] Step 4: The estimated dream content is sent to the advice generation module.

[1688] Step 5: The advice generation module detects stress from the content of the inferred dream and

[1689] We propose a solution to this problem.

[1690] "Example 2"

[1691] Step 1: The generative AI receives sleep talking data as input.

[1692] Step 2: The generative AI uses a deep learning model to infer the content of the dream.

[1693] Step 3: The deep learning model is trained using a large amount of sleep talking data and corresponding dream content data.

[1694] "Example 3"

[1695] Step 1: Infer the content of a psychiatric patient's dreams from their sleep talking.

[1696] Step 2: Discover the stress the patient is experiencing from the content of the dream you have guessed.

[1697] Step 3: Suggest ways to relieve the stress discovered and contribute to improving the patient's mental health.

[1698] Example 1

[1699] Next, a description will be given of Example 1 of Form 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."

[1700] Conventional sleep talking analysis systems have insufficient collection and analysis of sleep talking, making it difficult to accurately identify potential stress in users and suggest ways to alleviate it. Furthermore, the accuracy of converting voice data into text and predicting dream content is low, resulting in a lack of reliability in the advice given to users. This has led to the issue of insufficient efforts to improve users' mental health.

[1701] 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.

[1702] In this invention, the server includes means for collecting sleep talking during sleep, means for saving the collected sleep talking as audio data, means for transmitting the saved audio data to the server, means for converting the audio data into text data, means for inferring the content of dreams from the converted text data using a generative AI model, means for discovering potential stress the individual is experiencing from the inferred dream content, and means for suggesting ways to relieve the discovered stress. This makes it possible to analyze the user's sleep talking with high precision, accurately discover potential stress, and reliably suggest ways to relieve it.

[1703] The "means for collecting sleep talking during sleep" refers to a device or mechanism for collecting sleep talking uttered by a user while sleeping with high sensitivity.

[1704] The "means for saving collected sleep talking as audio data" refers to a device or mechanism for recording and saving collected sleep talking as digital audio data.

[1705] The "means for transmitting the stored voice data to the server" refers to a device or mechanism for transmitting the stored voice data to the server via a network.

[1706] "Means for converting voice data into text data" refers to a device or mechanism for converting voice data into text information using voice recognition technology.

[1707] A "means for inferring dream content using a generative AI model" is a device or mechanism that uses a deep learning algorithm to analyze and infer dream content from text data.

[1708] A "means for discovering potential stress experienced by an individual" is a device or mechanism for identifying potential stress experienced by a user based on the content of the inferred dream.

[1709] The "means for proposing a method for relieving the discovered stress" is a device or mechanism for proposing to the user a method for relieving the identified stress.

[1710] MODE FOR CARRYING OUT THE INVENTION

[1711] This invention is a system that collects sleep talking during sleep, analyzes the data to infer the content of dreams, discovers potential stress that an individual is experiencing, and suggests ways to alleviate that stress. Specific embodiments of this system are described below.

[1712] Hardware and software used

[1713] 1. Microphone

[1714] To collect sleep talking, a highly sensitive microphone is used, which is designed to clearly record the user's sleep talking.

[1715] 2. Terminal

[1716] A device is used to store the collected voice data and send it to a server. This device has the function of storing voice data in real time and sending it to a server using a secure protocol (e.g., HTTPS).

[1717] 3. Server

[1718] The server converts the received audio data into text using speech recognition software (e.g., Google Cloud Speech-to-Text API), which is then passed to a generative AI model and used to infer the content of the dream.

[1719] 4. Generative AI Models

[1720] The generative AI model uses a deep learning algorithm to infer dream content from text data. Specifically, it analyzes the text data using natural language processing technology and generates dream scenarios.

[1721] 5. Advice Generation Module

[1722] The advice generation module detects potential stresses the user may be experiencing based on the content of the inferred dream and suggests ways to alleviate them, including feasible and effective methods for the user.

[1723] Specific examples

[1724] If a user says in their sleep, "My work isn't going well," the microphone connected to the device will collect this speech and save it as audio data. The device will then send the collected audio data to a server, where the HTTPS protocol is used to ensure data security.

[1725] The server converts the received voice data into text data using the Google Cloud Speech-to-Text API. For example, the generated text data is "My work is not going well." The server then passes the generated text data to the generative AI model. The generative AI model analyzes this text data and infers that the user is experiencing work-related stress.

[1726] The dream content predicted by the generative AI model is sent to an advice generation module via the server. The advice generation module determines that the user is experiencing work-related stress and generates advice for the user, such as "We suggest that you set aside time for relaxation every day to reduce work stress." This suggestion is sent to the device via the server and notified to the user.

[1727] Prompt Sentence Examples

[1728] "If a user says in their sleep, 'My work isn't going well,' the generative AI model will analyze this sleep talking and infer that the user is experiencing work-related stress. What advice will the advice generation module provide to the user?"

[1729] In this way, the system analyzes the user's sleep talking, discovers potential stress, and suggests ways to relieve it.

[1730] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1731] Step 1:

[1732] While the user is sleeping, a microphone connected to the device collects the user's sleep talking. The microphone is designed to be highly sensitive and to clearly record the user's sleep talking. The input is the user's sleep talking audio, and the output is the collected audio data.

[1733] Step 2:

[1734] The device saves the collected voice data. The saved voice data is temporarily stored in the device for further processing. The input is the collected voice data, and the output is the saved voice data.

[1735] Step 3:

[1736] The device sends the stored audio data to the server using a secure protocol (e.g., HTTPS) to ensure data security. The input is the stored audio data, and the output is the audio data sent to the server.

[1737] Step 4:

[1738] The server converts the received voice data into text data using speech recognition software (e.g., Google Cloud Speech-to-Text API). The input is the voice data, and the output is the converted text data. Specifically, the speech recognition algorithm analyzes the voice data and generates the corresponding text.

[1739] Step 5:

[1740] The server passes the converted text data to the generative AI model, which uses a deep learning algorithm to infer the dream content from the text data. The input is text data, and the output is the inferred dream content. Specifically, it uses natural language processing technology to analyze the text data and generate a dream scenario.

[1741] Step 6:

[1742] The dream content inferred by the generative AI model is sent via the server to the advice generation module, which then discovers the user's potential stress based on the inferred dream content. The input is the inferred dream content, and the output is the discovered stress.

[1743] Step 7:

[1744] The advice generation module proposes ways to relieve the detected stress. These proposals include feasible and effective methods for the user. The input is the detected stress, and the output is the proposed ways to relieve it. Specific actions include relaxation techniques and ways to improve lifestyle habits.

[1745] Step 8:

[1746] The proposed solution is sent to the device via the server and notified to the user. The input is the proposed solution, and the output is a notification to the user. Specifically, the device displays a notification so that the user can confirm the proposed solution.

[1747] (Application example 1)

[1748] Next, a description will be given of Application Example 1 of Form 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."

[1749] In modern society, individuals experience a wide range of potential stresses, which can affect not only their mental health but also security risks. However, while conventional systems can infer the content of dreams from sleep talking and detect potential stress, they are unable to suggest appropriate security measures in response to that stress. This creates a need for systems that can simultaneously support both the user's mental health and security.

[1750] 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.

[1751] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for detecting potential stress in an individual from the inferred dream content, means for suggesting ways to relieve the detected stress, and means for recommending appropriate security measures according to the stress level, thereby supporting the user's mental health and reducing security risks.

[1752] "Somnitalking" refers to speech that is involuntarily uttered by an individual while they are asleep.

[1753] "Means for collecting" refers to a device or method for acquiring audio data.

[1754] "Generative AI" is a system that uses artificial intelligence technology to generate new information from data.

[1755] A "means for inferring the content of dreams" is a method or device for analyzing collected data to infer the content of dreams.

[1756] A "means for detecting potential stress" is a method or device for identifying the stress an individual is experiencing from the inferred content of dreams.

[1757] A "means for suggesting ways to relieve stress" is a method or device that suggests appropriate ways to deal with discovered stress.

[1758] The "means for recommending security measures" is a method or device that suggests appropriate security measures to a user depending on the user's stress level.

[1759] "Deep learning" is a technology that uses multi-layered neural networks to analyze data and learn patterns and features.

[1760] A "mentally ill person" is an individual who has a mental illness.

[1761] "Mental health care" is a medical procedure that evaluates mental health status and provides treatment and support.

[1762] The following system configuration will be described as an embodiment of the present invention.

[1763] System Configuration

[1764] The system is centered around the "Sleep Security Assistant" application installed on the user's smartphone. The system includes the following main components:

[1765] 1. A method for collecting sleep talking

[1766] The smartphone microphone is used to collect the sleep talking of the user. The audio data is converted into text data using a speech recognition library (e.g., speech_recognition library).

[1767] 2. A way to guess the content of dreams

[1768] The collected text data of sleep talking is sent to a generative AI model (e.g., the rinna / japanese-gpt2-medium model using the transformers library) to infer the content of dreams. This generative AI model uses deep learning to analyze the content of dreams from the text data.

[1769] 3. Ways to detect underlying stress

[1770] The inferred dream content is used to discover potential stress levels of individuals, assess stress levels, and understand the user's mental health.

[1771] 4. Suggesting ways to relieve stress

[1772] The system suggests appropriate ways to relieve stress, such as deep breathing to relax or recommending activities that will help relieve stress.

[1773] 5. Means of recommending security measures

[1774] The system suggests appropriate security measures to users based on their stress level. For example, if stress increases, the system suggests changing passwords and recommends setting up two-factor authentication.

[1775] Processing flow

[1776] 1. Audio collection and recognition

[1777] The sleep talking uttered by the user while sleeping is collected using the smartphone's microphone and converted into text data using a voice recognition library.

[1778] 2. Guessing the content of the dream

[1779] The collected text data is input into a generative AI model to infer the content of dreams. The generative AI model generates dream content using the following prompt sentences:

[1780] Prompt: "User says in their sleep: 'Work isn't going well.' Infer the content of their dream from this sleep rambling."

[1781] 3. Identifying and assessing stress

[1782] The content of the dream is inferred to identify the individual's potential stress and assess its level.

[1783] 4. Generating Advice

[1784] The system will suggest appropriate ways to relieve the stress detected, such as "Take a deep breath to relax."

[1785] 5. Proposal of security measures

[1786] The system suggests appropriate security measures to users depending on their stress level. For example, if stress increases, the system will recommend specific measures such as "change your password" or "enable two-factor authentication."

[1787] In this way, it is possible to support the mental health of users and reduce security risks.

[1788] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1789] Step 1:

[1790] Sleep talking uttered by a user while sleeping is collected.

[1791] Input: Audio data from the smartphone microphone.

[1792] Specific operation: The device (smartphone) uses a microphone to collect the user's sleep talking in real time. The collected voice data is converted into text data using a speech recognition library (e.g., speech_recognition library).

[1793] Output: Sleep talking as text data.

[1794] Step 2:

[1795] The collected text data of sleep talking is input into a generative AI model to infer the content of the dream.

[1796] Input: Text data obtained in step 1.

[1797] Specific operation: The server inputs text data into a generative AI model (e.g., the rinna / japanese-gpt2-medium model using the transformers library) to infer the content of the dream. The generative AI model generates the content of the dream using the following prompt sentence:

[1798] Prompt: "User says in their sleep: 'Work isn't going well.' Infer the content of their dream from this sleep rambling."

[1799] Output: Inferred dream content.

[1800] Step 3:

[1801] The content of the dream is inferred to identify the individual's potential stress and assess its level.

[1802] Input: Dream content obtained in step 2.

[1803] How it works: The server analyzes the inferred dream content to discover the individual's potential stress. It uses a pre-trained stress assessment model to assess the stress level.

[1804] Output: Stress level assessment results.

[1805] Step 4:

[1806] We will suggest appropriate ways to relieve any stress that is discovered.

[1807] Input: Stress level assessment results obtained in step 3.

[1808] What it does: The server suggests ways to relieve stress based on your stress level, such as taking deep breaths to relax or recommending activities that will help you relieve stress.

[1809] Output: Suggestions for stress relief.

[1810] Step 5:

[1811] Appropriate security measures are suggested to users based on their stress level.

[1812] Input: Stress level assessment results obtained in step 3.

[1813] Specific behavior: When the stress level increases, the server suggests appropriate security measures to the user, such as suggesting a password change or setting up two-factor authentication.

[1814] Output: Suggested security measures.

[1815] Example 2

[1816] Next, a description will be given of Example 2 of Form 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."

[1817] In conventional mental health treatment, it has been difficult to accurately grasp the potential stress that an individual is experiencing. In addition, there has been a lack of methods for analyzing the content of dreams, making it impossible to understand an individual's psychological state in detail. This has made it difficult to suggest appropriate stress relief methods.

[1818] 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.

[1819] In this invention, the server includes a means for collecting voice data during sleep, a means for converting the collected voice data into text data using voice recognition software, and a means for inputting the converted text data into a deep learning model to infer the content of dreams. This makes it possible to accurately infer the content of dreams from an individual's sleep talking, discover potential stress, and suggest appropriate stress relief methods.

[1820] "Audio data" refers to data that is a digital recording of the sounds made by the user while they are sleeping.

[1821] "Speech recognition software" is a program for converting voice data into text data.

[1822] "Text data" is text information converted by voice recognition software.

[1823] A "deep learning model" is an artificial intelligence model that is trained using large amounts of data to extract specific patterns and features from input data.

[1824] "Dream content" refers to the specific scenario or situation of the dream the user is having while sleeping.

[1825] "Latent stress" is stress that the user is not aware of but that has a psychological or physiological effect on the user.

[1826] "Stress relief methods" are specific measures or suggestions for reducing or eliminating the stress that has been identified.

[1827] "Mental health care" is a medical procedure that assesses mental health and provides treatment and support.

[1828] MODE FOR CARRYING OUT THE INVENTION

[1829] The present invention is a system for analyzing sleep talking uttered by a user while sleeping and predicting the content of the dream. A specific embodiment of this system will be described below.

[1830] 1. Program Generation

[1831] The server generates a generative AI program that uses deep learning to predict the content of dreams. Specifically, it builds a model using deep learning frameworks such as TensorFlow and PyTorch.

[1832] 2. Program processing explanation

[1833] The server receives the user's sleep talking data as input. This sleep talking data is converted into text data using speech recognition software (e.g., Google Speech-to-Text API). The server then inputs this text data into a deep learning model. This deep learning model has been trained using a large amount of sleep talking data and corresponding dream content data.

[1834] The deep learning model analyzes the input sleep talking data and infers the content of the dream. The inferred results are fed back to the user. Specifically, the server sends the inferred dream content in text format to the user's device.

[1835] 3. Specific Examples

[1836] For example, if a user says in their sleep, "I'm swimming in the ocean," the server converts this sleep-talking data into text using speech recognition software, and inputs the text data "I'm swimming in the ocean" into the deep learning model. The deep learning model analyzes this text data and infers that "the user is dreaming about swimming in the ocean." This inference is sent to the user's device in text format.

[1837] 4. Examples of prompts

[1838] An example of a prompt might be:

[1839] "If a user says in their sleep, 'I'm swimming in the ocean,' how would a generative AI infer the content of their dream?"

[1840] By inputting this prompt into the generative AI model, it is possible to provide a detailed explanation of the system's processing.

[1841] In this way, the system analyzes the user's sleep talking data and infers the content of their dreams, which allows for a detailed understanding of an individual's psychological state, the detection of potential stress, and the suggestion of appropriate stress relief methods.

[1842] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1843] Step 1:

[1844] The user uses a device (e.g., a smartphone or dedicated recording device) to record their sleep talking. The device collects the user's sleep talking as audio data. This audio data is then sent to a server.

[1845] Input: User's sleep talking (audio data)

[1846] Output: Audio data sent to the server

[1847] Step 2:

[1848] The server converts the received voice data into text data using voice recognition software (e.g., Google Speech-to-Text API). Specifically, the server sends the voice data to the API and obtains the returned text data.

[1849] Input: Audio data

[1850] Output: Text data

[1851] Step 3:

[1852] The server inputs the acquired text data into a deep learning model, which is built using frameworks such as TensorFlow and PyTorch and trained using a large amount of sleep talking data and corresponding dream content data.

[1853] Input: Text data

[1854] Output: Dream content guess result

[1855] Step 4:

[1856] The server uses a deep learning model to analyze the text data and predict the content of the dream. Specifically, the model analyzes patterns in the text data and compares them with past data to predict the most likely dream content.

[1857] Input: Text data

[1858] Output: Inferred dream content

[1859] Step 5:

[1860] The server sends the predicted dream content in text format to the user's device, where the user can check the results.

[1861] Input: Inferred dream content

[1862] Output: The text message sent to the user's device.

[1863] For example, if a user says in their sleep, "I'm swimming in the ocean," the following happens:

[1864] 1. Step 1: The user records their sleep talking using a recording app on their smartphone. The device then sends the recorded audio data to the server.

[1865] 2. Step 2: The server sends the received voice data to the Google Speech-to-Text API, which converts the voice data into text data such as "swimming in the ocean" and returns it to the server.

[1866] 3. Step 3: The server inputs the text data "swimming in the ocean" into the deep learning model. The model analyzes this text data.

[1867] 4. Step 4: The deep learning model analyzes the text data and infers that "the user is dreaming about swimming in the ocean."

[1868] 5. Step 5: The server sends the guess result to the user's device as a text message saying, "You were dreaming about swimming in the ocean." The user can check the notification on their device to find out the content of their dream.

[1869] (Application example 2)

[1870] Next, a description will be given of Application Example 2 of Form 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."

[1871] Conventional dream analysis systems focus on inferring the content of dreams, but lack the ability to link the results of these inferences to the user's interests and provide specific content. This means that users are unable to not only understand the content of their dreams, but also to enjoy useful information and entertainment based on that content.

[1872] 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.

[1873] In this invention, the server includes a means for collecting sleep talking during sleep, a means for inferring dream content from the collected sleep talking using a generative AI, a means for delivering customized content relevant to the individual based on the inferred dream content, and a means for analyzing the user's interests and concerns based on the delivered content. This allows the user to not only understand the content of their dreams, but also enjoy customized content based on that content.

[1874] A "means for collecting sleep talking" is a device or software that records the sounds a user makes while sleeping and saves the data.

[1875] "Means for inferring dream content from collected sleep talking using generative AI" refers to algorithms and systems that analyze recorded sleep talking data and use deep learning models to infer dream content.

[1876] "Means for delivering customized content relevant to individuals based on the content of inferred dreams" refers to a system or service for providing content such as music, video, or articles that are appropriate for users based on the content of their inferred dreams.

[1877] "Means for analyzing user interests and concerns based on delivered content" refers to algorithms or systems for analyzing users' reactions and behavior to received content and identifying users' interests and concerns.

[1878] The system for implementing this invention mainly consists of three elements: a server, a terminal, and a user. The function of each element will be explained below.

[1879] server

[1880] The server has the following main functions:

[1881] 1. Collection of sleep talking data: The server receives and stores the sleep talking data sent from the device. The sleep talking data is a recording of the user's speech while sleeping.

[1882] 2. Dream content prediction: The server inputs the stored sleep talking data into a generative AI model and uses deep learning to predict the dream content. This generative AI model is trained using a large amount of sleep talking data and corresponding dream content data.

[1883] 3. Delivery of customized content: Based on the content of the predicted dream, the server selects customized content (music, video, articles, etc.) relevant to the user and delivers it to the device.

[1884] 4. Analysis of user interests: The server analyzes the user's reactions and behavior to the delivered content to identify the user's interests.

[1885] Terminal

[1886] The terminal has the following main functions:

[1887] 1. Recording sleep talking data: The device is equipped with a microphone to record the user's sleep talking. The recorded sleep talking data is sent to the server.

[1888] 2. Receiving and displaying content: The terminal receives the customized content delivered from the server and displays it to the user.

[1889] User

[1890] Users interact with the system as follows:

[1891] 1. Providing sleep talking data: The user uses the device to record sleep talking while sleeping.

[1892] 2. Content Usage: The user uses the customized content delivered to the device, for example, listening to music or watching videos based on the content of the predicted dream.

[1893] Hardware and software used

[1894] Hardware: Devices such as smartphones and tablets, and servers

[1895] Software: TensorFlow (running deep learning models), SoundDevice (collecting audio data), Librosa (preprocessing audio data)

[1896] Specific examples

[1897] For example, if a user talks in their sleep with the keyword "adventure," the server analyzes the sleep-talking data and infers that the dream content is adventure-related.The server then delivers customized content, such as an adventure movie trailer or an introductory article for an adventure novel, to the device.The user can then enjoy this content on their device.

[1898] Prompt Sentence Examples

[1899] "Analyze last night's sleep talking data and infer the content of your dream. Based on the inferred content of your dream, suggest related content."

[1900] The above is an embodiment of the present invention.

[1901] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1902] Step 1:

[1903] Recording of sleep talking data

[1904] The device records the user's sleep speech. Specifically, it uses a built-in microphone to collect audio data at regular intervals. The input is the user's sleep talking, and the output is the recorded audio data.

[1905] Step 2:

[1906] Sending sleep talking data

[1907] The device sends the recorded sleep talking data to a server. Specifically, the device uploads the audio data to the server via an internet connection. The input is the recorded audio data, and the output is the audio data sent to the server.

[1908] Step 3:

[1909] Saving sleep talking data

[1910] The server receives the sleep talking data sent from the device and stores it in a database. The input is the voice data sent from the device, and the output is the stored voice data.

[1911] Step 4:

[1912] Audio data preprocessing

[1913] The server preprocesses the stored sleep talking data. Specifically, it converts the audio data into MFCC (Mel-Frequency Cepstrum Coefficients) using the Librosa library. The input is the stored audio data, and the output is the preprocessed audio data.

[1914] Step 5:

[1915] Guessing the content of the dream

[1916] The server inputs the preprocessed audio data into a generative AI model to infer the content of the dream. Specifically, it uses TensorFlow to run a deep learning model to infer the content of the dream. The input is the preprocessed audio data, and the output is the inferred dream content.

[1917] Step 6:

[1918] Customized content selection

[1919] The server selects customized content relevant to the user based on the inferred dream content. Specifically, it searches a database related to the inferred dream content for appropriate content (music, video, articles, etc.). The input is the inferred dream content, and the output is the selected content.

[1920] Step 7:

[1921] Content Delivery

[1922] The server delivers the selected customized content to the terminal, specifically by transmitting content data to the terminal via an Internet connection, with the input being the selected content and the output being the content delivered to the terminal.

[1923] Step 8:

[1924] View content

[1925] The terminal receives the customized content delivered from the server and displays it to the user. Specifically, the content is played back using the terminal's display and speakers. The input is the content delivered from the server, and the output is the content displayed to the user.

[1926] Step 9:

[1927] Collecting user responses

[1928] The device records the user's actions and reactions to the delivered content. Specifically, it collects data such as the user's taps, clicks, and playback time. The input is the user's actions and reactions, and the output is the collected reaction data.

[1929] Step 10:

[1930] Analysis of user interests and concerns

[1931] The server analyzes the user's reaction data sent from the device and identifies the user's interests. Specifically, the collected reaction data is input into an analysis algorithm to infer the user's interests. The input is the collected reaction data, and the output is the inferred user's interests.

[1932] Example 3

[1933] Next, a third embodiment of the third embodiment will be described. 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."

[1934] Traditional mental health treatments often rely on patients' subjective reports, making it difficult to accurately identify potential stressors. Furthermore, there are limited means of discovering stressors that patients are not aware of, making it difficult to propose effective stress relief methods. This can potentially delay the improvement of mental health in psychiatric patients.

[1935] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1936] In this invention, the server includes means for collecting voice data during sleep, means for transmitting the collected voice data to the server, means for converting the voice data into text data, means for inferring dream content from the text data using a generative AI model, means for discovering potential stress the individual is experiencing from the inferred dream content, and means for suggesting ways to relieve the discovered stress. This makes it possible to infer the content of dreams from a patient's sleep talking, identify potential stressors, and suggest appropriate ways to relieve stress.

[1937] "Sound data during sleep" refers to data recorded from sounds made by a patient while sleeping.

[1938] "Means of collection" refers to devices or software for recording audio data, including, for example, smartphones and dedicated recording devices.

[1939] "Means for sending to a server" refers to the functions and protocols for uploading collected voice data to a server via a network.

[1940] "Means for converting voice data into text data" refers to a process for converting voice data into text information using voice recognition technology, and includes, for example, voice recognition software.

[1941] A "generative AI model" refers to an artificial intelligence model that uses deep learning to perform a specific task, including, for example, a natural language processing model.

[1942] "Method of inferring dream content" refers to the process of using a generative AI model to analyze and infer dream content from text data.

[1943] "Methods for detecting potential stress" refers to the process of analyzing the content of inferred dreams to identify potential stressors for an individual.

[1944] "Means for suggesting ways to relieve stress" refers to the process of suggesting appropriate ways to relieve stress based on the stress factors identified.

[1945] This invention is a system that can be used in the mental health treatment of psychiatric patients, and it analyzes the patient's sleep talking to infer the content of their dreams, identify potential stress factors, and propose appropriate stress relief methods. A specific embodiment of this system will be described below.

[1946] 1. System Configuration

[1947] The system is mainly composed of the following hardware and software.

[1948] Device: A device used to record the patient's sleep talking. This includes smartphones and dedicated recording devices.

[1949] Server: A computer system that receives and analyzes audio data.

[1950] Speech recognition software: Software for converting voice data into text data. Examples include the Google Cloud Speech-to-Text API.

[1951] Generative AI models: Artificial intelligence models for inferring dream content from text data. Examples include OpenAI's GPT-4.

[1952] 2. System Operation

[1953] The operation of the system is as follows.

[1954] Audio data collection

[1955] The device records the patient's sleep talking while they sleep. For example, by activating a recording app on a smartphone and placing it next to the pillow at night, high-quality recordings of sleep talking can be made.

[1956] Sending audio data

[1957] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using Wi-Fi or mobile data communication. When the upload is complete, the device notifies the user.

[1958] Analysis of audio data

[1959] The server uses speech recognition software to convert the received voice data into text data, which is then used in the next step.

[1960] Guessing the content of the dream

[1961] The server uses a generative AI model to infer dream content from text data, specifically by inputting the text data as prompts and generating dream content.

[1962] The discovery of stress

[1963] The server analyzes the content of the predicted dream and identifies potential stressors the patient may be experiencing. For example, if the dream contains content about being chased, it will determine this as a sign of stress.

[1964] Suggestions for stress relief

[1965] The server then proposes appropriate stress relief methods based on the identified stress factors. The proposals are sent to the user's device. For example, it suggests relaxation techniques or counseling.

[1966] 3. Specific Examples

[1967] Users use their smartphones to record their sleep talking at night. The recorded audio data is sent to a server and converted into text using speech recognition software. A generative AI model analyzes the text data and infers the content of the dream. For example, if the dream contains content about being chased, the server will interpret this as a sign of stress and suggest relaxation techniques or counseling.

[1968] 4. Examples of prompts

[1969] "You must analyze the patient's sleep talking and infer the content of their dream. Then, you must identify the stressors that the patient is experiencing based on the content of the dream and suggest appropriate ways to relieve stress."

[1970] In this way, a system is realized that contributes to improving the mental health of psychiatric patients through cooperation between the server, the terminal, and the user. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1971] Step 1:

[1972] Audio data collection

[1973] The device records the patient's sleep talking. Specifically, it uses a smartphone or a dedicated recording device and automatically starts recording at night. The input is the patient's sleep talking, and the output is the recorded audio file. When recording is finished, the audio file is temporarily saved.

[1974] Step 2:

[1975] Sending audio data

[1976] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using Wi-Fi or mobile data communication. The input is the recorded audio file, and the output is the audio data sent to the server. When the upload is complete, the device notifies the user.

[1977] Step 3:

[1978] Analysis of audio data

[1979] The server converts the received voice data into text data using voice recognition software. Specifically, it uses Google Cloud Speech-to-Text API or similar. The input is the voice data sent to the server, and the output is the converted text data. The conversion results are stored in a database.

[1980] Step 4:

[1981] Guessing the content of the dream

[1982] The server uses a generative AI model to infer dream content from text data. Specifically, the text data is input as a prompt, and the generative AI model (e.g., OpenAI's GPT-4) is used to generate dream content. The input is the converted text data, and the output is the inferred dream content. The generated dream content is stored in a database.

[1983] Step 5:

[1984] The discovery of stress

[1985] The server analyzes the estimated dream content and identifies potential stressors for the patient. Specifically, it applies psychological algorithms to analyze the frequency of specific keywords and phrases. The input is the estimated dream content, and the output is the identified stressors.

[1986] Step 6:

[1987] Suggestions for stress relief

[1988] The server proposes appropriate stress relief methods based on the identified stress factors. Specifically, it generates suggestions for relaxation techniques and counseling. The input is the identified stress factors, and the output is the proposed stress relief methods. The suggestions are sent to the user's device as push notifications.

[1989] (Application example 3)

[1990] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1991] In modern society, managing the mental health of psychiatric patients is an important issue. However, conventional methods make it difficult to accurately grasp a patient's underlying stress and propose appropriate ways to relieve it. In particular, technology for inferring dream content from sleep talking and detecting stress is still in its infancy, and a system for realizing effective mental health treatment is needed.

[1992] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1993] In this invention, the server includes means for collecting sleep talking during sleep, means for inferring dream content from the collected sleep talking using a generative AI, means for detecting potential stress in an individual from the inferred dream content, means for suggesting ways to relieve the detected stress, means for notifying the user of the suggested stress relief methods, and means for recording the sleep talking, inferring dream content, detecting stress, and suggesting ways to relieve it on a smartphone. This makes it possible to effectively manage the mental health of psychiatric patients, and to detect potential stress early and suggest appropriate ways to relieve it.

[1994] A "means for collecting sleep talking" is a device or software that records the sounds a user makes while sleeping and saves them as data.

[1995] "Generative AI" is a system that uses artificial intelligence technology to generate new information from input data, and in particular, uses deep learning to infer the content of dreams.

[1996] A "means for inferring the content of dreams" is an algorithm or system that analyzes collected sleep talking data and infers the content of the dream the user is having.

[1997] A "means for detecting potential stress in an individual" is a method or system for identifying potential stress factors that a user may be experiencing from the content of their inferred dreams.

[1998] The "means for suggesting ways to relieve stress" is a system that suggests ways for users to relieve stress that they can implement in response to discovered stress factors.

[1999] A "means for notifying the user" is a communication means or interface for informing the user of the proposed stress relief method.

[2000] The "means to be executed on the smartphone" refers to software and hardware for executing the above-mentioned means as an application on the smartphone.

[2001] The following describes an embodiment of the present invention. The system is implemented as an application that runs on a user's smartphone. The main functions of the system and how they are implemented are described below.

[2002] 1. Collection of sleep talking

[2003] The app uses the smartphone's microphone to collect sleep talking while the user is asleep. The app automatically starts recording when the user talks in their sleep and saves the audio data.

[2004] 2. Guessing the content of the dream

[2005] The collected sleep talking data is input into a generative AI model (e.g., GPT-3). The generative AI model uses deep learning to infer the content of dreams from sleep talking. The following prompt sentences are used in this process:

[2006] Example prompt sentence:

[2007] "Please guess the content of the dream from the following sleep talking: A scary dog ​​is chasing me."

[2008] 3. Discovering stress

[2009] We again use a generative AI model to discover potential stressors from the inferred dream content. In this process, we use the following prompt:

[2010] Example prompt sentence:

[2011] "Identify the stress in the following dream and suggest ways to relieve it: The user dreams of being chased by a dog."

[2012] 4. Suggestions for ways to relieve stress

[2013] For each stress factor detected, the generative AI model will suggest an appropriate stress relief method, which will then be notified to the user.

[2014] 5. Notice to Users

[2015] The proposed stress relief methods are notified to users using the smartphone's notification function, allowing users to implement the stress relief methods at the ap...

Claims

1. A means for acquiring audio data of sleep talking uttered by a user while sleeping; means for converting the voice data into text data; A means for generating a dream scenario using a generative AI model based on the text data; A means for detecting potential stress experienced by the user based on the generated dream scenario; A means for suggesting ways to relieve the stress found; A system including:

2. A means for acquiring audio data of sleep talking uttered by a user while sleeping; A means for inferring the content of a dream using a generative AI model based on the voice data; A means for selecting content based on the inferred dream content; means for delivering selected content to said user; A system including:

3. A means for acquiring audio data of sleep talking uttered by a user while sleeping; A means for analyzing the user's emotions based on the user's voice, facial expression, or behavior; A means for predicting the content of a dream using a generative AI model based on the analysis results of the voice data and the user's emotions; A means for detecting potential stress of the user based on the content of the predicted dream; A means for suggesting ways to relieve the stress found; A system including:

Images