system

Abstract

We provide the system. [Solution] A means of receiving audio data, A means of converting received audio data into text data, A means of summarizing the converted text data, A means of outputting summarized text data, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In a modern business environment, multiple meetings are frequently held, but it is difficult to attend all of them. Also, when checking the content of a recorded meeting later, it often takes a huge amount of time. Conventional methods lack means to efficiently grasp the key points of a meeting in a short time, which leads to a decline in work efficiency. Therefore, an effective means that can quickly understand the key points of a meeting even when unable to attend the meeting is required.

Means for Solving the Problems

[0005] This invention provides a system that includes means for receiving audio data, means for converting the received audio data into text data, means for summarizing the converted text data, and means for outputting the summarized text data. This system enables the automatic transcription of meeting audio data into text and further summarization of that text, making it possible to grasp the key points of a meeting in a short amount of time. As a result, even if one cannot attend a meeting, important information can be quickly obtained, thereby improving work efficiency. Specifically, this problem is effectively solved by combining speech recognition technology, natural language processing technology, and summarization generation technology.

[0006] "Audio data" refers to data recorded in digital format from audio such as meetings.

[0007] "Means of receiving" refers to the technical elements for uploading audio data from a user to a server, which is usually transmitted over a network.

[0008] "Text data" refers to character information obtained by converting audio data using speech recognition technology.

[0009] "Means of converting to text data" refers to the technical elements of a speech recognition engine that analyzes audio data and converts it into corresponding text information.

[0010] A "summarization technique" is a technical element that extracts important information from converted text data and compiles it into a concise form.

[0011] "Output method" refers to the technical elements used to send summarized text data to the user, and typically involves methods such as email or database updates.

[0012] "Means of storage" refers to technical elements for temporarily or permanently storing received audio data in a storage device. [Brief explanation of the drawing]

[0013] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, when an emotion engine is combined. [Figure 14] This is a sequence diagram showing the processing flow of the data processing system in Application Example 2, which combines an emotion engine. [Modes for carrying out the invention]

[0014] An example of an embodiment of the system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0015] First, the terms used in the following description will be explained.

[0016] In the following embodiments, a numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0017] In the following embodiments, a numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0018] In the following embodiments, a numbered storage is one or more non-volatile storage devices that store various programs, various parameters, and the like. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, and the like.

[0019] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0020] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0021] [First Embodiment]

[0022] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0023] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0024] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0025] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0026] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0027] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0028] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0029] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0030] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0031] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0032] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0033] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0034] System Overview

[0035] This invention relates to a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user can upload audio data of a meeting to a server, the server analyzes the audio data, converts it into text data, and finally summarizes it before delivering it to the user.

[0036] Specific processing details of the program

[0037] 1. Receiving audio data

[0038] The user uploads the meeting audio data to the server. The server temporarily stores the received audio data. This process is initiated when the user selects and uploads the file.

[0039] 2. Text conversion of audio data

[0040] The server converts the received audio data into text data using a speech recognition engine. The speech recognition engine analyzes the audio waveform and generates corresponding character information.

[0041] 3. Text Data Summary

[0042] The server passes the text data to a natural language processing engine, which extracts and summarizes the important information. The natural language processing engine analyzes the meaning of the text and extracts key points.

[0043] 4. Distribution of summaries

[0044] The server sends the generated summary to the user. The user can receive this summary via email or other means.

[0045] Specific example

[0046] For example, if the following audio data was recorded during a meeting:

[0047] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0048] When this audio data is uploaded to the system, the system generates the following summary:

[0049] "The agenda included discussing the new product release, a new marketing strategy, expanding the target market, and confirming the date for the next meeting."

[0050] By reviewing the summarized content, users can quickly grasp the key points of a meeting. This system provides an efficient way to obtain important information even when unable to attend a meeting.

[0051] The following describes the processing flow.

[0052] Step 1:

[0053] The user uploads the meeting audio file to the server. The user selects the audio file using a web browser or dedicated application and clicks the upload button. This action sends the audio file to the server over the network.

[0054] Step 2:

[0055] The server saves the received audio files. Uploaded audio files are stored in a temporary storage area for subsequent processing. This saving process includes checking the size of the audio file and converting it to an appropriate format.

[0056] Step 3:

[0057] The server passes the stored audio files to the speech recognition engine, where they are converted into text data. Specifically, speech recognition software is used to analyze the audio files and convert the audio signals into corresponding strings of characters. This conversion process includes noise reduction and detection of speech boundaries.

[0058] Step 4:

[0059] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. The summary is compressed to the extent that the meaning of the original text is not lost.

[0060] Step 5:

[0061] The server distributes the generated summary to the user. The summary results are sent to the email address specified by the user, and the user can check the summary in the received email. It can also be displayed on the system dashboard or in a dedicated application.

[0062] (Example 1)

[0063] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0064] Conventional systems were time-consuming to analyze and summarize audio data, making it difficult for users to efficiently obtain information. Furthermore, insufficient temporary storage of audio data and verification of data integrity posed a risk of data reliability degradation. Therefore, there was a need for a means to quickly and accurately obtain important information from meetings and audio recordings.

[0065] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0066] In this invention, the server includes means for uploading audio data to a communication device, means for converting the audio data received by the communication device into text data using an analysis device, means for analyzing the converted text data using a natural language processing device to extract and summarize important information, and means for outputting the summarized text data via the communication device. This enables rapid analysis and accurate summarization of audio data, allowing users to efficiently obtain important information.

[0067] 1. "Audio data" refers to data recorded in digital format from audio information such as meetings and conversations.

[0068] 2. A "communication device" is a device that transmits and receives voice data and summarized text data, and is capable of uploading and downloading data via an internet connection.

[0069] 3. "Analysis device" refers to a device for converting audio data into text data, and includes a speech recognition engine and other analysis software.

[0070] 4. "Text data" refers to character information after it has been analyzed by a speech recognition engine, and is data that represents the content of speech data in text format.

[0071] 5. A "natural language processing device" is a device that analyzes text data, extracts important information, and generates a summary.

[0072] 6. "Summary text data" refers to text data in a concise format, generated by a natural language processing device, from which only the essential information has been extracted from the original text data.

[0073] 7. "Means of temporary storage" refers to means of temporarily storing received audio data in a storage device so that it can be used for subsequent processing.

[0074] 8. "Data integrity verification" is a verification method to confirm that the received data has been transmitted accurately from the sender without any changes.

[0075] Modes for carrying out the invention

[0076] System Overview

[0077] This invention relates to a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user uploads audio data of a meeting to a communication device, a server converts that audio data into text data using an analysis device, summarizes it using a natural language processing device, and delivers it to the user.

[0078] Specific processing details of the program

[0079] 1. Receiving audio data

[0080] Users upload meeting audio data to the server via a communication device. Users use a web application to select and upload audio files. Once the upload is complete, the server temporarily stores the audio data. At this time, a hash value is calculated to verify the integrity of the data.

[0081] 2. Text conversion of audio data

[0082] The server passes the stored audio data to the analysis device. This analysis device uses a speech recognition engine (e.g., Google® Speech-to-Text API). The speech recognition engine analyzes the waveform of the audio data and generates corresponding text data. The generated text data is received by the server and temporarily stored.

[0083] 3. Text Data Summary

[0084] The server passes the generated text data to a natural language processing unit (NLP). This NLP uses a generative AI model (e.g., OpenAI® GPT-4®). The NLP analyzes the text data, extracts important information, and generates a summary. This summary extracts key points from the original text data and is stored secondarily by the server.

[0085] 4. Distribution of summaries

[0086] The server delivers the generated summary text data to the user. Delivery methods include email and notifications via a dedicated portal. Users receive notifications and can click a link to view the summary.

[0087] Specific example

[0088] For example, if the following audio data is recorded during a meeting:

[0089] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0090] When this audio data is uploaded to the system, the system generates the following summary:

[0091] "The agenda included discussing the new product release, a new marketing strategy, expanding the target market, and confirming the date for the next meeting."

[0092] Example of a prompt

[0093] Examples of prompt statements are as follows:

[0094] Please summarize the following meeting audio transcript:

[0095] At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting.

[0096] By inputting this prompt into the generation AI model, an appropriate summary can be generated.

[0097] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0098] Step 1: Upload audio data

[0099] The user uploads audio data from their terminal to the communication device. As input, the user selects a conference audio file (e.g., MP3 or WAV format) and presses the upload button. As output, the audio file is sent to the communication device. Specifically, the user opens a file selection dialog through their browser, selects the file, and begins sending. The communication device then forwards the received audio file to the server.

[0100] Step 2: Receiving and temporarily storing audio data

[0101] The server receives audio data transmitted from the communication device and stores it in a temporary storage area. The input is audio data sent from the communication device. The output is that audio data stored in the server's temporary storage area. Specifically, the server calculates a hash value (e.g., SHA-256) to verify data integrity and records the file name and hash value in the temporary storage area.

[0102] Step 3: Convert speech to text

[0103] The server passes the audio data stored in the temporary storage area to the analysis device. The input is the stored audio file sent to the analysis device. The output is text data generated by a speech recognition engine (e.g., Google Speech-to-Text API). Specifically, the server sends the audio data as a request to the API and stores the returned text data in the temporary storage area.

[0104] Step 4: Natural language processing of text data

[0105] The server passes text data stored in the primary storage area to a natural language processing unit (e.g., a generative AI model). The generative AI model receives text data as input. The output is summarized text data with important information extracted. Specifically, the server sends text data as a prompt to the generative AI model and stores the generated summary in the secondary storage area.

[0106] Step 5: Distribution of the summary

[0107] The server delivers summarized text data stored in a secondary storage area to the user. The input is the summarized text data passed to the server's notification system. The output is an email notification or display on a web portal to the user. Specifically, the server delivers the summarized text according to the user's settings (e.g., email, push notification). The user receives the notification and clicks the link to view the summary.

[0108] (Application Example 1)

[0109] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0110] In autonomous vehicles, if a driver needs to record important notes or instructions by voice while driving, they must stop the vehicle to take notes. This interferes with driving and reduces efficiency. Furthermore, when reviewing the voice data later, extracting important information from the vast amount of data is time-consuming and hinders quick decision-making. To solve this, a system is needed that transcribes and summarizes voice data in real time, allowing drivers to quickly obtain the information they need.

[0111] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0112] In this invention, the server includes means for receiving voice data, means for converting the received voice data into text data, means for summarizing the converted text data, and means for outputting the summarized text data generated at a terminal installed in the vehicle. This makes it possible for drivers to give instructions or make notes by voice while driving, and for important information to be quickly obtained by converting and summarizing them into text in real time.

[0113] "Audio data" refers to information obtained by converting sound waves into digital signals via microphones or other acoustic sensors.

[0114] "Text data" refers to digital text obtained by analyzing audio data and converting it into corresponding textual information.

[0115] A "summary" is a text document that extracts important information from converted text data and presents it concisely.

[0116] A "terminal installed in a vehicle" refers to a computer or smart device installed inside a vehicle that performs voice recognition and displays and stores transcribed data.

[0117] A "user interface" is a screen, device, or software used by a system to exchange information with a user.

[0118] "Storage" means saving data to a storage medium or database so that it can be accessed later.

[0119] "Analysis" is the process of understanding the content of text data using natural language processing and algorithms, and extracting the necessary information.

[0120] "Reception" refers to acquiring audio data from an external source and incorporating it into the system.

[0121] "Conversion" refers to the process of reconstructing audio data into text data.

[0122] "Output" refers to displaying or sending summarized text data to the user.

[0123] This invention relates to a system that transcribes audio data into text in real time, summarizes the text, and displays it on a terminal inside an autonomous vehicle.

[0124] Hardware and software to be used

[0125] hardware

[0126] Devices installed in automobiles (smartphones and in-car smart devices)

[0127] Microphone for voice input

[0128] software

[0129] Google Cloud Speech-to-Text API

[0130] GPT-3(registered trademark).5

[0131] Python

[0132] Explanation of program processing

[0133] Receiving audio data

[0134] The user inputs voice memos and instructions via the microphone while driving. The device receives this voice data in real time.

[0135] Text conversion of audio data

[0136] The server uses the Google Cloud Speech-to-Text API to convert received audio data into text data. This API provides highly accurate speech recognition capabilities, analyzing the audio waveform to generate corresponding text information.

[0137] Summary of text data

[0138] The server summarizes the converted text data using GPT-3.5. GPT-3.5 is a natural language processing engine that analyzes the meaning of the text, extracts important information, and summarizes it concisely.

[0139] Output and save summary

[0140] The summarized text data is displayed on a terminal installed in the vehicle via the user interface. Additionally, this summarized data can be saved on the terminal as needed for later review by the user.

[0141] Specific example

[0142] For example, if a driver enters the following voice memo while driving:

[0143] "We departed at 10:00 AM and encountered traffic congestion on the way to our destination. I instructed everyone to take a detour. While driving, I received an important phone call and instructed them to prepare materials for the afternoon meeting."

[0144] This audio data is summarized by the system as follows:

[0145] "I instructed them to take a detour due to traffic congestion after departure. I also instructed them to prepare materials for the afternoon meeting via an important phone call."

[0146] Example of a prompt

[0147] Please summarize the following text: "We departed at 10:00 AM and encountered traffic congestion on the way to our destination. I instructed everyone to take a detour. I received an important phone call while driving and instructed everyone to prepare materials for the afternoon meeting."

[0148] This will allow users to efficiently record voice memos while driving and quickly retrieve important information.

[0149] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0150] Step 1:

[0151] While driving, the user inputs voice memos and instructions via the microphone. The input voice data is saved in real time to the terminal's storage device. Here, the user's voice is recorded as a digital signal.

[0152] Input: User's voice memos or instructions

[0153] Output: Digital audio data stored on the device

[0154] Step 2:

[0155] The terminal sends the stored audio data to the server. The server receives this audio data and stores it temporarily. Here, digital audio data is transferred.

[0156] Input: Digital audio data stored on the device

[0157] Output: Digital audio data stored on the server

[0158] Step 3:

[0159] The server uses the Google Cloud Speech-to-Text API to convert the received audio data into text data. The server sends the audio data to the API and receives and stores the returned text. In this process, the audio data is converted into text information.

[0160] Input: Digital audio data stored on the server

[0161] Output: Text data

[0162] Step 4:

[0163] The server passes the text data to GPT-3.5 for natural language processing. GPT-3.5 parses the text data based on the input prompt sentence, extracts important information, and summarizes it. The summary result is returned to the server. Here, the text data is compressed and converted into a shorter, more concise form of text.

[0164] Input: Text data and prompt text

[0165] Output: Summarized text data

[0166] Step 5:

[0167] The server sends the summarized text data to the terminal. The terminal displays the received summary through its user interface. The user can review the summary on the screen. Here, the final output is provided to the user.

[0168] Input: Summarized text data

[0169] Output: Summary data in a format that can be viewed by the user.

[0170] Step 6:

[0171] If necessary, the device saves the displayed summary data on its own. The user can then refer to and review this data later. Here, the summary data is saved for future reference.

[0172] Input: User-confirmed summary data

[0173] Output: Summary data saved on the device

[0174] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0175] System Overview

[0176] This invention combines a system that receives audio data, converts it into text data, and then summarizes it with an emotion engine that recognizes the user's emotions. Specifically, a user uploads audio data of a meeting to a server, the server analyzes the audio data and converts it into text data, analyzes the emotions using the emotion engine, and finally summarizes the text data and delivers it to the user.

[0177] Specific processing details of the program

[0178] 1. Receiving audio data

[0179] The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated application to select the audio file and click the upload button. This action sends the audio file to the server over the network.

[0180] 2. Saving audio data

[0181] The server saves the received audio files. Uploaded audio files are stored in a temporary storage area for use in subsequent processing.

[0182] 3. Converting audio data to text

[0183] The server passes the stored audio files to the speech recognition engine, which converts them into text data. The speech recognition engine analyzes the audio waveform and generates corresponding character information. This conversion process includes noise reduction and detection of speech boundaries.

[0184] 4. Emotion analysis

[0185] The server uses an emotion recognition engine to analyze the user's emotions from the audio data. It identifies and extracts information about the user's emotions (e.g., joy, anger, sadness) from the audio data and its text. This also allows for understanding the emotional tone of each statement made during a meeting.

[0186] 5. Text Data Summary

[0187] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. If necessary, the results of sentiment analysis are also reflected in the summary.

[0188] 6. Distribution of summaries

[0189] The server distributes the generated summary to the user. The summary results are sent to the email address specified by the user, and the user can check the summary in the received email. It can also be displayed on the system dashboard or in a dedicated application.

[0190] Specific example

[0191] For example, if the following audio data was recorded during a meeting:

[0192] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0193] When this audio data is uploaded to the system, the system processes it as follows:

[0194] 1. Convert audio data to text.

[0195] 2. Analyze the emotions expressed in each statement (for example, the joy felt when discussing the "release of a new product" and the tension felt when discussing the "review of the marketing strategy").

[0196] 3. Summarize the text, and reflect the emotional tone in the summary.

[0197] An example of a generated summary:

[0198] "There was a discussion about the new product release, and most opinions were optimistic. Following that, there was a discussion about reviewing the marketing strategy and expanding the target market, and some sense of urgency was evident. Finally, the date for the next meeting was confirmed."

[0199] By reviewing the emotional tone along with the summarized content, users can quickly grasp the key points of a meeting and their emotional context. This system not only provides an efficient way to obtain important information even when unable to attend a meeting, but also allows users to understand the atmosphere of the meeting and the emotional reactions of the participants.

[0200] The following describes the processing flow.

[0201] Step 1:

[0202] The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated app to select the audio file and clicks the upload button. This action sends the audio file to the server over the network.

[0203] Step 2:

[0204] The server saves the received audio files to a temporary storage area. The saving process also includes checking the format and size of the audio files and converting them to the appropriate format.

[0205] Step 3:

[0206] The server passes the stored audio files to the speech recognition engine, which converts them into text data. In this process, the speech recognition engine analyzes the audio waveform and sequentially generates corresponding text information. Pre-processing such as keyword recognition and noise reduction is also performed.

[0207] Step 4:

[0208] The server passes text data to an emotion recognition engine to analyze the user's emotions. The emotion recognition engine identifies the speaker's emotional state (e.g., joy, anger, sadness) from the text data and audio features, and adds the result to the text data. This emotion information is either added to individual sentences or grouped by segment.

[0209] Step 5:

[0210] The server passes text data, including sentiment data, to a natural language processing engine to generate a summary. The natural language processing engine extracts key keywords and phrases from the text, integrates that information with sentiment data, and creates a concise summary.

[0211] Step 6:

[0212] The server delivers the completed summary to the user. Delivery methods include sending to a specified email address and displaying it on a dedicated dashboard or app. The email contains the summary text, and important sentiment information is highlighted as appropriate.

[0213] (Example 2)

[0214] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0215] Traditional systems that convert audio data into text and summarize it have difficulty overlooking important information and understanding the emotional context of meeting participants. Furthermore, the lack of a means to reflect emotional information in the summary makes it difficult to grasp the overall atmosphere of the meeting.

[0216] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0217] In this invention, the server includes means for receiving audio data, means for storing the received audio data, means for converting the stored audio data into text data, means for analyzing the converted text data to extract emotional information, means for summarizing the text data taking the emotional information into consideration, and means for outputting the summarized text data. This enables accurate summarization that reflects emotional information.

[0218] "Audio data" refers to information recorded in digital or analog format.

[0219] "Text data" refers to information expressed as characters.

[0220] "Means of storage" refers to technology or equipment for temporarily or long-term storage of received data.

[0221] "Means of conversion" refers to a technology or device for converting data in one format to data in another format.

[0222] "Means of analysis" refers to techniques or devices for examining, evaluating, or breaking down data to extract specific information.

[0223] "Emotional information" refers to information that indicates the emotional tone or emotional state contained within the data.

[0224] "Means of summarization" refers to techniques or devices that extract important points from large amounts of data and convert them into a shorter format.

[0225] "Output method" refers to a technology or device that provides processed data in a format that can be read by the user.

[0226] "Natural language processing technology" refers to the technology that enables computers to understand and process natural language, which humans use on a daily basis.

[0227] A "prompt statement" is an input statement for a generative AI model, a text used to induce a specific output.

[0228] System Configuration

[0229] This invention combines a system that receives audio data, converts it to text data, and summarizes it with an emotion engine that recognizes the user's emotions. The user uploads meeting audio data to a server, the server analyzes the audio data, converts it to text data, analyzes the emotions, and finally summarizes the text data and delivers it to the user.

[0230] Hardware and software to be used

[0231] The system of this invention primarily uses the following hardware and software:

[0232] Server: Receives, stores, converts, analyzes, and summarizes audio data.

[0233] Web browser or dedicated application: A terminal for users to upload audio data.

[0234] Speech recognition engine: Uses the Google Cloud Speech-to-Text API to convert speech data into text data.

[0235] Emotion recognition engine: Uses IBM Watson® Tone Analyzer to extract emotional information from text data.

[0236] Summarizing text data using the natural language processing engine: OpenAI GPT-3.

[0237] Specific operation of the system

[0238] 1. The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated application to select the audio file and click the upload button. The audio file is sent to the server over the network using an HTTP POST request.

[0239] 2. The server saves the received audio files to a temporary storage area. For example, it stores them in cloud storage.

[0240] 3. The server sends the audio file to the Google Cloud Speech-to-Text API, where the audio data is converted into text data. The speech recognition engine analyzes the audio waveform and generates the corresponding text information.

[0241] 4. The server retrieves the converted text data and passes it to IBM Watson Tone Analyzer to analyze the emotional information. The analysis results are displayed numerically, for example, "Joy: 0.7, Anger: 0.1, Sadness: 0.2".

[0242] 5. The server sends the text data along with a prompt to OpenAI GPT-3 to generate a summary. An example of the prompt to be generated is "Please summarize the contents of the following meeting." + "Acquired text data".

[0243] 6. The server distributes the generated summary to the user. The summary results are sent via email to the email address specified by the user. Users can also view the summary results using the system dashboard or a dedicated application.

[0244] Specific example

[0245] For example, consider a case where the following audio data was recorded during a meeting:

[0246] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0247] When this audio data is uploaded to the system, it is processed as follows:

[0248] 1. The user uses a dedicated application to upload audio files to the server.

[0249] 2. The server saves the audio files to cloud storage (e.g., Amazon S3).

[0250] 3. The server sends the audio file to the Google Cloud Speech-to-Text API and retrieves the text data "We discussed the release of the new product...".

[0251] 4. The server passes the text data to IBM Watson Tone Analyzer and obtains emotion information: "Joy: 0.7, Anger: 0.1, Sadness: 0.2".

[0252] 5. The server sends the text data to OpenAI GPT-3 along with the prompt "Please summarize the contents of the following meeting." + "Acquired text data," generating the summary "There was a discussion about a new product release, and there were many optimistic opinions..."

[0253] 6. The server sends the generated summary to the user via email and also makes it available on the dashboard.

[0254] In this way, the present invention enables users to quickly grasp important information, including the overall emotional context of a meeting, through efficient analysis of audio data and summarization that reflects emotional information.

[0255] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0256] Step 1:

[0257] The user uploads the meeting audio data to the server. The input is the audio file selected by the user, and the output is the file sent to the server via an HTTP POST request. The user opens a web browser or dedicated application, selects the audio file, and clicks the upload button. This action sends the audio file to the server.

[0258] Step 2:

[0259] The server saves the received audio file to a temporary storage area. The input is the audio file sent by the user, and the output is the file saved in storage. When the server receives the audio file, it saves it to cloud storage (e.g., Amazon S3). This save operation makes it accessible in subsequent processing steps.

[0260] Step 3:

[0261] The server sends the stored audio file to the Google Cloud Speech-to-Text API and converts it into text data. The input is an audio file in cloud storage, and the output is text data obtained from the speech recognition engine. The server sends the audio file via the API, analyzes the audio waveform, and obtains the corresponding text information. Noise reduction and speech segmentation are also performed during this process.

[0262] Step 4:

[0263] The server sends text data to the IBM Watson Tone Analyzer for emotional analysis. The input is text data from the speech recognition engine, and the output is emotional information from the emotional recognition engine. The server passes the acquired text data to the emotional recognition engine to obtain emotional information (e.g., "Joy: 0.7, Anger: 0.1, Sadness: 0.2"). This information is then used in the subsequent summary.

[0264] Step 5:

[0265] The server sends text data along with a prompt to OpenAI GPT-3, which then generates a summary. The input consists of text data and a prompt, and the output is the summary obtained from the generating AI model. An example of a prompt is "Please summarize the contents of the following meeting." + "Acquired text data". The server then generates and saves the summary.

[0266] Step 6:

[0267] The server delivers the generated summary to the user. The input is the generated summary text, and the output is the summary result sent to the user. The server sends the summary result to an email address. Users can also view the summary result on the system dashboard or a dedicated application. Emails are sent using the SMTP protocol, and the dashboard displays the summary result securely using HTTPS.

[0268] (Application Example 2)

[0269] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0270] Systems that extract and summarize important information from audio data face the challenge of understanding the emotional context of meetings and discussions. Conventional technologies cannot extract emotional information from transcribed data, making it difficult to grasp the nuances and emotional shifts of conversations. To solve this problem, it is necessary to simultaneously extract emotional information from audio data and incorporate it into the summary.

[0271] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0272] In this invention, the server includes means for receiving audio data, means for converting the received audio data into text data, means for summarizing the converted text data, means for adding emotional information to the summarized text data, and means for outputting the summarized text data. This allows for the extraction of emotional information along with the summary of audio data, and by reflecting the results in the summary, it becomes possible to simultaneously understand the content of meetings and discussions and their emotional background.

[0273] "Voice data" refers to digital information that digitizes human voice, including conversations, instructions, and other audio files.

[0274] "Means" refers to the methods, apparatus, or techniques used to achieve a particular purpose.

[0275] "Text data" refers to data that is stored or displayed as character information, specifically information obtained by converting audio data into text.

[0276] A "summary" refers to a short and concise version of the content of a text or audio recording, extracting the most important information and making it more compact.

[0277] "Emotional information" refers to data on emotional states analyzed from audio or text, such as information representing emotions like joy, anger, or sadness.

[0278] "Output" refers to the act of displaying, printing, or sending processed data to another system.

[0279] "Receiving" refers to the process of receiving and processing data or information from an external source.

[0280] "Storage" refers to the act of keeping recorded data without losing it, so that it can be retrieved and used later.

[0281] "Analysis" refers to the process of breaking down data and understanding its structure and meaning.

[0282] "Extraction" refers to the act of extracting necessary information from a large amount of data.

[0283] Mode for Carrying out the Invention

[0284] Overview of the System

[0285] The present invention combines an emotion recognition engine with a system that receives voice data, converts it into text data, and further summarizes it. Specifically, a user uploads voice data of a meeting via smart glasses, the server analyzes the voice data and converts it into text data, analyzes the emotion using an emotion engine, and finally summarizes the text data and distributes it to the user.

[0286] Configuration of the Program

[0287] The server uses the following hardware and software:

[0288] Hardware: Smart glasses (e.g., Google Glass (registered trademark)), built-in microphone, audio interface

[0289] Software: Python, SpeechRecognition library, transformers library

[0290] Processing Flow of the Program

[0291] 1. Reception of Voice Data

[0292] The user uploads the voice data of the meeting to the server through smart glasses. The voice is recorded by the built-in microphone of the smart glasses, and the data is transmitted to the server via the network.

[0293] 2. Saving of Voice Data

[0294] The server saves the received audio file. The saved audio file is stored in a temporary storage area and used for subsequent processing.

[0295] 3. Converting audio data to text

[0296] The server passes the saved audio file to the speech recognition engine, which converts it into text data. The speech recognition engine analyzes the audio waveform and generates corresponding text information. This is done using the SpeechRecognition library.

[0297] 4. Emotion analysis

[0298] The server uses an emotion recognition engine to analyze the user's emotions from the audio data. It identifies the user's emotions (e.g., joy, anger, sadness, etc.) from the audio data and its text, and extracts that information. The transformers library is used for this purpose.

[0299] 5. Text Data Summary

[0300] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. If necessary, the results of sentiment analysis are also reflected in the summary.

[0301] 6. Distribution of summaries

[0302] The server delivers the generated summary to the user. The summary is sent to the user's specified email address, and the user can view the summary via the received email. It can also be displayed on the system dashboard or through a dedicated application.

[0303] Specific example

[0304] For example, if the following audio data was recorded during a meeting:

[0305] "We discussed the launch of the new production line. Many opinions were put forward to improve efficiency. Subsequently, we also talked about strengthening safety measures. It is planned to summarize improvement proposals from each department by the next meeting."

[0306] When this voice data is uploaded to the system, the system will process it as follows:

[0307] 1. Convert the voice data into text

[0308] 2. Analyze the sentiment of each speech (for example, the joy in the part discussing the launch of the new production line and the tension in the part discussing the strengthening of safety measures, etc.)

[0309] 3. Summarize the text, and the summary text should also reflect the sentiment tone

[0310] An example of the generated summary:

[0311] "Many opinions were put forward to improve the efficiency of launching the new production line. Strengthening safety measures was also discussed, and improvement proposals are planned to be summarized by the next meeting."

[0312] Example of a prompt sentence

[0313] "Please convert the following voice data into text, analyze the sentiment, and summarize it. The voice data contains discussions on the launch of a new production line and the strengthening of safety measures."

[0314] In this way, through the summary generation system for voice data centered on the server, users can efficiently grasp the main points and emotional background of the meeting.

[0315] The flow of the specific process in Application Example 2 will be described using Figure 14.

[0316] The flow of program processing

[0317] Step 1:

[0318] The user records audio data using the built-in microphone of the smart glasses. This audio data includes the content of meetings and discussions. After recording is complete, the user uploads the audio data to the server using the application on the smart glasses.

[0319] Input: Audio data recorded with smart glasses

[0320] Output: Audio data sent to the server via the network.

[0321] Step 2:

[0322] The server saves the received audio files to a temporary storage area. This saving process ensures that the audio data is stored properly to prevent loss.

[0323] Input: Received audio data

[0324] Output: Saved audio data

[0325] Step 3:

[0326] The server converts the stored audio data into text data using the SpeechRecognition library. Specifically, it analyzes the audio waveform and converts it into corresponding text information. This process also includes noise reduction and detection of speech boundaries.

[0327] Input: Saved audio data

[0328] Output: Text data

[0329] Step 4:

[0330] The server analyzes the audio data and its accompanying text data to extract emotional information. To achieve this, it uses the transformers library to automatically identify the emotions expressed in the user's statements (e.g., joy, anger, sadness, etc.) and extract that information.

[0331] Input: Text data

[0332] Output: Text data with emotional information added.

[0333] Step 5:

[0334] The server uses a natural language processing engine to summarize text data and sentiment information. It extracts important keywords and phrases and generates a concise summary. The summary also incorporates the results of sentiment analysis.

[0335] Input: Text data with emotional information attached

[0336] Output: Summarized text data

[0337] Step 6:

[0338] The server distributes the generated summary to the user. The summary can be sent to the user's specified email address, or it can be displayed on the system dashboard or a dedicated application.

[0339] Input: Summarized text data

[0340] Output: Summarized text data sent to and delivered to the user.

[0341] Through the above processing steps, users can efficiently grasp the content of the meeting and understand its emotional context.

[0342] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0343] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0344] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0345] [Second Embodiment]

[0346] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0347] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0348] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0349] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0350] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0351] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0352] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0353] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0354] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0355] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0356] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0357] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0358] System Overview

[0359] This invention relates to a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user can upload audio data of a meeting to a server, the server analyzes the audio data, converts it into text data, and finally summarizes it before delivering it to the user.

[0360] Specific processing details of the program

[0361] 1. Receiving audio data

[0362] The user uploads the meeting audio data to the server. The server temporarily stores the received audio data. This process is initiated when the user selects and uploads the file.

[0363] 2. Text conversion of audio data

[0364] The server converts the received audio data into text data using a speech recognition engine. The speech recognition engine analyzes the audio waveform and generates corresponding character information.

[0365] 3. Text Data Summary

[0366] The server passes the text data to a natural language processing engine, which extracts and summarizes the important information. The natural language processing engine analyzes the meaning of the text and extracts key points.

[0367] 4. Distribution of summaries

[0368] The server sends the generated summary to the user. The user can receive this summary via email or other means.

[0369] Specific example

[0370] For example, if the following audio data was recorded during a meeting:

[0371] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0372] When this audio data is uploaded to the system, the system generates the following summary:

[0373] "The agenda included discussing the new product release, a new marketing strategy, expanding the target market, and confirming the date for the next meeting."

[0374] By reviewing the summarized content, users can quickly grasp the key points of a meeting. This system provides an efficient way to obtain important information even when unable to attend a meeting.

[0375] The following describes the processing flow.

[0376] Step 1:

[0377] The user uploads the meeting audio file to the server. The user selects the audio file using a web browser or dedicated application and clicks the upload button. This action sends the audio file to the server over the network.

[0378] Step 2:

[0379] The server saves the received audio files. Uploaded audio files are stored in a temporary storage area for subsequent processing. This saving process includes checking the size of the audio file and converting it to an appropriate format.

[0380] Step 3:

[0381] The server passes the stored audio files to the speech recognition engine, where they are converted into text data. Specifically, speech recognition software is used to analyze the audio files and convert the audio signals into corresponding strings of characters. This conversion process includes noise reduction and detection of speech boundaries.

[0382] Step 4:

[0383] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. The summary is compressed to the extent that the meaning of the original text is not lost.

[0384] Step 5:

[0385] The server distributes the generated summary to the user. The summary results are sent to the email address specified by the user, and the user can check the summary in the received email. It can also be displayed on the system dashboard or in a dedicated application.

[0386] (Example 1)

[0387] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0388] Conventional systems were time-consuming to analyze and summarize audio data, making it difficult for users to efficiently obtain information. Furthermore, insufficient temporary storage of audio data and verification of data integrity posed a risk of data reliability degradation. Therefore, there was a need for a means to quickly and accurately obtain important information from meetings and audio recordings.

[0389] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0390] In this invention, the server includes means for uploading audio data to a communication device, means for converting the audio data received by the communication device into text data using an analysis device, means for analyzing the converted text data using a natural language processing device to extract and summarize important information, and means for outputting the summarized text data via the communication device. This enables rapid analysis and accurate summarization of audio data, allowing users to efficiently obtain important information.

[0391] 1. "Audio data" refers to data recorded in digital format from audio information such as meetings and conversations.

[0392] 2. A "communication device" is a device that transmits and receives voice data and summarized text data, and is capable of uploading and downloading data via an internet connection.

[0393] 3. "Analysis device" refers to a device for converting audio data into text data, and includes a speech recognition engine and other analysis software.

[0394] 4. "Text data" refers to character information after it has been analyzed by a speech recognition engine, and is data that represents the content of speech data in text format.

[0395] 5. A "natural language processing device" is a device that analyzes text data, extracts important information, and generates a summary.

[0396] 6. "Summary text data" refers to text data in a concise format, generated by a natural language processing device, from which only the essential information has been extracted from the original text data.

[0397] 7. "Means of temporary storage" refers to means of temporarily storing received audio data in a storage device so that it can be used for subsequent processing.

[0398] 8. "Data integrity verification" is a verification method to confirm that the received data has been transmitted accurately from the sender without any changes.

[0399] Modes for carrying out the invention

[0400] System Overview

[0401] This invention relates to a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user uploads audio data of a meeting to a communication device, a server converts that audio data into text data using an analysis device, summarizes it using a natural language processing device, and delivers it to the user.

[0402] Specific processing details of the program

[0403] 1. Receiving audio data

[0404] Users upload meeting audio data to the server via a communication device. Users use a web application to select and upload audio files. Once the upload is complete, the server temporarily stores the audio data. At this time, a hash value is calculated to verify the integrity of the data.

[0405] 2. Text conversion of audio data

[0406] The server passes the stored audio data to the analysis device. This analysis device uses a speech recognition engine (e.g., Google Speech-to-Text API). The speech recognition engine analyzes the waveform of the audio data and generates corresponding text data. The generated text data is received by the server and temporarily stored.

[0407] 3. Text Data Summary

[0408] The server passes the generated text data to a natural language processing unit (NLP). This NLP uses a generative AI model (e.g., OpenAI GPT-4). The NLP analyzes the text data, extracts important information, and generates a summary. This summary extracts key points from the original text data and is stored secondarily by the server.

[0409] 4. Distribution of summaries

[0410] The server delivers the generated summary text data to the user. Delivery methods include email and notifications via a dedicated portal. Users receive notifications and can click a link to view the summary.

[0411] Specific example

[0412] For example, if the following audio data is recorded during a meeting:

[0413] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0414] When this audio data is uploaded to the system, the system generates the following summary:

[0415] "The agenda included discussing the new product release, a new marketing strategy, expanding the target market, and confirming the date for the next meeting."

[0416] Example of a prompt

[0417] Examples of prompt statements are as follows:

[0418] Please summarize the following meeting audio transcript:

[0419] At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting.

[0420] By inputting this prompt into the generation AI model, an appropriate summary can be generated.

[0421] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0422] Step 1: Upload audio data

[0423] The user uploads audio data from their terminal to the communication device. As input, the user selects a conference audio file (e.g., MP3 or WAV format) and presses the upload button. As output, the audio file is sent to the communication device. Specifically, the user opens a file selection dialog through their browser, selects the file, and begins sending. The communication device then forwards the received audio file to the server.

[0424] Step 2: Receiving and temporarily storing audio data

[0425] The server receives audio data transmitted from the communication device and stores it in a temporary storage area. The input is audio data sent from the communication device. The output is that audio data stored in the server's temporary storage area. Specifically, the server calculates a hash value (e.g., SHA-256) to verify data integrity and records the file name and hash value in the temporary storage area.

[0426] Step 3: Convert speech to text

[0427] The server passes the audio data stored in the temporary storage area to the analysis device. The input is the stored audio file sent to the analysis device. The output is text data generated by a speech recognition engine (e.g., Google Speech-to-Text API). Specifically, the server sends the audio data as a request to the API and stores the returned text data in the temporary storage area.

[0428] Step 4: Natural language processing of text data

[0429] The server passes text data stored in the primary storage area to a natural language processing unit (e.g., a generative AI model). The generative AI model receives text data as input. The output is summarized text data with important information extracted. Specifically, the server sends text data as a prompt to the generative AI model and stores the generated summary in the secondary storage area.

[0430] Step 5: Distribution of the summary

[0431] The server delivers summarized text data stored in a secondary storage area to the user. The input is the summarized text data passed to the server's notification system. The output is an email notification or display on a web portal to the user. Specifically, the server delivers the summarized text according to the user's settings (e.g., email, push notification). The user receives the notification and clicks the link to view the summary.

[0432] (Application Example 1)

[0433] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0434] In autonomous vehicles, if a driver needs to record important notes or instructions by voice while driving, they must stop the vehicle to take notes. This interferes with driving and reduces efficiency. Furthermore, when reviewing the voice data later, extracting important information from the vast amount of data is time-consuming and hinders quick decision-making. To solve this, a system is needed that transcribes and summarizes voice data in real time, allowing drivers to quickly obtain the information they need.

[0435] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0436] In this invention, the server includes means for receiving voice data, means for converting the received voice data into text data, means for summarizing the converted text data, and means for outputting the summarized text data generated at a terminal installed in the vehicle. This makes it possible for drivers to give instructions or make notes by voice while driving, and for important information to be quickly obtained by converting and summarizing them into text in real time.

[0437] "Audio data" refers to information obtained by converting sound waves into digital signals via microphones or other acoustic sensors.

[0438] "Text data" refers to digital text obtained by analyzing audio data and converting it into corresponding textual information.

[0439] A "summary" is a text document that extracts important information from converted text data and presents it concisely.

[0440] A "terminal installed in a vehicle" refers to a computer or smart device installed inside a vehicle that performs voice recognition and displays and stores transcribed data.

[0441] A "user interface" is a screen, device, or software used by a system to exchange information with a user.

[0442] "Storage" means saving data to a storage medium or database so that it can be accessed later.

[0443] "Analysis" is the process of understanding the content of text data using natural language processing and algorithms, and extracting the necessary information.

[0444] "Reception" refers to acquiring audio data from an external source and incorporating it into the system.

[0445] "Conversion" refers to the process of reconstructing audio data into text data.

[0446] "Output" refers to displaying or sending summarized text data to the user.

[0447] This invention relates to a system that transcribes audio data into text in real time, summarizes the text, and displays it on a terminal inside an autonomous vehicle.

[0448] Hardware and software to be used

[0449] hardware

[0450] Devices installed in automobiles (smartphones and in-car smart devices)

[0451] Microphone for voice input

[0452] software

[0453] Google Cloud Speech-to-Text API

[0454] GPT-3.5

[0455] Python

[0456] Explanation of program processing

[0457] Receiving audio data

[0458] The user inputs voice memos and instructions via the microphone while driving. The device receives this voice data in real time.

[0459] Text conversion of audio data

[0460] The server uses the Google Cloud Speech-to-Text API to convert received audio data into text data. This API provides highly accurate speech recognition capabilities, analyzing the audio waveform to generate corresponding text information.

[0461] Summary of text data

[0462] The server summarizes the converted text data using GPT-3.5. GPT-3.5 is a natural language processing engine that analyzes the meaning of the text, extracts important information, and summarizes it concisely.

[0463] Output and save summary

[0464] The summarized text data is displayed on a terminal installed in the vehicle via the user interface. Additionally, this summarized data can be saved on the terminal as needed for later review by the user.

[0465] Specific example

[0466] For example, if a driver enters the following voice memo while driving:

[0467] "We departed at 10:00 AM and encountered traffic congestion on the way to our destination. I instructed everyone to take a detour. While driving, I received an important phone call and instructed them to prepare materials for the afternoon meeting."

[0468] This audio data is summarized by the system as follows:

[0469] "I instructed them to take a detour due to traffic congestion after departure. I also instructed them to prepare materials for the afternoon meeting via an important phone call."

[0470] Example of a prompt

[0471] Please summarize the following text: "We departed at 10:00 AM and encountered traffic congestion on the way to our destination. I instructed everyone to take a detour. I received an important phone call while driving and instructed everyone to prepare materials for the afternoon meeting."

[0472] This will allow users to efficiently record voice memos while driving and quickly retrieve important information.

[0473] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0474] Step 1:

[0475] While driving, the user inputs voice memos and instructions via the microphone. The input voice data is saved in real time to the terminal's storage device. Here, the user's voice is recorded as a digital signal.

[0476] Input: User's voice memos or instructions

[0477] Output: Digital audio data stored on the device

[0478] Step 2:

[0479] The terminal sends the stored audio data to the server. The server receives this audio data and stores it temporarily. Here, digital audio data is transferred.

[0480] Input: Digital audio data stored on the device

[0481] Output: Digital audio data stored on the server

[0482] Step 3:

[0483] The server uses the Google Cloud Speech-to-Text API to convert the received audio data into text data. The server sends the audio data to the API and receives and stores the returned text. In this process, the audio data is converted into text information.

[0484] Input: Digital audio data stored on the server

[0485] Output: Text data

[0486] Step 4:

[0487] The server passes the text data to GPT-3.5 for natural language processing. GPT-3.5 parses the text data based on the input prompt sentence, extracts important information, and summarizes it. The summary result is returned to the server. Here, the text data is compressed and converted into a shorter, more concise form of text.

[0488] Input: Text data and prompt text

[0489] Output: Summarized text data

[0490] Step 5:

[0491] The server sends the summarized text data to the terminal. The terminal displays the received summary through its user interface. The user can review the summary on the screen. Here, the final output is provided to the user.

[0492] Input: Summarized text data

[0493] Output: Summary data in a format that can be viewed by the user.

[0494] Step 6:

[0495] If necessary, the device saves the displayed summary data on its own. The user can then refer to and review this data later. Here, the summary data is saved for future reference.

[0496] Input: User-confirmed summary data

[0497] Output: Summary data saved on the device

[0498] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0499] System Overview

[0500] This invention combines a system that receives audio data, converts it into text data, and then summarizes it with an emotion engine that recognizes the user's emotions. Specifically, a user uploads audio data of a meeting to a server, the server analyzes the audio data and converts it into text data, analyzes the emotions using the emotion engine, and finally summarizes the text data and delivers it to the user.

[0501] Specific processing details of the program

[0502] 1. Receiving audio data

[0503] The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated application to select the audio file and click the upload button. This action sends the audio file to the server over the network.

[0504] 2. Saving audio data

[0505] The server saves the received audio files. Uploaded audio files are stored in a temporary storage area for use in subsequent processing.

[0506] 3. Converting audio data to text

[0507] The server passes the stored audio files to the speech recognition engine, which converts them into text data. The speech recognition engine analyzes the audio waveform and generates corresponding character information. This conversion process includes noise reduction and detection of speech boundaries.

[0508] 4. Emotion analysis

[0509] The server uses an emotion recognition engine to analyze the user's emotions from the audio data. It identifies and extracts information about the user's emotions (e.g., joy, anger, sadness) from the audio data and its text. This also allows for understanding the emotional tone of each statement made during a meeting.

[0510] 5. Text Data Summary

[0511] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. If necessary, the results of sentiment analysis are also reflected in the summary.

[0512] 6. Distribution of summaries

[0513] The server distributes the generated summary to the user. The summary results are sent to the email address specified by the user, and the user can check the summary in the received email. It can also be displayed on the system dashboard or in a dedicated application.

[0514] Specific example

[0515] For example, if the following audio data was recorded during a meeting:

[0516] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0517] When this audio data is uploaded to the system, the system processes it as follows:

[0518] 1. Convert audio data to text.

[0519] 2. Analyze the emotions expressed in each statement (for example, the joy felt when discussing the "release of a new product" and the tension felt when discussing the "review of the marketing strategy").

[0520] 3. Summarize the text, and reflect the emotional tone in the summary.

[0521] An example of a generated summary:

[0522] "There was a discussion about the new product release, and most opinions were optimistic. Following that, there was a discussion about reviewing the marketing strategy and expanding the target market, and some sense of urgency was evident. Finally, the date for the next meeting was confirmed."

[0523] By reviewing the emotional tone along with the summarized content, users can quickly grasp the key points of a meeting and their emotional context. This system not only provides an efficient way to obtain important information even when unable to attend a meeting, but also allows users to understand the atmosphere of the meeting and the emotional reactions of the participants.

[0524] The following describes the processing flow.

[0525] Step 1:

[0526] The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated app to select the audio file and clicks the upload button. This action sends the audio file to the server over the network.

[0527] Step 2:

[0528] The server saves the received audio files to a temporary storage area. The saving process also includes checking the format and size of the audio files and converting them to the appropriate format.

[0529] Step 3:

[0530] The server passes the stored audio files to the speech recognition engine, which converts them into text data. In this process, the speech recognition engine analyzes the audio waveform and sequentially generates corresponding text information. Pre-processing such as keyword recognition and noise reduction is also performed.

[0531] Step 4:

[0532] The server passes text data to an emotion recognition engine to analyze the user's emotions. The emotion recognition engine identifies the speaker's emotional state (e.g., joy, anger, sadness) from the text data and audio features, and adds the result to the text data. This emotion information is either added to individual sentences or grouped by segment.

[0533] Step 5:

[0534] The server passes text data, including sentiment data, to a natural language processing engine to generate a summary. The natural language processing engine extracts key keywords and phrases from the text, integrates that information with sentiment data, and creates a concise summary.

[0535] Step 6:

[0536] The server delivers the completed summary to the user. Delivery methods include sending to a specified email address and displaying it on a dedicated dashboard or app. The email contains the summary text, and important sentiment information is highlighted as appropriate.

[0537] (Example 2)

[0538] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0539] Traditional systems that convert audio data into text and summarize it have difficulty overlooking important information and understanding the emotional context of meeting participants. Furthermore, the lack of a means to reflect emotional information in the summary makes it difficult to grasp the overall atmosphere of the meeting.

[0540] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0541] In this invention, the server includes means for receiving audio data, means for storing the received audio data, means for converting the stored audio data into text data, means for analyzing the converted text data to extract emotional information, means for summarizing the text data taking the emotional information into consideration, and means for outputting the summarized text data. This enables accurate summarization that reflects emotional information.

[0542] "Audio data" refers to information recorded in digital or analog format.

[0543] "Text data" refers to information expressed as characters.

[0544] "Means of storage" refers to technology or equipment for temporarily or long-term storage of received data.

[0545] "Means of conversion" refers to a technology or device for converting data in one format to data in another format.

[0546] "Means of analysis" refers to techniques or devices for examining, evaluating, or breaking down data to extract specific information.

[0547] "Emotional information" refers to information that indicates the emotional tone or emotional state contained within the data.

[0548] "Means of summarization" refers to techniques or devices that extract important points from large amounts of data and convert them into a shorter format.

[0549] "Output method" refers to a technology or device that provides processed data in a format that can be read by the user.

[0550] "Natural language processing technology" refers to the technology that enables computers to understand and process natural language, which humans use on a daily basis.

[0551] A "prompt statement" is an input statement for a generative AI model, a text used to induce a specific output.

[0552] System Configuration

[0553] This invention combines a system that receives audio data, converts it to text data, and summarizes it with an emotion engine that recognizes the user's emotions. The user uploads meeting audio data to a server, the server analyzes the audio data, converts it to text data, analyzes the emotions, and finally summarizes the text data and delivers it to the user.

[0554] Hardware and software to be used

[0555] The system of this invention primarily uses the following hardware and software:

[0556] Server: Receives, stores, converts, analyzes, and summarizes audio data.

[0557] Web browser or dedicated application: A terminal for users to upload audio data.

[0558] Speech recognition engine: Uses the Google Cloud Speech-to-Text API to convert speech data into text data.

[0559] Emotion recognition engine: Uses IBM Watson Tone Analyzer to extract emotional information from text data.

[0560] Summarizing text data using the natural language processing engine: OpenAI GPT-3.

[0561] Specific operation of the system

[0562] 1. The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated application to select the audio file and click the upload button. The audio file is sent to the server over the network using an HTTP POST request.

[0563] 2. The server saves the received audio files to a temporary storage area. For example, it stores them in cloud storage.

[0564] 3. The server sends the audio file to the Google Cloud Speech-to-Text API, where the audio data is converted into text data. The speech recognition engine analyzes the audio waveform and generates the corresponding text information.

[0565] 4. The server retrieves the converted text data and passes it to IBM Watson Tone Analyzer to analyze the emotional information. The analysis results are displayed numerically, for example, "Joy: 0.7, Anger: 0.1, Sadness: 0.2".

[0566] 5. The server sends the text data along with a prompt to OpenAI GPT-3 to generate a summary. An example of the prompt to be generated is "Please summarize the contents of the following meeting." + "Acquired text data".

[0567] 6. The server distributes the generated summary to the user. The summary results are sent via email to the email address specified by the user. Users can also view the summary results using the system dashboard or a dedicated application.

[0568] Specific example

[0569] For example, consider a case where the following audio data was recorded during a meeting:

[0570] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0571] When this audio data is uploaded to the system, it is processed as follows:

[0572] 1. The user uses a dedicated application to upload audio files to the server.

[0573] 2. The server saves the audio files to cloud storage (e.g., Amazon S3).

[0574] 3. The server sends the audio file to the Google Cloud Speech-to-Text API and retrieves the text data "We discussed the release of the new product...".

[0575] 4. The server passes the text data to IBM Watson Tone Analyzer and obtains emotion information: "Joy: 0.7, Anger: 0.1, Sadness: 0.2".

[0576] 5. The server sends the text data to OpenAI GPT-3 along with the prompt "Please summarize the contents of the following meeting." + "Acquired text data," generating the summary "There was a discussion about a new product release, and there were many optimistic opinions..."

[0577] 6. The server sends the generated summary to the user via email and also makes it available on the dashboard.

[0578] In this way, the present invention enables users to quickly grasp important information, including the overall emotional context of a meeting, through efficient analysis of audio data and summarization that reflects emotional information.

[0579] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0580] Step 1:

[0581] The user uploads the meeting audio data to the server. The input is the audio file selected by the user, and the output is the file sent to the server via an HTTP POST request. The user opens a web browser or dedicated application, selects the audio file, and clicks the upload button. This action sends the audio file to the server.

[0582] Step 2:

[0583] The server saves the received audio file to a temporary storage area. The input is the audio file sent by the user, and the output is the file saved in storage. When the server receives the audio file, it saves it to cloud storage (e.g., Amazon S3). This save operation makes it accessible in subsequent processing steps.

[0584] Step 3:

[0585] The server sends the stored audio file to the Google Cloud Speech-to-Text API and converts it into text data. The input is an audio file in cloud storage, and the output is text data obtained from the speech recognition engine. The server sends the audio file via the API, analyzes the audio waveform, and obtains the corresponding text information. Noise reduction and speech segmentation are also performed during this process.

[0586] Step 4:

[0587] The server sends text data to the IBM Watson Tone Analyzer for emotional analysis. The input is text data from the speech recognition engine, and the output is emotional information from the emotional recognition engine. The server passes the acquired text data to the emotional recognition engine to obtain emotional information (e.g., "Joy: 0.7, Anger: 0.1, Sadness: 0.2"). This information is then used in the subsequent summary.

[0588] Step 5:

[0589] The server sends text data along with a prompt to OpenAI GPT-3, which then generates a summary. The input consists of text data and a prompt, and the output is the summary obtained from the generating AI model. An example of a prompt is "Please summarize the contents of the following meeting." + "Acquired text data". The server then generates and saves the summary.

[0590] Step 6:

[0591] The server delivers the generated summary to the user. The input is the generated summary text, and the output is the summary result sent to the user. The server sends the summary result to an email address. Users can also view the summary result on the system dashboard or a dedicated application. Emails are sent using the SMTP protocol, and the dashboard displays the summary result securely using HTTPS.

[0592] (Application Example 2)

[0593] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0594] Systems that extract and summarize important information from audio data face the challenge of understanding the emotional context of meetings and discussions. Conventional technologies cannot extract emotional information from transcribed data, making it difficult to grasp the nuances and emotional shifts of conversations. To solve this problem, it is necessary to simultaneously extract emotional information from audio data and incorporate it into the summary.

[0595] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0596] In this invention, the server includes means for receiving audio data, means for converting the received audio data into text data, means for summarizing the converted text data, means for adding emotional information to the summarized text data, and means for outputting the summarized text data. This allows for the extraction of emotional information along with the summary of audio data, and by reflecting the results in the summary, it becomes possible to simultaneously understand the content of meetings and discussions and their emotional background.

[0597] "Voice data" refers to digital information that digitizes human voice, including conversations, instructions, and other audio files.

[0598] "Means" refers to the methods, apparatus, or techniques used to achieve a particular purpose.

[0599] "Text data" refers to data that is stored or displayed as character information, specifically information obtained by converting audio data into text.

[0600] A "summary" refers to a short and concise version of the content of a text or audio recording, extracting the most important information and making it more compact.

[0601] "Emotional information" refers to data on emotional states analyzed from audio or text, such as information representing emotions like joy, anger, or sadness.

[0602] "Output" refers to the act of displaying, printing, or sending processed data to another system.

[0603] "Receiving" refers to the process of receiving and processing data or information from an external source.

[0604] "Storage" refers to the act of keeping recorded data without losing it, so that it can be retrieved and used later.

[0605] "Analysis" refers to the process of breaking down data and understanding its structure and meaning.

[0606] "Extraction" refers to the act of taking out necessary information from a large amount of data.

[0607] Modes for carrying out the invention

[0608] System Overview

[0609] This invention combines an emotion recognition engine with a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user can upload audio data of a meeting via smart glasses, a server can analyze the audio data and convert it into text data, analyze the emotions using the emotion engine, and finally summarize the text data and deliver it to the user.

[0610] Program Configuration

[0611] The server uses the following hardware and software:

[0612] Hardware: Smart glasses (e.g., Google Glass), built-in microphone, audio interface

[0613] Software: Python, SpeechRecognition library, transformers library

[0614] Program processing flow

[0615] 1. Receiving audio data

[0616] The user uploads the meeting audio data to the server via smart glasses. The audio is recorded using the smart glasses' built-in microphone, and that data is transmitted to the server over the network.

[0617] 2. Saving audio data

[0618] The server saves the received audio file. The saved audio file is stored in a temporary storage area and used for subsequent processing.

[0619] 3. Converting audio data to text

[0620] The server passes the saved audio file to the speech recognition engine, which converts it into text data. The speech recognition engine analyzes the audio waveform and generates corresponding text information. This is done using the SpeechRecognition library.

[0621] 4. Emotion analysis

[0622] The server uses an emotion recognition engine to analyze the user's emotions from the audio data. It identifies the user's emotions (e.g., joy, anger, sadness, etc.) from the audio data and its text, and extracts that information. The transformers library is used for this purpose.

[0623] 5. Text Data Summary

[0624] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. If necessary, the results of sentiment analysis are also reflected in the summary.

[0625] 6. Distribution of summaries

[0626] The server delivers the generated summary to the user. The summary is sent to the user's specified email address, and the user can view the summary via the received email. It can also be displayed on the system dashboard or through a dedicated application.

[0627] Specific example

[0628] For example, if the following audio data was recorded during a meeting:

[0629] "We discussed setting up a new production line. Many suggestions were made to improve efficiency. Afterwards, we also discussed strengthening safety measures. We plan to compile improvement proposals from each department before the next meeting."

[0630] When this audio data is uploaded to the system, the system processes it as follows:

[0631] 1. Convert audio data to text.

[0632] 2. Analyze the emotions expressed in each statement (for example, the joy felt when discussing "launching a new production line" and the tension felt when discussing "strengthening safety measures").

[0633] 3. Summarize the text, and reflect the emotional tone in the summary.

[0634] An example of a generated summary:

[0635] "Many suggestions were made to improve efficiency when setting up the new production line. Strengthening safety measures was also discussed, and improvement proposals are expected to be compiled by the next meeting."

[0636] Example of a prompt

[0637] "Please convert the following audio data into text, analyze the emotions, and summarize it. The audio data includes discussions about launching a new production line and strengthening safety measures."

[0638] In this way, a server-centric system for generating summaries from audio data allows users to efficiently grasp the key points and emotional context of a meeting.

[0639] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0640] Program processing flow

[0641] Step 1:

[0642] The user records audio data using the built-in microphone of the smart glasses. This audio data includes the content of meetings and discussions. After recording is complete, the user uploads the audio data to the server using the application on the smart glasses.

[0643] Input: Audio data recorded with smart glasses

[0644] Output: Audio data sent to the server via the network.

[0645] Step 2:

[0646] The server saves the received audio files to a temporary storage area. This saving process ensures that the audio data is stored properly to prevent loss.

[0647] Input: Received audio data

[0648] Output: Saved audio data

[0649] Step 3:

[0650] The server converts the stored audio data into text data using the SpeechRecognition library. Specifically, it analyzes the audio waveform and converts it into corresponding text information. This process also includes noise reduction and detection of speech boundaries.

[0651] Input: Saved audio data

[0652] Output: Text data

[0653] Step 4:

[0654] The server analyzes the audio data and its accompanying text data to extract emotional information. To achieve this, it uses the transformers library to automatically identify the emotions expressed in the user's statements (e.g., joy, anger, sadness, etc.) and extract that information.

[0655] Input: Text data

[0656] Output: Text data with emotional information added.

[0657] Step 5:

[0658] The server uses a natural language processing engine to summarize text data and sentiment information. It extracts important keywords and phrases and generates a concise summary. The summary also incorporates the results of sentiment analysis.

[0659] Input: Text data with emotional information attached

[0660] Output: Summarized text data

[0661] Step 6:

[0662] The server distributes the generated summary to the user. The summary can be sent to the user's specified email address, or it can be displayed on the system dashboard or a dedicated application.

[0663] Input: Summarized text data

[0664] Output: Summarized text data sent to and delivered to the user.

[0665] Through the above processing steps, users can efficiently grasp the content of the meeting and understand its emotional context.

[0666] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0667] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0668] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0669] [Third Embodiment]

[0670] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0671] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0672] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0673] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0674] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0675] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0676] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0677] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0678] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0679] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0680] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0681] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0682] System Overview

[0683] This invention relates to a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user can upload audio data of a meeting to a server, the server analyzes the audio data, converts it into text data, and finally summarizes it before delivering it to the user.

[0684] Specific processing details of the program

[0685] 1. Receiving audio data

[0686] The user uploads the meeting audio data to the server. The server temporarily stores the received audio data. This process is initiated when the user selects and uploads the file.

[0687] 2. Text conversion of audio data

[0688] The server converts the received audio data into text data using a speech recognition engine. The speech recognition engine analyzes the audio waveform and generates corresponding character information.

[0689] 3. Text Data Summary

[0690] The server passes the text data to a natural language processing engine, which extracts and summarizes the important information. The natural language processing engine analyzes the meaning of the text and extracts key points.

[0691] 4. Distribution of summaries

[0692] The server sends the generated summary to the user. The user can receive this summary via email or other means.

[0693] Specific example

[0694] For example, if the following audio data was recorded during a meeting:

[0695] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0696] When this audio data is uploaded to the system, the system generates the following summary:

[0697] "The agenda included discussing the new product release, a new marketing strategy, expanding the target market, and confirming the date for the next meeting."

[0698] By reviewing the summarized content, users can quickly grasp the key points of a meeting. This system provides an efficient way to obtain important information even when unable to attend a meeting.

[0699] The following describes the processing flow.

[0700] Step 1:

[0701] The user uploads the meeting audio file to the server. The user selects the audio file using a web browser or dedicated application and clicks the upload button. This action sends the audio file to the server over the network.

[0702] Step 2:

[0703] The server saves the received audio files. Uploaded audio files are stored in a temporary storage area for subsequent processing. This saving process includes checking the size of the audio file and converting it to an appropriate format.

[0704] Step 3:

[0705] The server passes the stored audio files to the speech recognition engine, where they are converted into text data. Specifically, speech recognition software is used to analyze the audio files and convert the audio signals into corresponding strings of characters. This conversion process includes noise reduction and detection of speech boundaries.

[0706] Step 4:

[0707] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. The summary is compressed to the extent that the meaning of the original text is not lost.

[0708] Step 5:

[0709] The server distributes the generated summary to the user. The summary results are sent to the email address specified by the user, and the user can check the summary in the received email. It can also be displayed on the system dashboard or in a dedicated application.

[0710] (Example 1)

[0711] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0712] Conventional systems were time-consuming to analyze and summarize audio data, making it difficult for users to efficiently obtain information. Furthermore, insufficient temporary storage of audio data and verification of data integrity posed a risk of data reliability degradation. Therefore, there was a need for a means to quickly and accurately obtain important information from meetings and audio recordings.

[0713] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0714] In this invention, the server includes means for uploading audio data to a communication device, means for converting the audio data received by the communication device into text data using an analysis device, means for analyzing the converted text data using a natural language processing device to extract and summarize important information, and means for outputting the summarized text data via the communication device. This enables rapid analysis and accurate summarization of audio data, allowing users to efficiently obtain important information.

[0715] 1. "Audio data" refers to data recorded in digital format from audio information such as meetings and conversations.

[0716] 2. A "communication device" is a device that transmits and receives voice data and summarized text data, and is capable of uploading and downloading data via an internet connection.

[0717] 3. "Analysis device" refers to a device for converting audio data into text data, and includes a speech recognition engine and other analysis software.

[0718] 4. "Text data" refers to character information after it has been analyzed by a speech recognition engine, and is data that represents the content of speech data in text format.

[0719] 5. A "natural language processing device" is a device that analyzes text data, extracts important information, and generates a summary.

[0720] 6. "Summary text data" refers to text data in a concise format, generated by a natural language processing device, from which only the essential information has been extracted from the original text data.

[0721] 7. "Means of temporary storage" refers to means of temporarily storing received audio data in a storage device so that it can be used for subsequent processing.

[0722] 8. "Data integrity verification" is a verification method to confirm that the received data has been transmitted accurately from the sender without any changes.

[0723] Modes for carrying out the invention

[0724] System Overview

[0725] This invention relates to a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user uploads audio data of a meeting to a communication device, a server converts that audio data into text data using an analysis device, summarizes it using a natural language processing device, and delivers it to the user.

[0726] Specific processing details of the program

[0727] 1. Receiving audio data

[0728] Users upload meeting audio data to the server via a communication device. Users use a web application to select and upload audio files. Once the upload is complete, the server temporarily stores the audio data. At this time, a hash value is calculated to verify the integrity of the data.

[0729] 2. Text conversion of audio data

[0730] The server passes the stored audio data to the analysis device. This analysis device uses a speech recognition engine (e.g., Google Speech-to-Text API). The speech recognition engine analyzes the waveform of the audio data and generates corresponding text data. The generated text data is received by the server and temporarily stored.

[0731] 3. Text Data Summary

[0732] The server passes the generated text data to a natural language processing unit (NLP). This NLP uses a generative AI model (e.g., OpenAI GPT-4). The NLP analyzes the text data, extracts important information, and generates a summary. This summary extracts key points from the original text data and is stored secondarily by the server.

[0733] 4. Distribution of summaries

[0734] The server delivers the generated summary text data to the user. Delivery methods include email and notifications via a dedicated portal. Users receive notifications and can click a link to view the summary.

[0735] Specific example

[0736] For example, if the following audio data is recorded during a meeting:

[0737] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0738] When this audio data is uploaded to the system, the system generates the following summary:

[0739] "The agenda included discussing the new product release, a new marketing strategy, expanding the target market, and confirming the date for the next meeting."

[0740] Example of a prompt

[0741] Examples of prompt statements are as follows:

[0742] Please summarize the following meeting audio transcript:

[0743] At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting.

[0744] By inputting this prompt into the generation AI model, an appropriate summary can be generated.

[0745] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0746] Step 1: Upload audio data

[0747] The user uploads audio data from their terminal to the communication device. As input, the user selects a conference audio file (e.g., MP3 or WAV format) and presses the upload button. As output, the audio file is sent to the communication device. Specifically, the user opens a file selection dialog through their browser, selects the file, and begins sending. The communication device then forwards the received audio file to the server.

[0748] Step 2: Receiving and temporarily storing audio data

[0749] The server receives audio data transmitted from the communication device and stores it in a temporary storage area. The input is audio data sent from the communication device. The output is that audio data stored in the server's temporary storage area. Specifically, the server calculates a hash value (e.g., SHA-256) to verify data integrity and records the file name and hash value in the temporary storage area.

[0750] Step 3: Convert speech to text

[0751] The server passes the audio data stored in the temporary storage area to the analysis device. The input is the stored audio file sent to the analysis device. The output is text data generated by a speech recognition engine (e.g., Google Speech-to-Text API). Specifically, the server sends the audio data as a request to the API and stores the returned text data in the temporary storage area.

[0752] Step 4: Natural language processing of text data

[0753] The server passes text data stored in the primary storage area to a natural language processing unit (e.g., a generative AI model). The generative AI model receives text data as input. The output is summarized text data with important information extracted. Specifically, the server sends text data as a prompt to the generative AI model and stores the generated summary in the secondary storage area.

[0754] Step 5: Distribution of the summary

[0755] The server delivers summarized text data stored in a secondary storage area to the user. The input is the summarized text data passed to the server's notification system. The output is an email notification or display on a web portal to the user. Specifically, the server delivers the summarized text according to the user's settings (e.g., email, push notification). The user receives the notification and clicks the link to view the summary.

[0756] (Application Example 1)

[0757] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0758] In autonomous vehicles, if a driver needs to record important notes or instructions by voice while driving, they must stop the vehicle to take notes. This interferes with driving and reduces efficiency. Furthermore, when reviewing the voice data later, extracting important information from the vast amount of data is time-consuming and hinders quick decision-making. To solve this, a system is needed that transcribes and summarizes voice data in real time, allowing drivers to quickly obtain the information they need.

[0759] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0760] In this invention, the server includes means for receiving voice data, means for converting the received voice data into text data, means for summarizing the converted text data, and means for outputting the summarized text data generated at a terminal installed in the vehicle. This makes it possible for drivers to give instructions or make notes by voice while driving, and for important information to be quickly obtained by converting and summarizing them into text in real time.

[0761] "Audio data" refers to information obtained by converting sound waves into digital signals via microphones or other acoustic sensors.

[0762] "Text data" refers to digital text obtained by analyzing audio data and converting it into corresponding textual information.

[0763] A "summary" is a text document that extracts important information from converted text data and presents it concisely.

[0764] A "terminal installed in a vehicle" refers to a computer or smart device installed inside a vehicle that performs voice recognition and displays and stores transcribed data.

[0765] A "user interface" is a screen, device, or software used by a system to exchange information with a user.

[0766] "Storage" means saving data to a storage medium or database so that it can be accessed later.

[0767] "Analysis" is the process of understanding the content of text data using natural language processing and algorithms, and extracting the necessary information.

[0768] "Reception" refers to acquiring audio data from an external source and incorporating it into the system.

[0769] "Conversion" refers to the process of reconstructing audio data into text data.

[0770] "Output" refers to displaying or sending summarized text data to the user.

[0771] This invention relates to a system that transcribes audio data into text in real time, summarizes the text, and displays it on a terminal inside an autonomous vehicle.

[0772] Hardware and software to be used

[0773] hardware

[0774] Devices installed in automobiles (smartphones and in-car smart devices)

[0775] Microphone for voice input

[0776] software

[0777] Google Cloud Speech-to-Text API

[0778] GPT-3.5

[0779] Python

[0780] Explanation of program processing

[0781] Receiving audio data

[0782] The user inputs voice memos and instructions via the microphone while driving. The device receives this voice data in real time.

[0783] Text conversion of audio data

[0784] The server uses the Google Cloud Speech-to-Text API to convert received audio data into text data. This API provides highly accurate speech recognition capabilities, analyzing the audio waveform to generate corresponding text information.

[0785] Summary of text data

[0786] The server summarizes the converted text data using GPT-3.5. GPT-3.5 is a natural language processing engine that analyzes the meaning of the text, extracts important information, and summarizes it concisely.

[0787] Output and save summary

[0788] The summarized text data is displayed on a terminal installed in the vehicle via the user interface. Additionally, this summarized data can be saved on the terminal as needed for later review by the user.

[0789] Specific example

[0790] For example, if a driver enters the following voice memo while driving:

[0791] "We departed at 10:00 AM and encountered traffic congestion on the way to our destination. I instructed everyone to take a detour. While driving, I received an important phone call and instructed them to prepare materials for the afternoon meeting."

[0792] This audio data is summarized by the system as follows:

[0793] "I instructed them to take a detour due to traffic congestion after departure. I also instructed them to prepare materials for the afternoon meeting via an important phone call."

[0794] Example of a prompt

[0795] Please summarize the following text: "We departed at 10:00 AM and encountered traffic congestion on the way to our destination. I instructed everyone to take a detour. I received an important phone call while driving and instructed everyone to prepare materials for the afternoon meeting."

[0796] This will allow users to efficiently record voice memos while driving and quickly retrieve important information.

[0797] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0798] Step 1:

[0799] While driving, the user inputs voice memos and instructions via the microphone. The input voice data is saved in real time to the terminal's storage device. Here, the user's voice is recorded as a digital signal.

[0800] Input: User's voice memos or instructions

[0801] Output: Digital audio data stored on the device

[0802] Step 2:

[0803] The terminal sends the stored audio data to the server. The server receives this audio data and stores it temporarily. Here, digital audio data is transferred.

[0804] Input: Digital audio data stored on the device

[0805] Output: Digital audio data stored on the server

[0806] Step 3:

[0807] The server uses the Google Cloud Speech-to-Text API to convert the received audio data into text data. The server sends the audio data to the API and receives and stores the returned text. In this process, the audio data is converted into text information.

[0808] Input: Digital audio data stored on the server

[0809] Output: Text data

[0810] Step 4:

[0811] The server passes the text data to GPT-3.5 for natural language processing. GPT-3.5 parses the text data based on the input prompt sentence, extracts important information, and summarizes it. The summary result is returned to the server. Here, the text data is compressed and converted into a shorter, more concise form of text.

[0812] Input: Text data and prompt text

[0813] Output: Summarized text data

[0814] Step 5:

[0815] The server sends the summarized text data to the terminal. The terminal displays the received summary through its user interface. The user can review the summary on the screen. Here, the final output is provided to the user.

[0816] Input: Summarized text data

[0817] Output: Summary data in a format that can be viewed by the user.

[0818] Step 6:

[0819] If necessary, the device saves the displayed summary data on its own. The user can then refer to and review this data later. Here, the summary data is saved for future reference.

[0820] Input: User-confirmed summary data

[0821] Output: Summary data saved on the device

[0822] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0823] System Overview

[0824] This invention combines a system that receives audio data, converts it into text data, and then summarizes it with an emotion engine that recognizes the user's emotions. Specifically, a user uploads audio data of a meeting to a server, the server analyzes the audio data and converts it into text data, analyzes the emotions using the emotion engine, and finally summarizes the text data and delivers it to the user.

[0825] Specific processing details of the program

[0826] 1. Receiving audio data

[0827] The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated application to select the audio file and click the upload button. This action sends the audio file to the server over the network.

[0828] 2. Saving audio data

[0829] The server saves the received audio files. Uploaded audio files are stored in a temporary storage area for use in subsequent processing.

[0830] 3. Converting audio data to text

[0831] The server passes the stored audio files to the speech recognition engine, which converts them into text data. The speech recognition engine analyzes the audio waveform and generates corresponding character information. This conversion process includes noise reduction and detection of speech boundaries.

[0832] 4. Emotion analysis

[0833] The server uses an emotion recognition engine to analyze the user's emotions from the audio data. It identifies and extracts information about the user's emotions (e.g., joy, anger, sadness) from the audio data and its text. This also allows for understanding the emotional tone of each statement made during a meeting.

[0834] 5. Text Data Summary

[0835] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. If necessary, the results of sentiment analysis are also reflected in the summary.

[0836] 6. Distribution of summaries

[0837] The server distributes the generated summary to the user. The summary results are sent to the email address specified by the user, and the user can check the summary in the received email. It can also be displayed on the system dashboard or in a dedicated application.

[0838] Specific example

[0839] For example, if the following audio data was recorded during a meeting:

[0840] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0841] When this audio data is uploaded to the system, the system processes it as follows:

[0842] 1. Convert audio data to text.

[0843] 2. Analyze the emotions expressed in each statement (for example, the joy felt when discussing the "release of a new product" and the tension felt when discussing the "review of the marketing strategy").

[0844] 3. Summarize the text, and reflect the emotional tone in the summary.

[0845] An example of a generated summary:

[0846] "There was a discussion about the new product release, and most opinions were optimistic. Following that, there was a discussion about reviewing the marketing strategy and expanding the target market, and some sense of urgency was evident. Finally, the date for the next meeting was confirmed."

[0847] By reviewing the emotional tone along with the summarized content, users can quickly grasp the key points of a meeting and their emotional context. This system not only provides an efficient way to obtain important information even when unable to attend a meeting, but also allows users to understand the atmosphere of the meeting and the emotional reactions of the participants.

[0848] The following describes the processing flow.

[0849] Step 1:

[0850] The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated app to select the audio file and clicks the upload button. This action sends the audio file to the server over the network.

[0851] Step 2:

[0852] The server saves the received audio files to a temporary storage area. The saving process also includes checking the format and size of the audio files and converting them to the appropriate format.

[0853] Step 3:

[0854] The server passes the stored audio files to the speech recognition engine, which converts them into text data. In this process, the speech recognition engine analyzes the audio waveform and sequentially generates corresponding text information. Pre-processing such as keyword recognition and noise reduction is also performed.

[0855] Step 4:

[0856] The server passes text data to an emotion recognition engine to analyze the user's emotions. The emotion recognition engine identifies the speaker's emotional state (e.g., joy, anger, sadness) from the text data and audio features, and adds the result to the text data. This emotion information is either added to individual sentences or grouped by segment.

[0857] Step 5:

[0858] The server passes text data, including sentiment data, to a natural language processing engine to generate a summary. The natural language processing engine extracts key keywords and phrases from the text, integrates that information with sentiment data, and creates a concise summary.

[0859] Step 6:

[0860] The server delivers the completed summary to the user. Delivery methods include sending to a specified email address and displaying it on a dedicated dashboard or app. The email contains the summary text, and important sentiment information is highlighted as appropriate.

[0861] (Example 2)

[0862] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0863] Traditional systems that convert audio data into text and summarize it have difficulty overlooking important information and understanding the emotional context of meeting participants. Furthermore, the lack of a means to reflect emotional information in the summary makes it difficult to grasp the overall atmosphere of the meeting.

[0864] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0865] In this invention, the server includes means for receiving audio data, means for storing the received audio data, means for converting the stored audio data into text data, means for analyzing the converted text data to extract emotional information, means for summarizing the text data taking the emotional information into consideration, and means for outputting the summarized text data. This enables accurate summarization that reflects emotional information.

[0866] "Audio data" refers to information recorded in digital or analog format.

[0867] "Text data" refers to information expressed as characters.

[0868] "Means of storage" refers to technology or equipment for temporarily or long-term storage of received data.

[0869] "Means of conversion" refers to a technology or device for converting data in one format to data in another format.

[0870] "Means of analysis" refers to techniques or devices for examining, evaluating, or breaking down data to extract specific information.

[0871] "Emotional information" refers to information that indicates the emotional tone or emotional state contained within the data.

[0872] "Means of summarization" refers to techniques or devices that extract important points from large amounts of data and convert them into a shorter format.

[0873] "Output method" refers to a technology or device that provides processed data in a format that can be read by the user.

[0874] "Natural language processing technology" refers to the technology that enables computers to understand and process natural language, which humans use on a daily basis.

[0875] A "prompt statement" is an input statement for a generative AI model, a text used to induce a specific output.

[0876] System Configuration

[0877] This invention combines a system that receives audio data, converts it to text data, and summarizes it with an emotion engine that recognizes the user's emotions. The user uploads meeting audio data to a server, the server analyzes the audio data, converts it to text data, analyzes the emotions, and finally summarizes the text data and delivers it to the user.

[0878] Hardware and software to be used

[0879] The system of this invention primarily uses the following hardware and software:

[0880] Server: Receives, stores, converts, analyzes, and summarizes audio data.

[0881] Web browser or dedicated application: A terminal for users to upload audio data.

[0882] Speech recognition engine: Uses the Google Cloud Speech-to-Text API to convert speech data into text data.

[0883] Emotion recognition engine: Uses IBM Watson Tone Analyzer to extract emotional information from text data.

[0884] Summarizing text data using the natural language processing engine: OpenAI GPT-3.

[0885] Specific operation of the system

[0886] 1. The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated application to select the audio file and click the upload button. The audio file is sent to the server over the network using an HTTP POST request.

[0887] 2. The server saves the received audio files to a temporary storage area. For example, it stores them in cloud storage.

[0888] 3. The server sends the audio file to the Google Cloud Speech-to-Text API, where the audio data is converted into text data. The speech recognition engine analyzes the audio waveform and generates the corresponding text information.

[0889] 4. The server retrieves the converted text data and passes it to IBM Watson Tone Analyzer to analyze the emotional information. The analysis results are displayed numerically, for example, "Joy: 0.7, Anger: 0.1, Sadness: 0.2".

[0890] 5. The server sends the text data along with a prompt to OpenAI GPT-3 to generate a summary. An example of the prompt to be generated is "Please summarize the contents of the following meeting." + "Acquired text data".

[0891] 6. The server distributes the generated summary to the user. The summary results are sent via email to the email address specified by the user. Users can also view the summary results using the system dashboard or a dedicated application.

[0892] Specific example

[0893] For example, consider a case where the following audio data was recorded during a meeting:

[0894] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[0895] When this audio data is uploaded to the system, it is processed as follows:

[0896] 1. The user uses a dedicated application to upload audio files to the server.

[0897] 2. The server saves the audio files to cloud storage (e.g., Amazon S3).

[0898] 3. The server sends the audio file to the Google Cloud Speech-to-Text API and retrieves the text data "We discussed the release of the new product...".

[0899] 4. The server passes the text data to IBM Watson Tone Analyzer and obtains emotion information: "Joy: 0.7, Anger: 0.1, Sadness: 0.2".

[0900] 5. The server sends the text data to OpenAI GPT-3 along with the prompt "Please summarize the contents of the following meeting." + "Acquired text data," generating the summary "There was a discussion about a new product release, and there were many optimistic opinions..."

[0901] 6. The server sends the generated summary to the user via email and also makes it available on the dashboard.

[0902] In this way, the present invention enables users to quickly grasp important information, including the overall emotional context of a meeting, through efficient analysis of audio data and summarization that reflects emotional information.

[0903] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0904] Step 1:

[0905] The user uploads the meeting audio data to the server. The input is the audio file selected by the user, and the output is the file sent to the server via an HTTP POST request. The user opens a web browser or dedicated application, selects the audio file, and clicks the upload button. This action sends the audio file to the server.

[0906] Step 2:

[0907] The server saves the received audio file to a temporary storage area. The input is the audio file sent by the user, and the output is the file saved in storage. When the server receives the audio file, it saves it to cloud storage (e.g., Amazon S3). This save operation makes it accessible in subsequent processing steps.

[0908] Step 3:

[0909] The server sends the stored audio file to the Google Cloud Speech-to-Text API and converts it into text data. The input is an audio file in cloud storage, and the output is text data obtained from the speech recognition engine. The server sends the audio file via the API, analyzes the audio waveform, and obtains the corresponding text information. Noise reduction and speech segmentation are also performed during this process.

[0910] Step 4:

[0911] The server sends text data to the IBM Watson Tone Analyzer for emotional analysis. The input is text data from the speech recognition engine, and the output is emotional information from the emotional recognition engine. The server passes the acquired text data to the emotional recognition engine to obtain emotional information (e.g., "Joy: 0.7, Anger: 0.1, Sadness: 0.2"). This information is then used in the subsequent summary.

[0912] Step 5:

[0913] The server sends text data along with a prompt to OpenAI GPT-3, which then generates a summary. The input consists of text data and a prompt, and the output is the summary obtained from the generating AI model. An example of a prompt is "Please summarize the contents of the following meeting." + "Acquired text data". The server then generates and saves the summary.

[0914] Step 6:

[0915] The server delivers the generated summary to the user. The input is the generated summary text, and the output is the summary result sent to the user. The server sends the summary result to an email address. Users can also view the summary result on the system dashboard or a dedicated application. Emails are sent using the SMTP protocol, and the dashboard displays the summary result securely using HTTPS.

[0916] (Application Example 2)

[0917] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0918] Systems that extract and summarize important information from audio data face the challenge of understanding the emotional context of meetings and discussions. Conventional technologies cannot extract emotional information from transcribed data, making it difficult to grasp the nuances and emotional shifts of conversations. To solve this problem, it is necessary to simultaneously extract emotional information from audio data and incorporate it into the summary.

[0919] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0920] In this invention, the server includes means for receiving audio data, means for converting the received audio data into text data, means for summarizing the converted text data, means for adding emotional information to the summarized text data, and means for outputting the summarized text data. This allows for the extraction of emotional information along with the summary of audio data, and by reflecting the results in the summary, it becomes possible to simultaneously understand the content of meetings and discussions and their emotional background.

[0921] "Voice data" refers to digital information that digitizes human voice, including conversations, instructions, and other audio files.

[0922] "Means" refers to the methods, apparatus, or techniques used to achieve a particular purpose.

[0923] "Text data" refers to data that is stored or displayed as character information, specifically information obtained by converting audio data into text.

[0924] A "summary" refers to a short and concise version of the content of a text or audio recording, extracting the most important information and making it more compact.

[0925] "Emotional information" refers to data on emotional states analyzed from audio or text, such as information representing emotions like joy, anger, or sadness.

[0926] "Output" refers to the act of displaying, printing, or sending processed data to another system.

[0927] "Receiving" refers to the process of receiving and processing data or information from an external source.

[0928] "Storage" refers to the act of keeping recorded data without losing it, so that it can be retrieved and used later.

[0929] "Analysis" refers to the process of breaking down data and understanding its structure and meaning.

[0930] "Extraction" refers to the act of taking out necessary information from a large amount of data.

[0931] Modes for carrying out the invention

[0932] System Overview

[0933] This invention combines an emotion recognition engine with a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user can upload audio data of a meeting via smart glasses, a server can analyze the audio data and convert it into text data, analyze the emotions using the emotion engine, and finally summarize the text data and deliver it to the user.

[0934] Program Configuration

[0935] The server uses the following hardware and software:

[0936] Hardware: Smart glasses (e.g., Google Glass), built-in microphone, audio interface

[0937] Software: Python, SpeechRecognition library, transformers library

[0938] Program processing flow

[0939] 1. Receiving audio data

[0940] The user uploads the meeting audio data to the server via smart glasses. The audio is recorded using the smart glasses' built-in microphone, and that data is transmitted to the server over the network.

[0941] 2. Saving audio data

[0942] The server saves the received audio file. The saved audio file is stored in a temporary storage area and used for subsequent processing.

[0943] 3. Converting audio data to text

[0944] The server passes the saved audio file to the speech recognition engine, which converts it into text data. The speech recognition engine analyzes the audio waveform and generates corresponding text information. This is done using the SpeechRecognition library.

[0945] 4. Emotion analysis

[0946] The server uses an emotion recognition engine to analyze the user's emotions from the audio data. It identifies the user's emotions (e.g., joy, anger, sadness, etc.) from the audio data and its text, and extracts that information. The transformers library is used for this purpose.

[0947] 5. Text Data Summary

[0948] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. If necessary, the results of sentiment analysis are also reflected in the summary.

[0949] 6. Distribution of summaries

[0950] The server delivers the generated summary to the user. The summary is sent to the user's specified email address, and the user can view the summary via the received email. It can also be displayed on the system dashboard or through a dedicated application.

[0951] Specific example

[0952] For example, if the following audio data was recorded during a meeting:

[0953] "We discussed setting up a new production line. Many suggestions were made to improve efficiency. Afterwards, we also discussed strengthening safety measures. We plan to compile improvement proposals from each department before the next meeting."

[0954] When this audio data is uploaded to the system, the system processes it as follows:

[0955] 1. Convert audio data to text.

[0956] 2. Analyze the emotions expressed in each statement (for example, the joy felt when discussing "launching a new production line" and the tension felt when discussing "strengthening safety measures").

[0957] 3. Summarize the text, and reflect the emotional tone in the summary.

[0958] An example of a generated summary:

[0959] "Many suggestions were made to improve efficiency when setting up the new production line. Strengthening safety measures was also discussed, and improvement proposals are expected to be compiled by the next meeting."

[0960] Example of a prompt

[0961] "Please convert the following audio data into text, analyze the emotions, and summarize it. The audio data includes discussions about launching a new production line and strengthening safety measures."

[0962] In this way, a server-centric system for generating summaries from audio data allows users to efficiently grasp the key points and emotional context of a meeting.

[0963] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0964] Program processing flow

[0965] Step 1:

[0966] The user records audio data using the built-in microphone of the smart glasses. This audio data includes the content of meetings and discussions. After recording is complete, the user uploads the audio data to the server using the application on the smart glasses.

[0967] Input: Audio data recorded with smart glasses

[0968] Output: Audio data sent to the server via the network.

[0969] Step 2:

[0970] The server saves the received audio files to a temporary storage area. This saving process ensures that the audio data is stored properly to prevent loss.

[0971] Input: Received audio data

[0972] Output: Saved audio data

[0973] Step 3:

[0974] The server converts the stored audio data into text data using the SpeechRecognition library. Specifically, it analyzes the audio waveform and converts it into corresponding text information. This process also includes noise reduction and detection of speech boundaries.

[0975] Input: Saved audio data

[0976] Output: Text data

[0977] Step 4:

[0978] The server analyzes the audio data and its accompanying text data to extract emotional information. To achieve this, it uses the transformers library to automatically identify the emotions expressed in the user's statements (e.g., joy, anger, sadness, etc.) and extract that information.

[0979] Input: Text data

[0980] Output: Text data with emotional information added.

[0981] Step 5:

[0982] The server uses a natural language processing engine to summarize text data and sentiment information. It extracts important keywords and phrases and generates a concise summary. The summary also incorporates the results of sentiment analysis.

[0983] Input: Text data with emotional information attached

[0984] Output: Summarized text data

[0985] Step 6:

[0986] The server distributes the generated summary to the user. The summary can be sent to the user's specified email address, or it can be displayed on the system dashboard or a dedicated application.

[0987] Input: Summarized text data

[0988] Output: Summarized text data sent to and delivered to the user.

[0989] Through the above processing steps, users can efficiently grasp the content of the meeting and understand its emotional context.

[0990] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0991] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0992] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0993] [Fourth Embodiment]

[0994] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0995] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[0996] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0997] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0998] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0999] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1000] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1001] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1002] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1003] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1004] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1005] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1006] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1007] System Overview

[1008] This invention relates to a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user can upload audio data of a meeting to a server, the server analyzes the audio data, converts it into text data, and finally summarizes it before delivering it to the user.

[1009] Specific processing details of the program

[1010] 1. Receiving audio data

[1011] The user uploads the meeting audio data to the server. The server temporarily stores the received audio data. This process is initiated when the user selects and uploads the file.

[1012] 2. Text conversion of audio data

[1013] The server converts the received audio data into text data using a speech recognition engine. The speech recognition engine analyzes the audio waveform and generates corresponding character information.

[1014] 3. Text Data Summary

[1015] The server passes the text data to a natural language processing engine, which extracts and summarizes the important information. The natural language processing engine analyzes the meaning of the text and extracts key points.

[1016] 4. Distribution of summaries

[1017] The server sends the generated summary to the user. The user can receive this summary via email or other means.

[1018] Specific example

[1019] For example, if the following audio data was recorded during a meeting:

[1020] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[1021] When this audio data is uploaded to the system, the system generates the following summary:

[1022] "The agenda included discussing the new product release, a new marketing strategy, expanding the target market, and confirming the date for the next meeting."

[1023] By reviewing the summarized content, users can quickly grasp the key points of a meeting. This system provides an efficient way to obtain important information even when unable to attend a meeting.

[1024] The following describes the processing flow.

[1025] Step 1:

[1026] The user uploads the meeting audio file to the server. The user selects the audio file using a web browser or dedicated application and clicks the upload button. This action sends the audio file to the server over the network.

[1027] Step 2:

[1028] The server saves the received audio files. Uploaded audio files are stored in a temporary storage area for subsequent processing. This saving process includes checking the size of the audio file and converting it to an appropriate format.

[1029] Step 3:

[1030] The server passes the stored audio files to the speech recognition engine, where they are converted into text data. Specifically, speech recognition software is used to analyze the audio files and convert the audio signals into corresponding strings of characters. This conversion process includes noise reduction and detection of speech boundaries.

[1031] Step 4:

[1032] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. The summary is compressed to the extent that the meaning of the original text is not lost.

[1033] Step 5:

[1034] The server distributes the generated summary to the user. The summary results are sent to the email address specified by the user, and the user can check the summary in the received email. It can also be displayed on the system dashboard or in a dedicated application.

[1035] (Example 1)

[1036] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1037] Conventional systems were time-consuming to analyze and summarize audio data, making it difficult for users to efficiently obtain information. Furthermore, insufficient temporary storage of audio data and verification of data integrity posed a risk of data reliability degradation. Therefore, there was a need for a means to quickly and accurately obtain important information from meetings and audio recordings.

[1038] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1039] In this invention, the server includes means for uploading audio data to a communication device, means for converting the audio data received by the communication device into text data using an analysis device, means for analyzing the converted text data using a natural language processing device to extract and summarize important information, and means for outputting the summarized text data via the communication device. This enables rapid analysis and accurate summarization of audio data, allowing users to efficiently obtain important information.

[1040] 1. "Audio data" refers to data recorded in digital format from audio information such as meetings and conversations.

[1041] 2. A "communication device" is a device that transmits and receives voice data and summarized text data, and is capable of uploading and downloading data via an internet connection.

[1042] 3. "Analysis device" refers to a device for converting audio data into text data, and includes a speech recognition engine and other analysis software.

[1043] 4. "Text data" refers to character information after it has been analyzed by a speech recognition engine, and is data that represents the content of speech data in text format.

[1044] 5. A "natural language processing device" is a device that analyzes text data, extracts important information, and generates a summary.

[1045] 6. "Summary text data" refers to text data in a concise format, generated by a natural language processing device, from which only the essential information has been extracted from the original text data.

[1046] 7. "Means of temporary storage" refers to means of temporarily storing received audio data in a storage device so that it can be used for subsequent processing.

[1047] 8. "Data integrity verification" is a verification method to confirm that the received data has been transmitted accurately from the sender without any changes.

[1048] Modes for carrying out the invention

[1049] System Overview

[1050] This invention relates to a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user uploads audio data of a meeting to a communication device, a server converts that audio data into text data using an analysis device, summarizes it using a natural language processing device, and delivers it to the user.

[1051] Specific processing details of the program

[1052] 1. Receiving audio data

[1053] Users upload meeting audio data to the server via a communication device. Users use a web application to select and upload audio files. Once the upload is complete, the server temporarily stores the audio data. At this time, a hash value is calculated to verify the integrity of the data.

[1054] 2. Text conversion of audio data

[1055] The server passes the stored audio data to the analysis device. This analysis device uses a speech recognition engine (e.g., Google Speech-to-Text API). The speech recognition engine analyzes the waveform of the audio data and generates corresponding text data. The generated text data is received by the server and temporarily stored.

[1056] 3. Text Data Summary

[1057] The server passes the generated text data to a natural language processing unit (NLP). This NLP uses a generative AI model (e.g., OpenAI GPT-4). The NLP analyzes the text data, extracts important information, and generates a summary. This summary extracts key points from the original text data and is stored secondarily by the server.

[1058] 4. Distribution of summaries

[1059] The server delivers the generated summary text data to the user. Delivery methods include email and notifications via a dedicated portal. Users receive notifications and can click a link to view the summary.

[1060] Specific example

[1061] For example, if the following audio data is recorded during a meeting:

[1062] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[1063] When this audio data is uploaded to the system, the system generates the following summary:

[1064] "The agenda included discussing the new product release, a new marketing strategy, expanding the target market, and confirming the date for the next meeting."

[1065] Example of a prompt

[1066] Examples of prompt statements are as follows:

[1067] Please summarize the following meeting audio transcript:

[1068] At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting.

[1069] By inputting this prompt into the generation AI model, an appropriate summary can be generated.

[1070] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1071] Step 1: Upload audio data

[1072] The user uploads audio data from their terminal to the communication device. As input, the user selects a conference audio file (e.g., MP3 or WAV format) and presses the upload button. As output, the audio file is sent to the communication device. Specifically, the user opens a file selection dialog through their browser, selects the file, and begins sending. The communication device then forwards the received audio file to the server.

[1073] Step 2: Receiving and temporarily storing audio data

[1074] The server receives audio data transmitted from the communication device and stores it in a temporary storage area. The input is audio data sent from the communication device. The output is that audio data stored in the server's temporary storage area. Specifically, the server calculates a hash value (e.g., SHA-256) to verify data integrity and records the file name and hash value in the temporary storage area.

[1075] Step 3: Convert speech to text

[1076] The server passes the audio data stored in the temporary storage area to the analysis device. The input is the stored audio file sent to the analysis device. The output is text data generated by a speech recognition engine (e.g., Google Speech-to-Text API). Specifically, the server sends the audio data as a request to the API and stores the returned text data in the temporary storage area.

[1077] Step 4: Natural language processing of text data

[1078] The server passes text data stored in the primary storage area to a natural language processing unit (e.g., a generative AI model). The generative AI model receives text data as input. The output is summarized text data with important information extracted. Specifically, the server sends text data as a prompt to the generative AI model and stores the generated summary in the secondary storage area.

[1079] Step 5: Distribution of the summary

[1080] The server delivers summarized text data stored in a secondary storage area to the user. The input is the summarized text data passed to the server's notification system. The output is an email notification or display on a web portal to the user. Specifically, the server delivers the summarized text according to the user's settings (e.g., email, push notification). The user receives the notification and clicks the link to view the summary.

[1081] (Application Example 1)

[1082] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1083] In autonomous vehicles, if a driver needs to record important notes or instructions by voice while driving, they must stop the vehicle to take notes. This interferes with driving and reduces efficiency. Furthermore, when reviewing the voice data later, extracting important information from the vast amount of data is time-consuming and hinders quick decision-making. To solve this, a system is needed that transcribes and summarizes voice data in real time, allowing drivers to quickly obtain the information they need.

[1084] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1085] In this invention, the server includes means for receiving voice data, means for converting the received voice data into text data, means for summarizing the converted text data, and means for outputting the summarized text data generated at a terminal installed in the vehicle. This makes it possible for drivers to give instructions or make notes by voice while driving, and for important information to be quickly obtained by converting and summarizing them into text in real time.

[1086] "Audio data" refers to information obtained by converting sound waves into digital signals via microphones or other acoustic sensors.

[1087] "Text data" refers to digital text obtained by analyzing audio data and converting it into corresponding textual information.

[1088] A "summary" is a text document that extracts important information from converted text data and presents it concisely.

[1089] A "terminal installed in a vehicle" refers to a computer or smart device installed inside a vehicle that performs voice recognition and displays and stores transcribed data.

[1090] A "user interface" is a screen, device, or software used by a system to exchange information with a user.

[1091] "Storage" means saving data to a storage medium or database so that it can be accessed later.

[1092] "Analysis" is the process of understanding the content of text data using natural language processing and algorithms, and extracting the necessary information.

[1093] "Reception" refers to acquiring audio data from an external source and incorporating it into the system.

[1094] "Conversion" refers to the process of reconstructing audio data into text data.

[1095] "Output" refers to displaying or sending summarized text data to the user.

[1096] This invention relates to a system that transcribes audio data into text in real time, summarizes the text, and displays it on a terminal inside an autonomous vehicle.

[1097] Hardware and software to be used

[1098] hardware

[1099] Devices installed in automobiles (smartphones and in-car smart devices)

[1100] Microphone for voice input

[1101] software

[1102] Google Cloud Speech-to-Text API

[1103] GPT-3.5

[1104] Python

[1105] Explanation of program processing

[1106] Receiving audio data

[1107] The user inputs voice memos and instructions via the microphone while driving. The device receives this voice data in real time.

[1108] Text conversion of audio data

[1109] The server uses the Google Cloud Speech-to-Text API to convert received audio data into text data. This API provides highly accurate speech recognition capabilities, analyzing the audio waveform to generate corresponding text information.

[1110] Summary of text data

[1111] The server summarizes the converted text data using GPT-3.5. GPT-3.5 is a natural language processing engine that analyzes the meaning of the text, extracts important information, and summarizes it concisely.

[1112] Output and save summary

[1113] The summarized text data is displayed on a terminal installed in the vehicle via the user interface. Additionally, this summarized data can be saved on the terminal as needed for later review by the user.

[1114] Specific example

[1115] For example, if a driver enters the following voice memo while driving:

[1116] "We departed at 10:00 AM and encountered traffic congestion on the way to our destination. I instructed everyone to take a detour. While driving, I received an important phone call and instructed them to prepare materials for the afternoon meeting."

[1117] This audio data is summarized by the system as follows:

[1118] "I instructed them to take a detour due to traffic congestion after departure. I also instructed them to prepare materials for the afternoon meeting via an important phone call."

[1119] Example of a prompt

[1120] Please summarize the following text: "We departed at 10:00 AM and encountered traffic congestion on the way to our destination. I instructed everyone to take a detour. I received an important phone call while driving and instructed everyone to prepare materials for the afternoon meeting."

[1121] This will allow users to efficiently record voice memos while driving and quickly retrieve important information.

[1122] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1123] Step 1:

[1124] While driving, the user inputs voice memos and instructions via the microphone. The input voice data is saved in real time to the terminal's storage device. Here, the user's voice is recorded as a digital signal.

[1125] Input: User's voice memos or instructions

[1126] Output: Digital audio data stored on the device

[1127] Step 2:

[1128] The terminal sends the stored audio data to the server. The server receives this audio data and stores it temporarily. Here, digital audio data is transferred.

[1129] Input: Digital audio data stored on the device

[1130] Output: Digital audio data stored on the server

[1131] Step 3:

[1132] The server uses the Google Cloud Speech-to-Text API to convert the received audio data into text data. The server sends the audio data to the API and receives and stores the returned text. In this process, the audio data is converted into text information.

[1133] Input: Digital audio data stored on the server

[1134] Output: Text data

[1135] Step 4:

[1136] The server passes the text data to GPT-3.5 for natural language processing. GPT-3.5 parses the text data based on the input prompt sentence, extracts important information, and summarizes it. The summary result is returned to the server. Here, the text data is compressed and converted into a shorter, more concise form of text.

[1137] Input: Text data and prompt text

[1138] Output: Summarized text data

[1139] Step 5:

[1140] The server sends the summarized text data to the terminal. The terminal displays the received summary through its user interface. The user can review the summary on the screen. Here, the final output is provided to the user.

[1141] Input: Summarized text data

[1142] Output: Summary data in a format that can be viewed by the user.

[1143] Step 6:

[1144] If necessary, the device saves the displayed summary data on its own. The user can then refer to and review this data later. Here, the summary data is saved for future reference.

[1145] Input: User-confirmed summary data

[1146] Output: Summary data saved on the device

[1147] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1148] System Overview

[1149] This invention combines a system that receives audio data, converts it into text data, and then summarizes it with an emotion engine that recognizes the user's emotions. Specifically, a user uploads audio data of a meeting to a server, the server analyzes the audio data and converts it into text data, analyzes the emotions using the emotion engine, and finally summarizes the text data and delivers it to the user.

[1150] Specific processing details of the program

[1151] 1. Receiving audio data

[1152] The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated application to select the audio file and click the upload button. This action sends the audio file to the server over the network.

[1153] 2. Saving audio data

[1154] The server saves the received audio files. Uploaded audio files are stored in a temporary storage area for use in subsequent processing.

[1155] 3. Converting audio data to text

[1156] The server passes the stored audio files to the speech recognition engine, which converts them into text data. The speech recognition engine analyzes the audio waveform and generates corresponding character information. This conversion process includes noise reduction and detection of speech boundaries.

[1157] 4. Emotion analysis

[1158] The server uses an emotion recognition engine to analyze the user's emotions from the audio data. It identifies and extracts information about the user's emotions (e.g., joy, anger, sadness) from the audio data and its text. This also allows for understanding the emotional tone of each statement made during a meeting.

[1159] 5. Text Data Summary

[1160] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. If necessary, the results of sentiment analysis are also reflected in the summary.

[1161] 6. Distribution of summaries

[1162] The server distributes the generated summary to the user. The summary results are sent to the email address specified by the user, and the user can check the summary in the received email. It can also be displayed on the system dashboard or in a dedicated application.

[1163] Specific example

[1164] For example, if the following audio data was recorded during a meeting:

[1165] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[1166] When this audio data is uploaded to the system, the system processes it as follows:

[1167] 1. Convert audio data to text.

[1168] 2. Analyze the emotions expressed in each statement (for example, the joy felt when discussing the "release of a new product" and the tension felt when discussing the "review of the marketing strategy").

[1169] 3. Summarize the text, and reflect the emotional tone in the summary.

[1170] An example of a generated summary:

[1171] "There was a discussion about the new product release, and most opinions were optimistic. Following that, there was a discussion about reviewing the marketing strategy and expanding the target market, and some sense of urgency was evident. Finally, the date for the next meeting was confirmed."

[1172] By reviewing the emotional tone along with the summarized content, users can quickly grasp the key points of a meeting and their emotional context. This system not only provides an efficient way to obtain important information even when unable to attend a meeting, but also allows users to understand the atmosphere of the meeting and the emotional reactions of the participants.

[1173] The following describes the processing flow.

[1174] Step 1:

[1175] The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated app to select the audio file and clicks the upload button. This action sends the audio file to the server over the network.

[1176] Step 2:

[1177] The server saves the received audio files to a temporary storage area. The saving process also includes checking the format and size of the audio files and converting them to the appropriate format.

[1178] Step 3:

[1179] The server passes the stored audio files to the speech recognition engine, which converts them into text data. In this process, the speech recognition engine analyzes the audio waveform and sequentially generates corresponding text information. Pre-processing such as keyword recognition and noise reduction is also performed.

[1180] Step 4:

[1181] The server passes text data to an emotion recognition engine to analyze the user's emotions. The emotion recognition engine identifies the speaker's emotional state (e.g., joy, anger, sadness) from the text data and audio features, and adds the result to the text data. This emotion information is either added to individual sentences or grouped by segment.

[1182] Step 5:

[1183] The server passes text data, including sentiment data, to a natural language processing engine to generate a summary. The natural language processing engine extracts key keywords and phrases from the text, integrates that information with sentiment data, and creates a concise summary.

[1184] Step 6:

[1185] The server delivers the completed summary to the user. Delivery methods include sending to a specified email address and displaying it on a dedicated dashboard or app. The email contains the summary text, and important sentiment information is highlighted as appropriate.

[1186] (Example 2)

[1187] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1188] Traditional systems that convert audio data into text and summarize it have difficulty overlooking important information and understanding the emotional context of meeting participants. Furthermore, the lack of a means to reflect emotional information in the summary makes it difficult to grasp the overall atmosphere of the meeting.

[1189] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1190] In this invention, the server includes means for receiving audio data, means for storing the received audio data, means for converting the stored audio data into text data, means for analyzing the converted text data to extract emotional information, means for summarizing the text data taking the emotional information into consideration, and means for outputting the summarized text data. This enables accurate summarization that reflects emotional information.

[1191] "Audio data" refers to information recorded in digital or analog format.

[1192] "Text data" refers to information expressed as characters.

[1193] "Means of storage" refers to technology or equipment for temporarily or long-term storage of received data.

[1194] "Means of conversion" refers to a technology or device for converting data in one format to data in another format.

[1195] "Means of analysis" refers to techniques or devices for examining, evaluating, or breaking down data to extract specific information.

[1196] "Emotional information" refers to information that indicates the emotional tone or emotional state contained within the data.

[1197] "Means of summarization" refers to techniques or devices that extract important points from large amounts of data and convert them into a shorter format.

[1198] "Output method" refers to a technology or device that provides processed data in a format that can be read by the user.

[1199] "Natural language processing technology" refers to the technology that enables computers to understand and process natural language, which humans use on a daily basis.

[1200] A "prompt statement" is an input statement for a generative AI model, a text used to induce a specific output.

[1201] System Configuration

[1202] This invention combines a system that receives audio data, converts it to text data, and summarizes it with an emotion engine that recognizes the user's emotions. The user uploads meeting audio data to a server, the server analyzes the audio data, converts it to text data, analyzes the emotions, and finally summarizes the text data and delivers it to the user.

[1203] Hardware and software to be used

[1204] The system of this invention primarily uses the following hardware and software:

[1205] Server: Receives, stores, converts, analyzes, and summarizes audio data.

[1206] Web browser or dedicated application: A terminal for users to upload audio data.

[1207] Speech recognition engine: Uses the Google Cloud Speech-to-Text API to convert speech data into text data.

[1208] Emotion recognition engine: Uses IBM Watson Tone Analyzer to extract emotional information from text data.

[1209] Summarizing text data using the natural language processing engine: OpenAI GPT-3.

[1210] Specific operation of the system

[1211] 1. The user uploads the meeting audio data to the server. The user uses a web browser or a dedicated application to select the audio file and click the upload button. The audio file is sent to the server over the network using an HTTP POST request.

[1212] 2. The server saves the received audio files to a temporary storage area. For example, it stores them in cloud storage.

[1213] 3. The server sends the audio file to the Google Cloud Speech-to-Text API, where the audio data is converted into text data. The speech recognition engine analyzes the audio waveform and generates the corresponding text information.

[1214] 4. The server retrieves the converted text data and passes it to IBM Watson Tone Analyzer to analyze the emotional information. The analysis results are displayed numerically, for example, "Joy: 0.7, Anger: 0.1, Sadness: 0.2".

[1215] 5. The server sends the text data along with a prompt to OpenAI GPT-3 to generate a summary. An example of the prompt to be generated is "Please summarize the contents of the following meeting." + "Acquired text data".

[1216] 6. The server distributes the generated summary to the user. The summary results are sent via email to the email address specified by the user. Users can also view the summary results using the system dashboard or a dedicated application.

[1217] Specific example

[1218] For example, consider a case where the following audio data was recorded during a meeting:

[1219] "At the beginning of the meeting, we discussed the release of the new product. In the next session, we discussed reviewing the marketing strategy and expanding the target market. Finally, we confirmed the date for the next meeting and concluded the meeting."

[1220] When this audio data is uploaded to the system, it is processed as follows:

[1221] 1. The user uses a dedicated application to upload audio files to the server.

[1222] 2. The server saves the audio files to cloud storage (e.g., Amazon S3).

[1223] 3. The server sends the audio file to the Google Cloud Speech-to-Text API and retrieves the text data "We discussed the release of the new product...".

[1224] 4. The server passes the text data to IBM Watson Tone Analyzer and obtains emotion information: "Joy: 0.7, Anger: 0.1, Sadness: 0.2".

[1225] 5. The server sends the text data to OpenAI GPT-3 along with the prompt "Please summarize the contents of the following meeting." + "Acquired text data," generating the summary "There was a discussion about a new product release, and there were many optimistic opinions..."

[1226] 6. The server sends the generated summary to the user via email and also makes it available on the dashboard.

[1227] In this way, the present invention enables users to quickly grasp important information, including the overall emotional context of a meeting, through efficient analysis of audio data and summarization that reflects emotional information.

[1228] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1229] Step 1:

[1230] The user uploads the meeting audio data to the server. The input is the audio file selected by the user, and the output is the file sent to the server via an HTTP POST request. The user opens a web browser or dedicated application, selects the audio file, and clicks the upload button. This action sends the audio file to the server.

[1231] Step 2:

[1232] The server saves the received audio file to a temporary storage area. The input is the audio file sent by the user, and the output is the file saved in storage. When the server receives the audio file, it saves it to cloud storage (e.g., Amazon S3). This save operation makes it accessible in subsequent processing steps.

[1233] Step 3:

[1234] The server sends the stored audio file to the Google Cloud Speech-to-Text API and converts it into text data. The input is an audio file in cloud storage, and the output is text data obtained from the speech recognition engine. The server sends the audio file via the API, analyzes the audio waveform, and obtains the corresponding text information. Noise reduction and speech segmentation are also performed during this process.

[1235] Step 4:

[1236] The server sends text data to the IBM Watson Tone Analyzer for emotional analysis. The input is text data from the speech recognition engine, and the output is emotional information from the emotional recognition engine. The server passes the acquired text data to the emotional recognition engine to obtain emotional information (e.g., "Joy: 0.7, Anger: 0.1, Sadness: 0.2"). This information is then used in the subsequent summary.

[1237] Step 5:

[1238] The server sends text data along with a prompt to OpenAI GPT-3, which then generates a summary. The input consists of text data and a prompt, and the output is the summary obtained from the generating AI model. An example of a prompt is "Please summarize the contents of the following meeting." + "Acquired text data". The server then generates and saves the summary.

[1239] Step 6:

[1240] The server delivers the generated summary to the user. The input is the generated summary text, and the output is the summary result sent to the user. The server sends the summary result to an email address. Users can also view the summary result on the system dashboard or a dedicated application. Emails are sent using the SMTP protocol, and the dashboard displays the summary result securely using HTTPS.

[1241] (Application Example 2)

[1242] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1243] Systems that extract and summarize important information from audio data face the challenge of understanding the emotional context of meetings and discussions. Conventional technologies cannot extract emotional information from transcribed data, making it difficult to grasp the nuances and emotional shifts of conversations. To solve this problem, it is necessary to simultaneously extract emotional information from audio data and incorporate it into the summary.

[1244] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1245] In this invention, the server includes means for receiving audio data, means for converting the received audio data into text data, means for summarizing the converted text data, means for adding emotional information to the summarized text data, and means for outputting the summarized text data. This allows for the extraction of emotional information along with the summary of audio data, and by reflecting the results in the summary, it becomes possible to simultaneously understand the content of meetings and discussions and their emotional background.

[1246] "Voice data" refers to digital information that digitizes human voice, including conversations, instructions, and other audio files.

[1247] "Means" refers to the methods, apparatus, or techniques used to achieve a particular purpose.

[1248] "Text data" refers to data that is stored or displayed as character information, specifically information obtained by converting audio data into text.

[1249] A "summary" refers to a short and concise version of the content of a text or audio recording, extracting the most important information and making it more compact.

[1250] "Emotional information" refers to data on emotional states analyzed from audio or text, such as information representing emotions like joy, anger, or sadness.

[1251] "Output" refers to the act of displaying, printing, or sending processed data to another system.

[1252] "Receiving" refers to the process of receiving and processing data or information from an external source.

[1253] "Storage" refers to the act of keeping recorded data without losing it, so that it can be retrieved and used later.

[1254] "Analysis" refers to the process of breaking down data and understanding its structure and meaning.

[1255] "Extraction" refers to the act of taking out necessary information from a large amount of data.

[1256] Modes for carrying out the invention

[1257] System Overview

[1258] This invention combines an emotion recognition engine with a system that receives audio data, converts it into text data, and then summarizes it. Specifically, a user can upload audio data of a meeting via smart glasses, a server can analyze the audio data and convert it into text data, analyze the emotions using the emotion engine, and finally summarize the text data and deliver it to the user.

[1259] Program Configuration

[1260] The server uses the following hardware and software:

[1261] Hardware: Smart glasses (e.g., Google Glass), built-in microphone, audio interface

[1262] Software: Python, SpeechRecognition library, transformers library

[1263] Program processing flow

[1264] 1. Receiving audio data

[1265] The user uploads the meeting audio data to the server via smart glasses. The audio is recorded using the smart glasses' built-in microphone, and that data is transmitted to the server over the network.

[1266] 2. Saving audio data

[1267] The server saves the received audio file. The saved audio file is stored in a temporary storage area and used for subsequent processing.

[1268] 3. Converting audio data to text

[1269] The server passes the saved audio file to the speech recognition engine, which converts it into text data. The speech recognition engine analyzes the audio waveform and generates corresponding text information. This is done using the SpeechRecognition library.

[1270] 4. Emotion analysis

[1271] The server uses an emotion recognition engine to analyze the user's emotions from the audio data. It identifies the user's emotions (e.g., joy, anger, sadness, etc.) from the audio data and its text, and extracts that information. The transformers library is used for this purpose.

[1272] 5. Text Data Summary

[1273] The server passes the converted text data to a natural language processing engine for analysis and generates a summary. The natural language processing engine extracts important keywords and phrases from the text and uses them to create a summary. If necessary, the results of sentiment analysis are also reflected in the summary.

[1274] 6. Distribution of summaries

[1275] The server delivers the generated summary to the user. The summary is sent to the user's specified email address, and the user can view the summary via the received email. It can also be displayed on the system dashboard or through a dedicated application.

[1276] Specific example

[1277] For example, if the following audio data was recorded during a meeting:

[1278] "We discussed setting up a new production line. Many suggestions were made to improve efficiency. Afterwards, we also discussed strengthening safety measures. We plan to compile improvement proposals from each department before the next meeting."

[1279] When this audio data is uploaded to the system, the system processes it as follows:

[1280] 1. Convert audio data to text.

[1281] 2. Analyze the emotions expressed in each statement (for example, the joy felt when discussing "launching a new production line" and the tension felt when discussing "strengthening safety measures").

[1282] 3. Summarize the text, and reflect the emotional tone in the summary.

[1283] An example of a generated summary:

[1284] "Many suggestions were made to improve efficiency when setting up the new production line. Strengthening safety measures was also discussed, and improvement proposals are expected to be compiled by the next meeting."

[1285] Example of a prompt

[1286] "Please convert the following audio data into text, analyze the emotions, and summarize it. The audio data includes discussions about launching a new production line and strengthening safety measures."

[1287] In this way, a server-centric system for generating summaries from audio data allows users to efficiently grasp the key points and emotional context of a meeting.

[1288] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1289] Program processing flow

[1290] Step 1:

[1291] The user records audio data using the built-in microphone of the smart glasses. This audio data includes the content of meetings and discussions. After recording is complete, the user uploads the audio data to the server using the application on the smart glasses.

[1292] Input: Audio data recorded with smart glasses

[1293] Output: Audio data sent to the server via the network.

[1294] Step 2:

[1295] The server saves the received audio files to a temporary storage area. This saving process ensures that the audio data is stored properly to prevent loss.

[1296] Input: Received audio data

[1297] Output: Saved audio data

[1298] Step 3:

[1299] The server converts the stored audio data into text data using the SpeechRecognition library. Specifically, it analyzes the audio waveform and converts it into corresponding text information. This process also includes noise reduction and detection of speech boundaries.

[1300] Input: Saved audio data

[1301] Output: Text data

[1302] Step 4:

[1303] The server analyzes the audio data and its accompanying text data to extract emotional information. To achieve this, it uses the transformers library to automatically identify the emotions expressed in the user's statements (e.g., joy, anger, sadness, etc.) and extract that information.

[1304] Input: Text data

[1305] Output: Text data with emotional information added.

[1306] Step 5:

[1307] The server uses a natural language processing engine to summarize text data and sentiment information. It extracts important keywords and phrases and generates a concise summary. The summary also incorporates the results of sentiment analysis.

[1308] Input: Text data with emotional information attached

[1309] Output: Summarized text data

[1310] Step 6:

[1311] The server distributes the generated summary to the user. The summary can be sent to the user's specified email address, or it can be displayed on the system dashboard or a dedicated application.

[1312] Input: Summarized text data

[1313] Output: Summarized text data sent to and delivered to the user.

[1314] Through the above processing steps, users can efficiently grasp the content of the meeting and understand its emotional context.

[1315] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1316] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1317] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1318] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1319] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1320] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1321] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1322] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1323] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1324] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1325] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1326] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1327] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1328] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1329] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1330] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1331] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1332] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1333] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1334] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1335] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[1336] The following is further disclosed regarding the embodiments described above.

[1337] (Claim 1)

[1338] A means of receiving audio data,

[1339] A means of converting received audio data into text data,

[1340] A means of summarizing the converted text data,

[1341] A means of outputting summarized text data,

[1342] A system that includes this.

[1343] (Claim 2)

[1344] The system according to claim 1, further comprising means for storing received audio data.

[1345] (Claim 3)

[1346] The system according to claim 1, further comprising means for extracting and summarizing important information by analyzing the converted text data.

[1347] "Example 1"

[1348] (Claim 1)

[1349] A means of uploading audio data to a communication device,

[1350] A means for converting audio data received by a communication device into text data using an analysis device,

[1351] A means for analyzing the converted text data using a natural language processing device to extract and summarize important information,

[1352] A means for outputting summarized text data via a communication device,

[1353] A system that includes this.

[1354] (Claim 2)

[1355] The system according to claim 1, further comprising means for temporarily storing received audio data.

[1356] (Claim 3)

[1357] The system according to claim 1, comprising a natural language processing device that analyzes the converted text data to extract and summarize important information.

[1358] "Application Example 1"

[1359] (Claim 1)

[1360] A means of receiving audio data,

[1361] A means of converting received audio data into text data,

[1362] A means of summarizing the converted text data,

[1363] A means for outputting summarized text data generated on a terminal installed in a vehicle,

[1364] A system that includes this.

[1365] (Claim 2)

[1366] The system according to claim 1, further comprising means for storing received audio data.

[1367] (Claim 3)

[1368] The system according to claim 1, further comprising means for extracting and summarizing important information by analyzing the converted text data.

[1369] (Claim 4)

[1370] The system according to claim 1, which displays the generated summary through a user interface.

[1371] (Claim 5)

[1372] A system according to claim 1 for storing a generated summary.

[1373] "Example 2 of combining an emotion engine"

[1374] (Claim 1)

[1375] A means of receiving audio data,

[1376] A means of saving the received audio data,

[1377] A means of converting saved audio data into text data,

[1378] A method for analyzing converted text data and extracting emotional information,

[1379] A method for summarizing text data while taking emotional information into consideration,

[1380] A means of outputting summarized text data,

[1381] A system that includes this.

[1382] (Claim 2)

[1383] The system according to claim 1, characterized in that it summarizes the converted text data using natural language processing technology.

[1384] (Claim 3)

[1385] The system according to claim 1, characterized in that emotional information is reflected in the summary.

[1386] "Application example 2 when combining with an emotional engine"

[1387] (Claim 1)

[1388] A means of receiving audio data,

[1389] A means of converting received audio data into text data,

[1390] A means of summarizing the converted text data,

[1391] A means of adding emotional information to summarized text data,

[1392] A means of outputting summarized text data,

[1393] A system that includes this.

[1394] (Claim 2)

[1395] The system according to claim 1, further comprising means for storing received audio data.

[1396] (Claim 3)

[1397] The system according to claim 1, further comprising means for extracting and summarizing important information by analyzing the converted text data. [Explanation of Symbols]

[1398] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of receiving audio data, A means of converting received audio data into text data, A means of summarizing the converted text data, A means of outputting summarized text data, A system that includes this.

2. The system according to claim 1, further comprising means for storing received audio data.

3. The system according to claim 1, further comprising means for extracting and summarizing important information by analyzing the converted text data.

Images