system

Abstract

We provide the system. [Solution] A voice acquisition means that receives the user's voice in real time and sends it to a server, A speech recognition means for converting speech data transmitted from the speech acquisition means into text, An analytical means for analyzing the converted text and detecting keywords or phrases that indicate fraud or scams, A warning and question provision means that notifies the user of a warning based on detected keywords or phrases and provides more specific question suggestions, A feedback receiving means having an interface that allows users to provide feedback, 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, including 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 modern society, fraud and illegal business activities are becoming increasingly sophisticated, especially causing serious harm to the elderly and general consumers. In such a situation, effective means for protecting users from these risks are required. However, it is difficult to detect fraud in real time with conventional methods, and there is a possibility that users may suffer damage before they recognize the illegal act. Therefore, there is a need for a new method to immediately determine whether the conversation that a user is currently having is fraudulent and provide appropriate warnings and countermeasures.

Means for Solving the Problems

[0005] This invention provides a system for monitoring user conversations in real time. This system collects user voice data using voice acquisition means and transmits the voice data to a server. The server converts the voice to text using speech recognition means and detects keywords and phrases indicating fraud or wrongdoing through analysis means. After detection, warning and question provision means notify the user of an alert and present questions to help them confirm the fraud. Furthermore, feedback receiving means receive feedback from the user and uses it to improve the system's accuracy. In this way, real-time fraud detection and user support are achieved, making it possible to prevent damage before it occurs.

[0006] A "user" is an individual or organization that uses the system and provides voice data.

[0007] "Voice acquisition means" refers to a device or program that has the function of collecting the user's voice in real time and transferring it as data to a server.

[0008] "Speech recognition means" refers to software or a system for analyzing speech data and converting speech into corresponding text data.

[0009] "Analysis means" refers to technology or software for processing text data and detecting keywords or phrases that indicate fraud or scams.

[0010] "Warning and Question Provisioning Means" refers to a mechanism or system for efficiently issuing warnings to users and providing appropriate questions based on detected potential fraud.

[0011] A "feedback receiving mechanism" refers to a function or interface used to receive opinions and information from users and to improve or adjust the system.

[0012] A "server" is a central processing unit or computer network system that receives voice data and performs speech recognition and analysis. [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, which incorporates an emotion engine. [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] Hereinafter, an example of an embodiment of the system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

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

[0016] In the following embodiments, a numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of a plurality of arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of a plurality of 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 and various parameters, etc. 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] The system of the present invention has the function of monitoring a user's voice conversation in real time and detecting and issuing warnings for fraudulent or inappropriate behavior. This system is realized through the following main components and their operation.

[0035] First, when a user begins speaking through the device, the device uses voice acquisition tools to collect voice data in real time. This voice data is immediately sent to the server.

[0036] The server accurately converts the audio data received through the speech recognition means into text, and the text information is processed by the analysis means. The analysis means uses natural language processing techniques to analyze the text data in detail and detect words and phrases that may contain fraudulent or malicious elements.

[0037] If detection occurs, the server will utilize warning and question provision mechanisms to display a warning message to the user and provide the user with specific questions that can be used to verify the veracity of the other party's statements.

[0038] The device uses this information to immediately notify the user of any potential risks. The user can then choose to continue the conversation or interrupt it by following the instructions received.

[0039] Furthermore, the system includes a means for receiving user feedback, and the feedback provided by users is used to improve the system's accuracy and functionality.

[0040] As a concrete example, consider a scenario where a user is told over the phone, "We need to verify your bank information, so please tell us your account number." In this case, the system detects keywords such as "bank" and "account number," issues a warning to the user, and further asks questions such as, "Which bank is calling from?" to help reduce the likelihood of becoming a victim of fraud.

[0041] In this way, the system of the present invention provides an effective means for protecting users from fraud and illegal activities.

[0042] The following describes the processing flow.

[0043] Step 1:

[0044] The device acquires the user's voice in real time through the microphone. This voice data is converted into packets at regular time intervals (for example, every few seconds) and sent to the server.

[0045] Step 2:

[0046] When the server receives audio data, it converts the audio into text data via speech recognition. This conversion process also removes background noise and corrects for differences in accent and intonation.

[0047] Step 3:

[0048] The server processes the converted text data using analysis tools. During this analysis process, natural language processing techniques are used to detect contextually relevant keywords and phrases, identifying elements that may be fraudulent or malicious.

[0049] Step 4:

[0050] If detection occurs, the server generates a warning message for the user using warning and question provision mechanisms. At the same time, it prepares to formulate and provide the user with specific and easy-to-understand questions.

[0051] Step 5:

[0052] The terminal receives warning messages and suggested questions sent from the server and immediately notifies the user through a visual or auditory interface.

[0053] Step 6:

[0054] Users can use the received warnings and suggested questions to decide whether to continue or discontinue the conversation with the other party. Users can also use the suggested questions to verify the other party's identity and intentions.

[0055] Step 7:

[0056] When a user provides feedback, the device collects that information and sends it to the server. The server then analyzes the received feedback and uses it to improve the system and increase its accuracy.

[0057] (Example 1)

[0058] 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."

[0059] In modern society, fraudulent activities and scams targeting users are on the rise, creating a need for real-time monitoring methods to effectively address them. However, current methods make it difficult to immediately warn users and provide concrete countermeasures. Furthermore, continuous system improvement utilizing feedback is not being adequately implemented. These challenges need to be addressed.

[0060] 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.

[0061] In this invention, the server includes voice acquisition means, voice recognition means, analysis means, warning and inquiry suggestion provision means, evaluation receiving means, and means for improving analysis accuracy and warning quality using a generated AI model. This enables real-time protection of users from fraud and misconduct, immediate provision of countermeasures, and continuous improvement of the system based on feedback.

[0062] "Voice acquisition means" refers to a device or method that acquires and transmits voice information from a user in real time.

[0063] "Speech recognition means" refers to a technology or process for converting speech information into symbolic information.

[0064] "Analysis means" refers to a technique or process that analyzes symbolic information to detect identifiers that indicate fraud or fraudulent activity.

[0065] "Warning and Inquiry Provision Methods" refers to technologies or processes that warn users of potential fraudulent activity and provide specific methods for making inquiries.

[0066] "Evaluation receiving means" refers to an interface or method for receiving evaluations from users.

[0067] A "generative AI model" is an artificial intelligence technology used to improve analysis accuracy and the quality of warnings based on feedback.

[0068] The system of this invention is designed to protect users from fraud and illegal activities by analyzing voice information in real time and providing immediate countermeasures. The specific implementation of the system is described below.

[0069] The terminal is equipped with a microphone and voice input device, and when the user begins to speak, voice data is collected in real time through these voice acquisition means. This voice data is transmitted to the server using a secure communication protocol. The communication means incorporate commonly used encryption technologies.

[0070] The server uses a speech recognition engine (e.g., a commercial speech recognition API) to convert the received audio data into text. At this stage, a high-precision language model is utilized to remove noise and support multiple languages. The converted text data is then analyzed in detail using natural language processing techniques. This process utilizes Python's NLTK and spaCy libraries to detect keywords and phrases that could potentially lead to fraud or misconduct.

[0071] If any malicious elements are detected, the server uses an AI model to generate a warning message. This message includes an immediate alert to the user and specific questions to help them respond in the conversation. The questions provided may include those that allow the user to verify the legitimacy of the other party, such as, "Which organization is contacting you?"

[0072] Based on the warning messages and suggested questions displayed on the device, users can choose the appropriate course of action to take with the other party. This information is communicated visually and audibly, allowing users to respond immediately.

[0073] In addition, the terminal is equipped with an interface for receiving user feedback. This feedback is sent to a server and used by an AI model to improve the system's accuracy. For example, prompts such as "Suggest improvements to the warning system based on recent fraud patterns" are input to the AI ​​to improve the system.

[0074] In this way, the system of the present invention provides a means to guide the user into a safe communication environment.

[0075] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0076] Step 1:

[0077] The terminal collects the user's speech in real time using a speech acquisition device. The input data is the user's voice, which the terminal converts into a digital audio signal and immediately sends to the server. The device used at this stage is the microphone built into the terminal. The output is digital audio data.

[0078] Step 2:

[0079] The server converts transmitted audio data into text data using a speech recognition engine. The input is digital audio data, and the server applies speech recognition technology to generate corresponding text from the audio signal through a language model. The output is the converted text data. This process involves background noise removal and accurate pronunciation recognition.

[0080] Step 3:

[0081] The server applies natural language processing techniques to analyze text data and detect specific keywords and phrases. The input is text data, and the data analysis process considers context and intent. It utilizes Python's NLTK and spaCy libraries to identify elements that indicate fraud or misconduct. The output consists of the detected fraudulent keywords and phrases.

[0082] Step 4:

[0083] The server generates warning messages and specific questions based on the detection results. The input is identified keywords or phrases, and a generative AI model is used to design appropriate warnings and questions. The warning message serves as a reminder to the user, and the questions help verify the legitimacy of the other party. The output is the warning message and questions to be displayed to the user.

[0084] Step 5:

[0085] The terminal receives warning messages and suggested questions from the server and notifies the user visually and audibly. Input is message data from the server, which the terminal presents to the user through its interface. Based on this, the user can select an appropriate response. Output is screen displays and audio notifications that the user can see.

[0086] Step 6:

[0087] Based on the information provided, the user decides whether to continue or interrupt the conversation and provides feedback to the system via the terminal. This feedback, containing information about the user's choices and actions, is sent to the server. The system receives this feedback and uses a generative AI model to improve the system's analysis accuracy and the quality of its warnings. The output represents the improved system performance.

[0088] (Application Example 1)

[0089] 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."

[0090] In modern society, fraud and dishonest practices are commonplace, increasing the risk of users becoming victims. In particular, detecting fraudulent behavior in advance and responding quickly is difficult in voice communication. Current technology is insufficient to detect fraudulent activity in real time and issue immediate warnings to users; therefore, an efficient solution is needed to ensure user safety.

[0091] 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.

[0092] In this invention, the server includes voice data acquisition means for receiving user voice in real time and transmitting it to the server, voice recognition means for converting the voice information transmitted from the voice data acquisition means into text, and analysis means for analyzing the converted text and detecting words or expressions that indicate fraud or misconduct. This enables early detection of fraudulent activity and rapid notification of warnings to the user.

[0093] A "user" is an individual or legal entity that uses the system to conduct voice communication.

[0094] "Voice data acquisition means" refers to devices or software that receive the user's voice in real time and transmit it to a server.

[0095] A "server" is a computer system used for data processing and communication.

[0096] "Speech recognition means" refers to a technology or device for converting speech data into text.

[0097] "Analysis means" refers to techniques or devices for analyzing converted text and detecting words or expressions that may indicate fraud or misconduct.

[0098] A "warning and question supply means" is a means of notifying the user of an alert based on detected words or expressions and providing specific examples of questions.

[0099] A "feedback receiving means" is a device or system equipped with an interface that allows users to provide comments.

[0100] A "cloud environment" is an environment that utilizes computing resources provided over the internet.

[0101] This invention is a system that protects users by detecting fraudulent or illegal behavior in real time and issuing warnings when users engage in voice communication using devices such as smartphones. Specific embodiments of this system are shown below.

[0102] When a user initiates a call via their smartphone, the audio is received in real time by an audio data acquisition system. The received audio is immediately transmitted to the server. At this stage, the smartphone uses a communication module to maintain a stable connection.

[0103] The server instantly converts audio data into text using speech recognition tools such as Google® Cloud Speech-to-Text. This allows for real-time transcription of voice communication. The converted text is then analyzed by natural language processing tools (e.g., SpaCy and BERT). In this analysis, the text is scrutinized for potentially fraudulent or deceptive words and phrases based on its context.

[0104] If the analysis detects fraudulent content, the server will issue visual and auditory warnings to the user's smartphone using warning and question-providing mechanisms. At the same time, it will provide the user with specific questions, such as "Which organization is calling from?", to help them identify suspicious points in the conversation.

[0105] Users can provide feedback to the system through feedback receiving mechanisms. This feedback is stored and analyzed on the server for continuous improvement of the system. The data for improvement is accumulated in a cloud environment and used to train the generated AI model.

[0106] For example, if a user calls customer support for an online shopping site and is asked for their credit card information, the system will detect this phrase and display a warning to the user saying, "We recommend using the official website." This helps protect the user from unintentional fraud.

[0107] An example of a prompt message is: "Identify the phrase indicating fraudulent activity from the following phone conversation: 'An error has occurred in our system. We need to verify your credit card information immediately.'"

[0108] In this way, this invention provides a means to prevent fraudulent activities in users' voice communications.

[0109] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0110] Step 1:

[0111] The terminal receives audio in real time using an audio data acquisition mechanism as soon as the user initiates a call. The input is an audio signal, and the output is digitized audio data. The audio data is immediately transmitted to the server.

[0112] Step 2:

[0113] The server converts received digital audio data into text using speech recognition technology. The input is digital audio data, and the output is text data. Here, the Google Cloud Speech-to-Text service is used to perform the audio-to-text conversion.

[0114] Step 3:

[0115] The server analyzes the converted text data using parsing tools. The input is text data, and the output is a flag indicating whether the text indicates fraudulent or illegal behavior. SpaCy and BERT are used for the analysis, and natural language processing techniques are employed to identify words and phrases.

[0116] Step 4:

[0117] Based on the analysis results, the server uses warning and question-providing mechanisms to send visual and auditory warnings to the terminal. The input is a flag indicating inappropriate behavior, and the output is a warning message and specific questions for the user. This results in a notification being displayed and played on the user's device.

[0118] Step 5:

[0119] The user chooses whether to continue or interrupt the conversation based on the provided warnings and suggested questions. The input consists of warning messages and suggested questions, while the output is the action chosen by the user. The user can ask additional questions to ensure safety.

[0120] Step 6:

[0121] The terminal sends feedback collected from users to a server via a feedback receiving mechanism. The input is feedback information provided by the user, and the output is the cloud storage of the feedback data. This data will be used for future system improvements.

[0122] Step 7:

[0123] The server analyzes feedback data to train and improve the AI ​​model generated in the cloud environment. The input is feedback data, and the output is an updated model for system improvement. The analysis results will be used to adjust the model and improve prompt messages in the future.

[0124] 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.

[0125] The system of the present invention not only monitors user conversations in real time and detects signs of fraud or misconduct, but also recognizes the user's emotional state and adjusts its response accordingly. This system is implemented with the following components:

[0126] First, when a user initiates a conversation via the device, the device uses voice acquisition means to collect the user's voice data in real time and sends the data to the server.

[0127] The server uses speech recognition to convert the audio data into text. During this process, an emotion engine analyzes the audio data and uses factors such as tone of voice, speaking speed, and emphasis patterns to identify the user's emotional state.

[0128] The converted text data is processed by an analysis tool to detect keywords or phrases indicating fraud or misleading expressions. This detection takes into account emotional state information from an emotion engine, enabling more contextually accurate analysis.

[0129] Based on the detection results, the server generates warnings and suggested questions for the user using warning and question provision mechanisms. At this time, based on information from the emotion engine, notifications are made in a tone and content that matches the user's current emotional state.

[0130] The device receives the generated warning messages and suggested questions and notifies the user visually or audibly. The user can then use this information to make informed decisions.

[0131] Furthermore, by using emotion engine data when users provide feedback, the emotional nuances of the feedback are also taken into consideration, which can be used to improve the system.

[0132] For example, if a user receives a suspicious phone call and is asked, "Would you like to receive a limited offer?", the system will detect the keyword "limited offer," which suggests potential fraud. If the emotion engine senses tension in the user's voice, it will respond by providing a gentle warning such as, "Please relax and think about it. Please check specifically what the offer is." In this way, the system can not only effectively protect users from fraud but also provide emotional support.

[0133] The following describes the processing flow.

[0134] Step 1:

[0135] The device acquires the user's voice in real time via the microphone and prepares to send this audio data to the server. During this process, noise reduction is also performed to maintain audio clarity.

[0136] Step 2:

[0137] The server converts the received audio data into text format using speech recognition technology. This conversion uses a language model and focuses on accurately transcribing the user's utterances into text.

[0138] Step 3:

[0139] The server analyzes the converted text data using analytical tools. This analysis detects keywords and phrases that indicate fraud or scams. Natural language processing techniques are used to consider the context and aim for higher accuracy in detection.

[0140] Step 4:

[0141] Simultaneously, the server activates an emotion engine, analyzing the tone, speed, and pacing of the audio data to determine the user's emotional state. The analysis results indicate the user's current emotions and influence the next steps.

[0142] Step 5:

[0143] The server generates warnings and suggested questions based on information from the emotion engine. This generation process takes into account the user's emotional state and includes reassuring language and suggestions to support the user.

[0144] Step 6:

[0145] The terminal receives warning messages and suggested questions sent from the server and presents them to the user through visual and auditory interfaces. For example, it may display text on the screen while simultaneously providing important information via audio.

[0146] Step 7:

[0147] Based on the information presented, users can decide whether to continue or end the conversation with the other party. They can also report their thoughts and suggestions for improvement through the system's feedback function as needed.

[0148] Step 8:

[0149] The server collects user feedback and uses it as data to improve the system's analytical capabilities and user interface. This feedback, in particular, contributes to improving the accuracy of the emotion engine.

[0150] (Example 2)

[0151] 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".

[0152] While it is crucial to detect signs of fraud and misconduct early in user conversations, traditional systems often issue warnings without considering the user's emotional state, potentially leading to misunderstandings and stress. Therefore, there is a need to more effectively protect users by providing appropriate warnings and feedback that are tailored to their emotional state.

[0153] 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.

[0154] In this invention, the server includes speech recognition means, emotion analysis means, analysis means, warning and question provision means, and feedback receiving means. This enables the provision of accurate warnings and questions that take into account the user's emotional state, thereby more effectively protecting the user and facilitating system improvement using feedback.

[0155] A "voice acquisition means" is a device that receives the user's voice in real time and transmits it to an information processing device.

[0156] "Speech recognition means" refers to a technology or device that analyzes received speech data and converts it into text data.

[0157] "Emotional analysis means" refers to a technology or device that analyzes voice data and uses voice tone, speaking speed, and emphasis patterns to identify the user's emotional state.

[0158] "Analysis means" refers to a technology or device that analyzes text data, detects keywords or phrases indicating fraud or scams, and performs contextual analysis.

[0159] "Warning and Question Provisioning Means" refers to a technology or device that notifies the user of a customized warning and provides specific questions based on detected keywords or phrases and emotional states.

[0160] "Feedback receiving means" refers to a technology or device that has an interface for users to provide feedback and to receive that feedback while taking emotional nuances into consideration.

[0161] The system of this invention not only analyzes user conversations from audio to detect signs of fraud or dishonest behavior, but also has the function of understanding the user's emotional state and providing appropriate responses. The specific configuration of the system includes a terminal, a server, and various analysis engines.

[0162] When a user begins a conversation, the device collects voice data in real time through its built-in microphone and voice acquisition software. This voice data is compressed and packetized as initial processing before being sent to the server. The server has powerful processing capabilities and utilizes common speech recognition APIs, particularly for speech recognition. Specific examples include Google Cloud Speech-to-Text and Amazon Transcribe.

[0163] The server converts the received audio data into text data, then uses an emotion analysis engine to analyze the tone, volume, and speaking speed of the voice to identify the user's emotional state. This analysis engine is specialized in extracting emotions from speech using machine learning models and specific algorithms.

[0164] The converted text data is analyzed using natural language processing (NLP) techniques to detect keywords and phrases that indicate tendencies toward fraudulent activity or deception. This analysis also considers emotional state information, enabling more accurate contextual analysis. Based on the analysis results, the server generates warning messages and suggested questions tailored to the user's emotional state. This process utilizes a generative AI model to generate user-specific responses.

[0165] The generated warning messages and proposed questions are sent to the terminal and notified to the user visually or audibly. The user can then make their own decisions based on this notification. Furthermore, user feedback is sent to the server via the terminal, where sentiment analysis is also performed. This feedback is used for the continuous improvement of the system.

[0166] For example, if a user receives a sales call and is asked, "Would you like to receive a limited offer?", the system will detect the keyword "limited offer." At the same time, if it detects tension in the user's voice, it will provide a warning such as, "Please relax and think carefully about what you're going to say. We recommend that you clearly confirm the details." In this way, the system ensures the user's safety while also providing emotional support.

[0167] An example of a prompt message would be, "How should the system generate a warning if a user receives a potentially fraudulent phone call?"

[0168] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0169] Step 1:

[0170] The user uses a device to initiate a conversation. The device collects the conversation in real time using a voice acquisition device. The input is the user's voice, and voice data is generated as output. This voice data undergoes noise reduction and echo cancellation through digital signal processing before being sent to the server.

[0171] Step 2:

[0172] The server passes the audio data received from the terminal to the speech recognition system. The input is audio data, which is converted into text data as output. During this conversion process, speech recognition algorithms are applied to recognize phonemes and words. Techniques used for the conversion include phoneme decomposition and hidden Markov models.

[0173] Step 3:

[0174] The text data is analyzed by sentiment analysis tools on the server. The input is the text data generated in step 2, and the output is information indicating the user's emotional state. In this process, the analysis engine evaluates the voice tone, speaking speed, and emphasis patterns to determine the user's emotions. For example, a machine learning model using an annotated dataset can predict emotions.

[0175] Step 4:

[0176] The server further analyzes the text data using analysis tools. The input is the sentiment information and text data from step 3, and the output is the detection results of keywords and phrases indicating signs of fraud or wrongdoing. This process uses natural language processing techniques to scan the data and perform dictionary matching and contextual analysis. Sentiment information improves contextual understanding.

[0177] Step 5:

[0178] The server generates appropriate responses to the user using warning and question-providing mechanisms. The input is the detection results and emotional state obtained in step 4, and the output is a customized warning message and question. This generative AI model enables communication in an emotionally sensitive tone. For example, it can provide voice messages in a gentle tone.

[0179] Step 6:

[0180] The terminal visually or audibly notifies the user of warning messages and suggested questions received from the server. Input is the generated response message, and output is the notification to the user. The terminal communicates information to the user through screen display and speech synthesis. Notifications are provided in an easy-to-understand format using a user interface.

[0181] Step 7:

[0182] Users provide feedback to the system through a feedback receiving mechanism. The input is the feedback information delivered by the user, and the output is feedback data that contributes to system improvement. The server analyzes this feedback along with sentiment analysis and uses the findings to improve the system's accuracy. The submitted feedback is stored in a database.

[0183] (Application Example 2)

[0184] 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".

[0185] Fraud and dishonest behavior are significant issues in daily life, and it is essential for users to be able to detect these behaviors during conversations and take appropriate action. However, conventional technologies do not provide warnings or instructions that take into account the user's emotional state, which can cause stress. Therefore, the present invention aims to provide a system that takes the user's emotional state into consideration and provides more accurate and appropriate warnings and support.

[0186] 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.

[0187] In this invention, the server includes voice data collection means, voice data analysis means, data analysis means, notification and question provision means, opinion receiving means, and sentiment analysis means. This makes it possible to detect fraudulent activity and provide warnings in an appropriate tone while taking into account the user's emotional state in real time.

[0188] "Voice data collection means" refers to means for acquiring the user's voice in real time and transmitting it to a processing device.

[0189] "Voice data analysis means" refers to means for converting acquired voice information into text data.

[0190] "Data analysis means" refers to means for analyzing converted text data to detect terms or phrases that may indicate fraudulent or dishonest activity.

[0191] "Notification and Question Provisioning Means" are means of alerting users and providing more detailed questions based on detected terms or phrases.

[0192] A "means for receiving opinions" refers to a means equipped with a user interaction interface that allows users to submit opinions.

[0193] "Emotion analysis means" refers to a means for detecting the user's emotional state and adjusting the tone of notification messages based on that state.

[0194] The system for realizing this invention consists of a voice data collection means, a voice data analysis means, a data analysis means, a notification and question provision means, an opinion receiving means, and an emotion analysis means. The operation of the system is as follows.

[0195] First, the voice spoken by the user is collected as audio data using the microphone built into the device, and the audio data collection means transmits it to the processing unit. The audio data analysis means converts this audio data into text data using speech recognition software such as the Google Speech-to-Text API. At this time, the server processes the converted text data using the data analysis means and detects terms and phrases that suggest the possibility of fraudulent or dishonest activity using natural language processing technology.

[0196] Next, the server uses sentiment analysis tools to analyze the user's emotional state using software such as Microsoft® Azure®'s Text Analytics. This allows warning messages related to detected terms to be adjusted based on the emotional state. The adjusted messages are then visually and audibly communicated to the user through the terminal's display and speakers via notification and question-providing tools.

[0197] Users can provide feedback through various means of receiving it. This feedback will be used to improve the system's algorithms.

[0198] For example, if a user receives a solicitation call and the conversation contains suspicious language, the system will detect this and provide a message such as, "There may be some risks involved in what you're offering. Please consider it carefully." Furthermore, prompts such as, "Generate a reassuring message based on emotion data indicating whether the user is excited or nervous," will be used for the generative AI model.

[0199] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0200] Step 1:

[0201] The device collects the user's voice in real time via a microphone. The input is the user's voice data, which is converted into digital data as an audio signal. The output is digital audio data ready to be sent to the server.

[0202] Step 2:

[0203] The server receives audio data and uses an audio data analysis tool to execute speech recognition software such as the Google Speech-to-Text API. Here, the input is digital audio data, and this data is analyzed to output text data.

[0204] Step 3:

[0205] The server receives text data obtained by the voice data analysis system and uses data analysis tools to detect terms and phrases that suggest potential fraudulent or dishonest activities using natural language processing techniques. The input is text data, and the output is a list of detected keywords.

[0206] Step 4:

[0207] The server uses sentiment analysis tools to perform sentiment analysis on the obtained text data using software such as Microsoft Azure's Text Analytics. The input is text data, and the output is data indicating the user's emotional state.

[0208] Step 5:

[0209] The server integrates the information obtained in steps 3 and 4 and utilizes notification and question provision mechanisms to generate warnings and questions for the user. The input consists of detected keywords and sentiment state data, which are then output as warning messages adjusted to an appropriate tone.

[0210] Step 6:

[0211] The device receives the generated warning message and notifies the user visually and audibly using its display and speaker. The input is the warning message, and the output is the visual and audible notification to the user.

[0212] Step 7:

[0213] Users can provide feedback through a feedback reception system. This feedback is sent to the server and used to improve the system's algorithms. The input is user feedback data, and the output is data for system improvement.

[0214] 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.

[0215] 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.

[0216] 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.

[0217] [Second Embodiment]

[0218] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0219] 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.

[0220] 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).

[0221] 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.

[0222] 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.

[0223] 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).

[0224] 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.

[0225] 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.

[0226] 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.

[0227] 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.

[0228] 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.

[0229] 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".

[0230] The system of the present invention has the function of monitoring a user's voice conversation in real time and detecting and issuing warnings for fraudulent or inappropriate behavior. This system is realized through the following main components and their operation.

[0231] First, when a user begins speaking through the device, the device uses voice acquisition tools to collect voice data in real time. This voice data is immediately sent to the server.

[0232] The server accurately converts the audio data received through the speech recognition means into text, and the text information is processed by the analysis means. The analysis means uses natural language processing techniques to analyze the text data in detail and detect words and phrases that may contain fraudulent or malicious elements.

[0233] If detection occurs, the server will utilize warning and question provision mechanisms to display a warning message to the user and provide the user with specific questions that can be used to verify the veracity of the other party's statements.

[0234] The device uses this information to immediately notify the user of any potential risks. The user can then choose to continue the conversation or interrupt it by following the instructions received.

[0235] Furthermore, the system includes a means for receiving user feedback, and the feedback provided by users is used to improve the system's accuracy and functionality.

[0236] As a concrete example, consider a scenario where a user is told over the phone, "We need to verify your bank information, so please tell us your account number." In this case, the system detects keywords such as "bank" and "account number," issues a warning to the user, and further asks questions such as, "Which bank is calling from?" to help reduce the likelihood of becoming a victim of fraud.

[0237] In this way, the system of the present invention provides an effective means for protecting users from fraud and illegal activities.

[0238] The following describes the processing flow.

[0239] Step 1:

[0240] The device acquires the user's voice in real time through the microphone. This voice data is converted into packets at regular time intervals (for example, every few seconds) and sent to the server.

[0241] Step 2:

[0242] When the server receives audio data, it converts the audio into text data via speech recognition. This conversion process also removes background noise and corrects for differences in accent and intonation.

[0243] Step 3:

[0244] The server processes the converted text data using analysis tools. During this analysis process, natural language processing techniques are used to detect contextually relevant keywords and phrases, identifying elements that may be fraudulent or malicious.

[0245] Step 4:

[0246] If detection occurs, the server generates a warning message for the user using warning and question provision mechanisms. At the same time, it prepares to formulate and provide the user with specific and easy-to-understand questions.

[0247] Step 5:

[0248] The terminal receives warning messages and suggested questions sent from the server and immediately notifies the user through a visual or auditory interface.

[0249] Step 6:

[0250] Users can use the received warnings and suggested questions to decide whether to continue or discontinue the conversation with the other party. Users can also use the suggested questions to verify the other party's identity and intentions.

[0251] Step 7:

[0252] When a user provides feedback, the device collects that information and sends it to the server. The server then analyzes the received feedback and uses it to improve the system and increase its accuracy.

[0253] (Example 1)

[0254] 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."

[0255] In modern society, fraudulent activities and scams targeting users are on the rise, creating a need for real-time monitoring methods to effectively address them. However, current methods make it difficult to immediately warn users and provide concrete countermeasures. Furthermore, continuous system improvement utilizing feedback is not being adequately implemented. These challenges need to be addressed.

[0256] 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.

[0257] In this invention, the server includes voice acquisition means, voice recognition means, analysis means, warning and inquiry suggestion provision means, evaluation receiving means, and means for improving analysis accuracy and warning quality using a generated AI model. This enables real-time protection of users from fraud and misconduct, immediate provision of countermeasures, and continuous improvement of the system based on feedback.

[0258] "Voice acquisition means" refers to a device or method that acquires and transmits voice information from a user in real time.

[0259] "Speech recognition means" refers to a technology or process for converting speech information into symbolic information.

[0260] "Analysis means" refers to a technique or process that analyzes symbolic information to detect identifiers that indicate fraud or fraudulent activity.

[0261] "Warning and Inquiry Provision Methods" refers to technologies or processes that warn users of potential fraudulent activity and provide specific methods for making inquiries.

[0262] "Evaluation receiving means" refers to an interface or method for receiving evaluations from users.

[0263] A "generative AI model" is an artificial intelligence technology used to improve analysis accuracy and the quality of warnings based on feedback.

[0264] The system of this invention is designed to protect users from fraud and illegal activities by analyzing voice information in real time and providing immediate countermeasures. The specific implementation of the system is described below.

[0265] The terminal is equipped with a microphone and voice input device, and when the user begins to speak, voice data is collected in real time through these voice acquisition means. This voice data is transmitted to the server using a secure communication protocol. The communication means incorporate commonly used encryption technologies.

[0266] The server uses a speech recognition engine (e.g., a commercial speech recognition API) to convert the received audio data into text. At this stage, a high-precision language model is utilized to remove noise and support multiple languages. The converted text data is then analyzed in detail using natural language processing techniques. This process utilizes Python's NLTK and spaCy libraries to detect keywords and phrases that could potentially lead to fraud or misconduct.

[0267] If any malicious elements are detected, the server uses an AI model to generate a warning message. This message includes an immediate alert to the user and specific questions to help them respond in the conversation. The questions provided may include those that allow the user to verify the legitimacy of the other party, such as, "Which organization is contacting you?"

[0268] Based on the warning messages and suggested questions displayed on the device, users can choose the appropriate course of action to take with the other party. This information is communicated visually and audibly, allowing users to respond immediately.

[0269] In addition, the terminal is equipped with an interface for receiving user feedback. This feedback is sent to a server and used by an AI model to improve the system's accuracy. For example, prompts such as "Suggest improvements to the warning system based on recent fraud patterns" are input to the AI ​​to improve the system.

[0270] In this way, the system of the present invention provides a means to guide the user into a safe communication environment.

[0271] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0272] Step 1:

[0273] The terminal collects the user's speech in real time using a speech acquisition device. The input data is the user's voice, which the terminal converts into a digital audio signal and immediately sends to the server. The device used at this stage is the microphone built into the terminal. The output is digital audio data.

[0274] Step 2:

[0275] The server converts transmitted audio data into text data using a speech recognition engine. The input is digital audio data, and the server applies speech recognition technology to generate corresponding text from the audio signal through a language model. The output is the converted text data. This process involves background noise removal and accurate pronunciation recognition.

[0276] Step 3:

[0277] The server applies natural language processing techniques to analyze text data and detect specific keywords and phrases. The input is text data, and the data analysis process considers context and intent. It utilizes Python's NLTK and spaCy libraries to identify elements that indicate fraud or misconduct. The output consists of the detected fraudulent keywords and phrases.

[0278] Step 4:

[0279] The server generates warning messages and specific questions based on the detection results. The input is identified keywords or phrases, and a generative AI model is used to design appropriate warnings and questions. The warning message serves as a reminder to the user, and the questions help verify the legitimacy of the other party. The output is the warning message and questions to be displayed to the user.

[0280] Step 5:

[0281] The terminal receives warning messages and suggested questions from the server and notifies the user visually and audibly. Input is message data from the server, which the terminal presents to the user through its interface. Based on this, the user can select an appropriate response. Output is screen displays and audio notifications that the user can see.

[0282] Step 6:

[0283] Based on the provided information, the user decides whether to continue or interrupt the conversation and inputs feedback to the system through the terminal. The input is information regarding the user's selection and actions, which is sent to the server. The system receives this feedback and uses a generative AI model to improve the system's analysis accuracy and the quality of warnings. The output is improved system performance.

[0284] (Application Example 1)

[0285] Next, Application Example 1 will be described. In the following description, the data processing device 12 is referred to as the "server", and the smart glasses 214 are referred to as the "terminal".

[0286] In modern society, fraud and illegal activities are carried out daily, and the risk for users to suffer from such harm is increasing. Especially in voice communication, it is difficult to detect fraudulent behavior in advance and respond promptly. With current technologies, it is not sufficient to detect illegal activities in real time and issue immediate warnings to users, so an efficient solution for ensuring user safety is being sought.

[0287] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0288] In this invention, the server includes a voice data acquisition means for receiving the user's voice in real time and transmitting it to the server, a voice recognition means for converting the voice information transmitted from the voice data acquisition means into text, and an analysis means for analyzing the converted text to detect words and expressions indicating signs of illegality or fraud. This enables early detection of illegal activities and prompt warning notifications to users.

[0289] <00​​​​"Voice data acquisition means" refers to devices or software that receive the user's voice in real time and transmit it to a server.

[0291] A "server" is a computer system used for data processing and communication.

[0292] "Speech recognition means" refers to a technology or device for converting speech data into text.

[0293] "Analysis means" refers to techniques or devices for analyzing converted text and detecting words or expressions that may indicate fraud or misconduct.

[0294] A "warning and question supply means" is a means of notifying the user of an alert based on detected words or expressions and providing specific examples of questions.

[0295] A "feedback receiving means" is a device or system equipped with an interface that allows users to provide comments.

[0296] A "cloud environment" is an environment that utilizes computing resources provided over the internet.

[0297] This invention is a system that protects users by detecting fraudulent or illegal behavior in real time and issuing warnings when users engage in voice communication using devices such as smartphones. Specific embodiments of this system are shown below.

[0298] When a user initiates a call via their smartphone, the audio is received in real time by an audio data acquisition system. The received audio is immediately transmitted to the server. At this stage, the smartphone uses a communication module to maintain a stable connection.

[0299] The server instantly converts audio data into text using speech recognition tools such as Google Cloud Speech-to-Text. This allows for real-time transcription of voice communication. The converted text is then analyzed using natural language processing techniques (e.g., SpaCy and BERT). In this analysis, the text is scrutinized for potentially fraudulent or deceptive words and phrases based on their context.

[0300] If the analysis detects fraudulent content, the server will issue visual and auditory warnings to the user's smartphone using warning and question-providing mechanisms. At the same time, it will provide the user with specific questions, such as "Which organization is calling from?", to help them identify suspicious points in the conversation.

[0301] Users can provide feedback to the system through feedback receiving mechanisms. This feedback is stored and analyzed on the server for continuous improvement of the system. The data for improvement is accumulated in a cloud environment and used to train the generated AI model.

[0302] For example, if a user calls customer support for an online shopping site and is asked for their credit card information, the system will detect this phrase and display a warning to the user saying, "We recommend using the official website." This helps protect the user from unintentional fraud.

[0303] An example of a prompt message is: "Identify the phrase indicating fraudulent activity from the following phone conversation: 'An error has occurred in our system. We need to verify your credit card information immediately.'"

[0304] In this way, this invention provides a means to prevent fraudulent activities in users' voice communications.

[0305] The flow of the specific process in Application Example 1 will be described using FIG. 12.

[0306] Step 1:

[0307] Upon the user's call start, the terminal uses the voice data acquisition means to receive voice in real time. The input is a voice signal, and the output is digitized voice data. The voice data is immediately transmitted to the server.

[0308] Step 2:

[0309] The server uses the voice recognition means to convert the received digital voice data into text. The input is digital voice data, and the output is text data. Here, the operation of converting voice into text is performed using the Google Cloud Speech-to-Text service.

[0310] Step 3:

[0311] The server analyzes the converted text data using the analysis means. The input is text data, and the output is a flag indicating whether there are fraudulent or improper words or actions. For the analysis, SpaCy or BERT is used to utilize natural language processing technology to identify words and phrases.

[0312] Step 4:

[0313] Based on the result of the analysis, the server uses the warning and question proposal means to send visual and auditory warnings to the terminal. The input is the improper words or actions detection flag, and the output is a warning message and specific question proposals to the user. As a result, notifications are displayed and played on the user's device.

[0314] Step 5:

[0315] The user chooses whether to continue or interrupt the conversation based on the provided warnings and suggested questions. The input consists of warning messages and suggested questions, while the output is the action chosen by the user. The user can ask additional questions to ensure safety.

[0316] Step 6:

[0317] The terminal sends feedback collected from users to a server via a feedback receiving mechanism. The input is feedback information provided by the user, and the output is the cloud storage of the feedback data. This data will be used for future system improvements.

[0318] Step 7:

[0319] The server analyzes feedback data to train and improve the AI ​​model generated in the cloud environment. The input is feedback data, and the output is an updated model for system improvement. The analysis results will be used to adjust the model and improve prompt messages in the future.

[0320] 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.

[0321] The system of the present invention not only monitors user conversations in real time and detects signs of fraud or misconduct, but also recognizes the user's emotional state and adjusts its response accordingly. This system is implemented with the following components:

[0322] First, when a user initiates a conversation via the device, the device uses voice acquisition means to collect the user's voice data in real time and sends the data to the server.

[0323] The server uses speech recognition to convert the audio data into text. During this process, an emotion engine analyzes the audio data and uses factors such as tone of voice, speaking speed, and emphasis patterns to identify the user's emotional state.

[0324] The converted text data is processed by an analysis tool to detect keywords or phrases indicating fraud or misleading expressions. This detection takes into account emotional state information from an emotion engine, enabling more contextually accurate analysis.

[0325] Based on the detection results, the server generates warnings and suggested questions for the user using warning and question provision mechanisms. At this time, based on information from the emotion engine, notifications are made in a tone and content that matches the user's current emotional state.

[0326] The device receives the generated warning messages and suggested questions and notifies the user visually or audibly. The user can then use this information to make informed decisions.

[0327] Furthermore, by using emotion engine data when users provide feedback, the emotional nuances of the feedback are also taken into consideration, which can be used to improve the system.

[0328] For example, if a user receives a suspicious phone call and is asked, "Would you like to receive a limited offer?", the system will detect the keyword "limited offer," which suggests potential fraud. If the emotion engine senses tension in the user's voice, it will respond by providing a gentle warning such as, "Please relax and think about it. Please check specifically what the offer is." In this way, the system can not only effectively protect users from fraud but also provide emotional support.

[0329] The following describes the processing flow.

[0330] Step 1:

[0331] The device acquires the user's voice in real time via the microphone and prepares to send this audio data to the server. During this process, noise reduction is also performed to maintain audio clarity.

[0332] Step 2:

[0333] The server converts the received audio data into text format using speech recognition technology. This conversion uses a language model and focuses on accurately transcribing the user's utterances into text.

[0334] Step 3:

[0335] The server analyzes the converted text data using analytical tools. This analysis detects keywords and phrases that indicate fraud or scams. Natural language processing techniques are used to consider the context and aim for higher accuracy in detection.

[0336] Step 4:

[0337] Simultaneously, the server activates an emotion engine, analyzing the tone, speed, and pacing of the audio data to determine the user's emotional state. The analysis results indicate the user's current emotions and influence the next steps.

[0338] Step 5:

[0339] The server generates warnings and suggested questions based on information from the emotion engine. This generation process takes into account the user's emotional state and includes reassuring language and suggestions to support the user.

[0340] Step 6:

[0341] The terminal receives warning messages and suggested questions sent from the server and presents them to the user through visual and auditory interfaces. For example, it may display text on the screen while simultaneously providing important information via audio.

[0342] Step 7:

[0343] Based on the information presented, users can decide whether to continue or end the conversation with the other party. They can also report their thoughts and suggestions for improvement through the system's feedback function as needed.

[0344] Step 8:

[0345] The server collects user feedback and uses it as data to improve the system's analytical capabilities and user interface. This feedback, in particular, contributes to improving the accuracy of the emotion engine.

[0346] (Example 2)

[0347] 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".

[0348] While it is crucial to detect signs of fraud and misconduct early in user conversations, traditional systems often issue warnings without considering the user's emotional state, potentially leading to misunderstandings and stress. Therefore, there is a need to more effectively protect users by providing appropriate warnings and feedback that are tailored to their emotional state.

[0349] 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.

[0350] In this invention, the server includes speech recognition means, emotion analysis means, analysis means, warning and question provision means, and feedback receiving means. This enables the provision of accurate warnings and questions that take into account the user's emotional state, thereby more effectively protecting the user and facilitating system improvement using feedback.

[0351] A "voice acquisition means" is a device that receives the user's voice in real time and transmits it to an information processing device.

[0352] "Speech recognition means" refers to a technology or device that analyzes received speech data and converts it into text data.

[0353] "Emotional analysis means" refers to a technology or device that analyzes voice data and uses voice tone, speaking speed, and emphasis patterns to identify the user's emotional state.

[0354] "Analysis means" refers to a technology or device that analyzes text data, detects keywords or phrases indicating fraud or scams, and performs contextual analysis.

[0355] "Warning and Question Provisioning Means" refers to a technology or device that notifies the user of a customized warning and provides specific questions based on detected keywords or phrases and emotional states.

[0356] "Feedback receiving means" refers to a technology or device that has an interface for users to provide feedback and to receive that feedback while taking emotional nuances into consideration.

[0357] The system of this invention not only analyzes user conversations from audio to detect signs of fraud or dishonest behavior, but also has the function of understanding the user's emotional state and providing appropriate responses. The specific configuration of the system includes a terminal, a server, and various analysis engines.

[0358] When a user begins a conversation, the device collects voice data in real time through its built-in microphone and voice acquisition software. This voice data is compressed and packetized as initial processing before being sent to the server. The server has powerful processing capabilities and utilizes common speech recognition APIs, particularly for speech recognition. Specific examples include Google Cloud Speech-to-Text and Amazon Transcribe.

[0359] The server converts the received audio data into text data, then uses an emotion analysis engine to analyze the tone, volume, and speaking speed of the voice to identify the user's emotional state. This analysis engine is specialized in extracting emotions from speech using machine learning models and specific algorithms.

[0360] The converted text data is analyzed using natural language processing (NLP) techniques to detect keywords and phrases that indicate tendencies toward fraudulent activity or deception. This analysis also considers emotional state information, enabling more accurate contextual analysis. Based on the analysis results, the server generates warning messages and suggested questions tailored to the user's emotional state. This process utilizes a generative AI model to generate user-specific responses.

[0361] The generated warning messages and proposed questions are sent to the terminal and notified to the user visually or audibly. The user can then make their own decisions based on this notification. Furthermore, user feedback is sent to the server via the terminal, where sentiment analysis is also performed. This feedback is used for the continuous improvement of the system.

[0362] For example, if a user receives a sales call and is asked, "Would you like to receive a limited offer?", the system will detect the keyword "limited offer." At the same time, if it detects tension in the user's voice, it will provide a warning such as, "Please relax and think carefully about what you're going to say. We recommend that you clearly confirm the details." In this way, the system ensures the user's safety while also providing emotional support.

[0363] An example of a prompt message would be, "How should the system generate a warning if a user receives a potentially fraudulent phone call?"

[0364] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0365] Step 1:

[0366] The user uses a device to initiate a conversation. The device collects the conversation in real time using a voice acquisition device. The input is the user's voice, and voice data is generated as output. This voice data undergoes noise reduction and echo cancellation through digital signal processing before being sent to the server.

[0367] Step 2:

[0368] The server passes the audio data received from the terminal to the speech recognition system. The input is audio data, which is converted into text data as output. During this conversion process, speech recognition algorithms are applied to recognize phonemes and words. Techniques used for the conversion include phoneme decomposition and hidden Markov models.

[0369] Step 3:

[0370] The text data is analyzed by sentiment analysis tools on the server. The input is the text data generated in step 2, and the output is information indicating the user's emotional state. In this process, the analysis engine evaluates the voice tone, speaking speed, and emphasis patterns to determine the user's emotions. For example, a machine learning model using an annotated dataset can predict emotions.

[0371] Step 4:

[0372] The server further analyzes the text data using analysis tools. The input is the sentiment information and text data from step 3, and the output is the detection results of keywords and phrases indicating signs of fraud or wrongdoing. This process uses natural language processing techniques to scan the data and perform dictionary matching and contextual analysis. Sentiment information improves contextual understanding.

[0373] Step 5:

[0374] The server generates appropriate responses to the user using warning and question-providing mechanisms. The input is the detection results and emotional state obtained in step 4, and the output is a customized warning message and question. This generative AI model enables communication in an emotionally sensitive tone. For example, it can provide voice messages in a gentle tone.

[0375] Step 6:

[0376] The terminal visually or audibly notifies the user of warning messages and suggested questions received from the server. Input is the generated response message, and output is the notification to the user. The terminal communicates information to the user through screen display and speech synthesis. Notifications are provided in an easy-to-understand format using a user interface.

[0377] Step 7:

[0378] Users provide feedback to the system through a feedback receiving mechanism. The input is the feedback information delivered by the user, and the output is feedback data that contributes to system improvement. The server analyzes this feedback along with sentiment analysis and uses the findings to improve the system's accuracy. The submitted feedback is stored in a database.

[0379] (Application Example 2)

[0380] 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."

[0381] Fraud and dishonest behavior are significant issues in daily life, and it is essential for users to be able to detect these behaviors during conversations and take appropriate action. However, conventional technologies do not provide warnings or instructions that take into account the user's emotional state, which can cause stress. Therefore, the present invention aims to provide a system that takes the user's emotional state into consideration and provides more accurate and appropriate warnings and support.

[0382] 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.

[0383] In this invention, the server includes voice data collection means, voice data analysis means, data analysis means, notification and question provision means, opinion receiving means, and sentiment analysis means. This makes it possible to detect fraudulent activity and provide warnings in an appropriate tone while taking into account the user's emotional state in real time.

[0384] "Voice data collection means" refers to means for acquiring the user's voice in real time and transmitting it to a processing device.

[0385] "Voice data analysis means" refers to means for converting acquired voice information into text data.

[0386] "Data analysis means" refers to means for analyzing converted text data to detect terms or phrases that may indicate fraudulent or dishonest activity.

[0387] "Notification and Question Provisioning Means" are means of alerting users and providing more detailed questions based on detected terms or phrases.

[0388] A "means for receiving opinions" refers to a means equipped with a user interaction interface that allows users to submit opinions.

[0389] "Emotion analysis means" refers to a means for detecting the user's emotional state and adjusting the tone of notification messages based on that state.

[0390] The system for realizing this invention consists of a voice data collection means, a voice data analysis means, a data analysis means, a notification and question provision means, an opinion receiving means, and an emotion analysis means. The operation of the system is as follows.

[0391] First, the voice spoken by the user is collected as audio data using the microphone built into the device, and the audio data collection means transmits it to the processing unit. The audio data analysis means converts this audio data into text data using speech recognition software such as the Google Speech-to-Text API. At this time, the server processes the converted text data using the data analysis means and detects terms and phrases that suggest the possibility of fraudulent or dishonest activity using natural language processing technology.

[0392] Next, the server uses sentiment analysis tools to analyze the user's emotional state using software such as Microsoft Azure's Text Analytics. This allows warning messages related to detected terms to be adjusted based on the emotional state. The adjusted messages are then visually and audibly communicated to the user through the terminal's display and speakers via notification and question-providing tools.

[0393] Users can provide feedback through various means of receiving it. This feedback will be used to improve the system's algorithms.

[0394] For example, if a user receives a solicitation call and the conversation contains suspicious language, the system will detect this and provide a message such as, "There may be some risks involved in what you're offering. Please consider it carefully." Furthermore, prompts such as, "Generate a reassuring message based on emotion data indicating whether the user is excited or nervous," will be used for the generative AI model.

[0395] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0396] Step 1:

[0397] The device collects the user's voice in real time via a microphone. The input is the user's voice data, which is converted into digital data as an audio signal. The output is digital audio data ready to be sent to the server.

[0398] Step 2:

[0399] The server receives audio data and uses an audio data analysis tool to execute speech recognition software such as the Google Speech-to-Text API. Here, the input is digital audio data, and this data is analyzed to output text data.

[0400] Step 3:

[0401] The server receives text data obtained by the voice data analysis system and uses data analysis tools to detect terms and phrases that suggest potential fraudulent or dishonest activities using natural language processing techniques. The input is text data, and the output is a list of detected keywords.

[0402] Step 4:

[0403] The server uses sentiment analysis tools to perform sentiment analysis on the obtained text data using software such as Microsoft Azure's Text Analytics. The input is text data, and the output is data indicating the user's emotional state.

[0404] Step 5:

[0405] The server integrates the information obtained in steps 3 and 4 and utilizes notification and question provision mechanisms to generate warnings and questions for the user. The input consists of detected keywords and sentiment state data, which are then output as warning messages adjusted to an appropriate tone.

[0406] Step 6:

[0407] The device receives the generated warning message and notifies the user visually and audibly using its display and speaker. The input is the warning message, and the output is the visual and audible notification to the user.

[0408] Step 7:

[0409] Users can provide feedback through a feedback reception system. This feedback is sent to the server and used to improve the system's algorithms. The input is user feedback data, and the output is data for system improvement.

[0410] 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.

[0411] 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.

[0412] 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.

[0413] [Third Embodiment]

[0414] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0415] 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.

[0416] 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).

[0417] 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.

[0418] 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.

[0419] 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).

[0420] 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.

[0421] 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.

[0422] 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.

[0423] 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.

[0424] 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.

[0425] 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".

[0426] The system of the present invention has the function of monitoring a user's voice conversation in real time and detecting and issuing warnings for fraudulent or inappropriate behavior. This system is realized through the following main components and their operation.

[0427] First, when a user begins speaking through the device, the device uses voice acquisition tools to collect voice data in real time. This voice data is immediately sent to the server.

[0428] The server accurately converts the audio data received through the speech recognition means into text, and the text information is processed by the analysis means. The analysis means uses natural language processing techniques to analyze the text data in detail and detect words and phrases that may contain fraudulent or malicious elements.

[0429] If detection occurs, the server will utilize warning and question provision mechanisms to display a warning message to the user and provide the user with specific questions that can be used to verify the veracity of the other party's statements.

[0430] The device uses this information to immediately notify the user of any potential risks. The user can then choose to continue the conversation or interrupt it by following the instructions received.

[0431] Furthermore, the system includes a means for receiving user feedback, and the feedback provided by users is used to improve the system's accuracy and functionality.

[0432] As a concrete example, consider a scenario where a user is told over the phone, "We need to verify your bank information, so please tell us your account number." In this case, the system detects keywords such as "bank" and "account number," issues a warning to the user, and further asks questions such as, "Which bank is calling from?" to help reduce the likelihood of becoming a victim of fraud.

[0433] In this way, the system of the present invention provides an effective means for protecting users from fraud and illegal activities.

[0434] The following describes the processing flow.

[0435] Step 1:

[0436] The device acquires the user's voice in real time through the microphone. This voice data is converted into packets at regular time intervals (for example, every few seconds) and sent to the server.

[0437] Step 2:

[0438] When the server receives audio data, it converts the audio into text data via speech recognition. This conversion process also removes background noise and corrects for differences in accent and intonation.

[0439] Step 3:

[0440] The server processes the converted text data using analysis tools. During this analysis process, natural language processing techniques are used to detect contextually relevant keywords and phrases, identifying elements that may be fraudulent or malicious.

[0441] Step 4:

[0442] If detection occurs, the server generates a warning message for the user using warning and question provision mechanisms. At the same time, it prepares to formulate and provide the user with specific and easy-to-understand questions.

[0443] Step 5:

[0444] The terminal receives warning messages and suggested questions sent from the server and immediately notifies the user through a visual or auditory interface.

[0445] Step 6:

[0446] Users can use the received warnings and suggested questions to decide whether to continue or discontinue the conversation with the other party. Users can also use the suggested questions to verify the other party's identity and intentions.

[0447] Step 7:

[0448] When a user provides feedback, the device collects that information and sends it to the server. The server then analyzes the received feedback and uses it to improve the system and increase its accuracy.

[0449] (Example 1)

[0450] 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."

[0451] In modern society, fraudulent activities and scams targeting users are on the rise, creating a need for real-time monitoring methods to effectively address them. However, current methods make it difficult to immediately warn users and provide concrete countermeasures. Furthermore, continuous system improvement utilizing feedback is not being adequately implemented. These challenges need to be addressed.

[0452] 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.

[0453] In this invention, the server includes voice acquisition means, voice recognition means, analysis means, warning and inquiry suggestion provision means, evaluation receiving means, and means for improving analysis accuracy and warning quality using a generated AI model. This enables real-time protection of users from fraud and misconduct, immediate provision of countermeasures, and continuous improvement of the system based on feedback.

[0454] "Voice acquisition means" refers to a device or method that acquires and transmits voice information from a user in real time.

[0455] "Speech recognition means" refers to a technology or process for converting speech information into symbolic information.

[0456] "Analysis means" refers to a technique or process that analyzes symbolic information to detect identifiers that indicate fraud or fraudulent activity.

[0457] "Warning and Inquiry Provision Methods" refers to technologies or processes that warn users of potential fraudulent activity and provide specific methods for making inquiries.

[0458] "Evaluation receiving means" refers to an interface or method for receiving evaluations from users.

[0459] A "generative AI model" is an artificial intelligence technology used to improve analysis accuracy and the quality of warnings based on feedback.

[0460] The system of this invention is designed to protect users from fraud and illegal activities by analyzing voice information in real time and providing immediate countermeasures. The specific implementation of the system is described below.

[0461] The terminal is equipped with a microphone and voice input device, and when the user begins to speak, voice data is collected in real time through these voice acquisition means. This voice data is transmitted to the server using a secure communication protocol. The communication means incorporate commonly used encryption technologies.

[0462] The server uses a speech recognition engine (e.g., a commercial speech recognition API) to convert the received audio data into text. At this stage, a high-precision language model is utilized to remove noise and support multiple languages. The converted text data is then analyzed in detail using natural language processing techniques. This process utilizes Python's NLTK and spaCy libraries to detect keywords and phrases that could potentially lead to fraud or misconduct.

[0463] If any malicious elements are detected, the server uses an AI model to generate a warning message. This message includes an immediate alert to the user and specific questions to help them respond in the conversation. The questions provided may include those that allow the user to verify the legitimacy of the other party, such as, "Which organization is contacting you?"

[0464] Based on the warning messages and suggested questions displayed on the device, users can choose the appropriate course of action to take with the other party. This information is communicated visually and audibly, allowing users to respond immediately.

[0465] In addition, the terminal is equipped with an interface for receiving user feedback. This feedback is sent to a server and used by an AI model to improve the system's accuracy. For example, prompts such as "Suggest improvements to the warning system based on recent fraud patterns" are input to the AI ​​to improve the system.

[0466] In this way, the system of the present invention provides a means to guide the user into a safe communication environment.

[0467] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0468] Step 1:

[0469] The terminal collects the user's speech in real time using a speech acquisition device. The input data is the user's voice, which the terminal converts into a digital audio signal and immediately sends to the server. The device used at this stage is the microphone built into the terminal. The output is digital audio data.

[0470] Step 2:

[0471] The server converts transmitted audio data into text data using a speech recognition engine. The input is digital audio data, and the server applies speech recognition technology to generate corresponding text from the audio signal through a language model. The output is the converted text data. This process involves background noise removal and accurate pronunciation recognition.

[0472] Step 3:

[0473] The server applies natural language processing techniques to analyze text data and detect specific keywords and phrases. The input is text data, and the data analysis process considers context and intent. It utilizes Python's NLTK and spaCy libraries to identify elements that indicate fraud or misconduct. The output consists of the detected fraudulent keywords and phrases.

[0474] Step 4:

[0475] The server generates warning messages and specific questions based on the detection results. The input is identified keywords or phrases, and a generative AI model is used to design appropriate warnings and questions. The warning message serves as a reminder to the user, and the questions help verify the legitimacy of the other party. The output is the warning message and questions to be displayed to the user.

[0476] Step 5:

[0477] The terminal receives warning messages and suggested questions from the server and notifies the user visually and audibly. Input is message data from the server, which the terminal presents to the user through its interface. Based on this, the user can select an appropriate response. Output is screen displays and audio notifications that the user can see.

[0478] Step 6:

[0479] Based on the information provided, the user decides whether to continue or interrupt the conversation and provides feedback to the system via the terminal. This feedback, containing information about the user's choices and actions, is sent to the server. The system receives this feedback and uses a generative AI model to improve the system's analysis accuracy and the quality of its warnings. The output represents the improved system performance.

[0480] (Application Example 1)

[0481] 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."

[0482] In modern society, fraud and dishonest practices are commonplace, increasing the risk of users becoming victims. In particular, detecting fraudulent behavior in advance and responding quickly is difficult in voice communication. Current technology is insufficient to detect fraudulent activity in real time and issue immediate warnings to users; therefore, an efficient solution is needed to ensure user safety.

[0483] 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.

[0484] In this invention, the server includes voice data acquisition means for receiving user voice in real time and transmitting it to the server, voice recognition means for converting the voice information transmitted from the voice data acquisition means into text, and analysis means for analyzing the converted text and detecting words or expressions that indicate fraud or misconduct. This enables early detection of fraudulent activity and rapid notification of warnings to the user.

[0485] A "user" is an individual or legal entity that uses the system to conduct voice communication.

[0486] "Voice data acquisition means" refers to devices or software that receive the user's voice in real time and transmit it to a server.

[0487] A "server" is a computer system used for data processing and communication.

[0488] "Speech recognition means" refers to a technology or device for converting speech data into text.

[0489] "Analysis means" refers to techniques or devices for analyzing converted text and detecting words or expressions that may indicate fraud or misconduct.

[0490] A "warning and question supply means" is a means of notifying the user of an alert based on detected words or expressions and providing specific examples of questions.

[0491] A "feedback receiving means" is a device or system equipped with an interface that allows users to provide comments.

[0492] A "cloud environment" is an environment that utilizes computing resources provided over the internet.

[0493] This invention is a system that protects users by detecting fraudulent or illegal behavior in real time and issuing warnings when users engage in voice communication using devices such as smartphones. Specific embodiments of this system are shown below.

[0494] When a user initiates a call via their smartphone, the audio is received in real time by an audio data acquisition system. The received audio is immediately transmitted to the server. At this stage, the smartphone uses a communication module to maintain a stable connection.

[0495] The server instantly converts audio data into text using speech recognition tools such as Google Cloud Speech-to-Text. This allows for real-time transcription of voice communication. The converted text is then analyzed using natural language processing techniques (e.g., SpaCy and BERT). In this analysis, the text is scrutinized for potentially fraudulent or deceptive words and phrases based on their context.

[0496] If the analysis detects fraudulent content, the server will issue visual and auditory warnings to the user's smartphone using warning and question-providing mechanisms. At the same time, it will provide the user with specific questions, such as "Which organization is calling from?", to help them identify suspicious points in the conversation.

[0497] Users can provide feedback to the system through feedback receiving mechanisms. This feedback is stored and analyzed on the server for continuous improvement of the system. The data for improvement is accumulated in a cloud environment and used to train the generated AI model.

[0498] For example, if a user calls customer support for an online shopping site and is asked for their credit card information, the system will detect this phrase and display a warning to the user saying, "We recommend using the official website." This helps protect the user from unintentional fraud.

[0499] An example of a prompt message is: "Identify the phrase indicating fraudulent activity from the following phone conversation: 'An error has occurred in our system. We need to verify your credit card information immediately.'"

[0500] In this way, this invention provides a means to prevent fraudulent activities in users' voice communications.

[0501] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0502] Step 1:

[0503] The terminal receives audio in real time using an audio data acquisition mechanism as soon as the user initiates a call. The input is an audio signal, and the output is digitized audio data. The audio data is immediately transmitted to the server.

[0504] Step 2:

[0505] The server converts received digital audio data into text using speech recognition technology. The input is digital audio data, and the output is text data. Here, the Google Cloud Speech-to-Text service is used to perform the audio-to-text conversion.

[0506] Step 3:

[0507] The server analyzes the converted text data using parsing tools. The input is text data, and the output is a flag indicating whether the text indicates fraudulent or illegal behavior. SpaCy and BERT are used for the analysis, and natural language processing techniques are employed to identify words and phrases.

[0508] Step 4:

[0509] Based on the analysis results, the server uses warning and question-providing mechanisms to send visual and auditory warnings to the terminal. The input is a flag indicating inappropriate behavior, and the output is a warning message and specific questions for the user. This results in a notification being displayed and played on the user's device.

[0510] Step 5:

[0511] The user chooses whether to continue or interrupt the conversation based on the provided warnings and suggested questions. The input consists of warning messages and suggested questions, while the output is the action chosen by the user. The user can ask additional questions to ensure safety.

[0512] Step 6:

[0513] The terminal sends feedback collected from users to a server via a feedback receiving mechanism. The input is feedback information provided by the user, and the output is the cloud storage of the feedback data. This data will be used for future system improvements.

[0514] Step 7:

[0515] The server analyzes feedback data to train and improve the AI ​​model generated in the cloud environment. The input is feedback data, and the output is an updated model for system improvement. The analysis results will be used to adjust the model and improve prompt messages in the future.

[0516] 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.

[0517] The system of the present invention not only monitors user conversations in real time and detects signs of fraud or misconduct, but also recognizes the user's emotional state and adjusts its response accordingly. This system is implemented with the following components:

[0518] First, when a user initiates a conversation via the device, the device uses voice acquisition means to collect the user's voice data in real time and sends the data to the server.

[0519] The server uses speech recognition to convert the audio data into text. During this process, an emotion engine analyzes the audio data and uses factors such as tone of voice, speaking speed, and emphasis patterns to identify the user's emotional state.

[0520] The converted text data is processed by an analysis tool to detect keywords or phrases indicating fraud or misleading expressions. This detection takes into account emotional state information from an emotion engine, enabling more contextually accurate analysis.

[0521] Based on the detection results, the server generates warnings and suggested questions for the user using warning and question provision mechanisms. At this time, based on information from the emotion engine, notifications are made in a tone and content that matches the user's current emotional state.

[0522] The device receives the generated warning messages and suggested questions and notifies the user visually or audibly. The user can then use this information to make informed decisions.

[0523] Furthermore, by using emotion engine data when users provide feedback, the emotional nuances of the feedback are also taken into consideration, which can be used to improve the system.

[0524] For example, if a user receives a suspicious phone call and is asked, "Would you like to receive a limited offer?", the system will detect the keyword "limited offer," which suggests potential fraud. If the emotion engine senses tension in the user's voice, it will respond by providing a gentle warning such as, "Please relax and think about it. Please check specifically what the offer is." In this way, the system can not only effectively protect users from fraud but also provide emotional support.

[0525] The following describes the processing flow.

[0526] Step 1:

[0527] The device acquires the user's voice in real time via the microphone and prepares to send this audio data to the server. During this process, noise reduction is also performed to maintain audio clarity.

[0528] Step 2:

[0529] The server converts the received audio data into text format using speech recognition technology. This conversion uses a language model and focuses on accurately transcribing the user's utterances into text.

[0530] Step 3:

[0531] The server analyzes the converted text data using analytical tools. This analysis detects keywords and phrases that indicate fraud or scams. Natural language processing techniques are used to consider the context and aim for higher accuracy in detection.

[0532] Step 4:

[0533] Simultaneously, the server activates an emotion engine, analyzing the tone, speed, and pacing of the audio data to determine the user's emotional state. The analysis results indicate the user's current emotions and influence the next steps.

[0534] Step 5:

[0535] The server generates warnings and suggested questions based on information from the emotion engine. This generation process takes into account the user's emotional state and includes reassuring language and suggestions to support the user.

[0536] Step 6:

[0537] The terminal receives warning messages and suggested questions sent from the server and presents them to the user through visual and auditory interfaces. For example, it may display text on the screen while simultaneously providing important information via audio.

[0538] Step 7:

[0539] Based on the information presented, users can decide whether to continue or end the conversation with the other party. They can also report their thoughts and suggestions for improvement through the system's feedback function as needed.

[0540] Step 8:

[0541] The server collects user feedback and uses it as data to improve the system's analytical capabilities and user interface. This feedback, in particular, contributes to improving the accuracy of the emotion engine.

[0542] (Example 2)

[0543] 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."

[0544] While it is crucial to detect signs of fraud and misconduct early in user conversations, traditional systems often issue warnings without considering the user's emotional state, potentially leading to misunderstandings and stress. Therefore, there is a need to more effectively protect users by providing appropriate warnings and feedback that are tailored to their emotional state.

[0545] 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.

[0546] In this invention, the server includes speech recognition means, emotion analysis means, analysis means, warning and question provision means, and feedback receiving means. This enables the provision of accurate warnings and questions that take into account the user's emotional state, thereby more effectively protecting the user and facilitating system improvement using feedback.

[0547] A "voice acquisition means" is a device that receives the user's voice in real time and transmits it to an information processing device.

[0548] "Speech recognition means" refers to a technology or device that analyzes received speech data and converts it into text data.

[0549] "Emotional analysis means" refers to a technology or device that analyzes voice data and uses voice tone, speaking speed, and emphasis patterns to identify the user's emotional state.

[0550] "Analysis means" refers to a technology or device that analyzes text data, detects keywords or phrases indicating fraud or scams, and performs contextual analysis.

[0551] "Warning and Question Provisioning Means" refers to a technology or device that notifies the user of a customized warning and provides specific questions based on detected keywords or phrases and emotional states.

[0552] "Feedback receiving means" refers to a technology or device that has an interface for users to provide feedback and to receive that feedback while taking emotional nuances into consideration.

[0553] The system of this invention not only analyzes user conversations from audio to detect signs of fraud or dishonest behavior, but also has the function of understanding the user's emotional state and providing appropriate responses. The specific configuration of the system includes a terminal, a server, and various analysis engines.

[0554] When a user begins a conversation, the device collects voice data in real time through its built-in microphone and voice acquisition software. This voice data is compressed and packetized as initial processing before being sent to the server. The server has powerful processing capabilities and utilizes common speech recognition APIs, particularly for speech recognition. Specific examples include Google Cloud Speech-to-Text and Amazon Transcribe.

[0555] The server converts the received audio data into text data, then uses an emotion analysis engine to analyze the tone, volume, and speaking speed of the voice to identify the user's emotional state. This analysis engine is specialized in extracting emotions from speech using machine learning models and specific algorithms.

[0556] The converted text data is analyzed using natural language processing (NLP) techniques to detect keywords and phrases that indicate tendencies toward fraudulent activity or deception. This analysis also considers emotional state information, enabling more accurate contextual analysis. Based on the analysis results, the server generates warning messages and suggested questions tailored to the user's emotional state. This process utilizes a generative AI model to generate user-specific responses.

[0557] The generated warning messages and proposed questions are sent to the terminal and notified to the user visually or audibly. The user can then make their own decisions based on this notification. Furthermore, user feedback is sent to the server via the terminal, where sentiment analysis is also performed. This feedback is used for the continuous improvement of the system.

[0558] For example, if a user receives a sales call and is asked, "Would you like to receive a limited offer?", the system will detect the keyword "limited offer." At the same time, if it detects tension in the user's voice, it will provide a warning such as, "Please relax and think carefully about what you're going to say. We recommend that you clearly confirm the details." In this way, the system ensures the user's safety while also providing emotional support.

[0559] An example of a prompt message would be, "How should the system generate a warning if a user receives a potentially fraudulent phone call?"

[0560] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0561] Step 1:

[0562] The user uses a device to initiate a conversation. The device collects the conversation in real time using a voice acquisition device. The input is the user's voice, and voice data is generated as output. This voice data undergoes noise reduction and echo cancellation through digital signal processing before being sent to the server.

[0563] Step 2:

[0564] The server passes the audio data received from the terminal to the speech recognition system. The input is audio data, which is converted into text data as output. During this conversion process, speech recognition algorithms are applied to recognize phonemes and words. Techniques used for the conversion include phoneme decomposition and hidden Markov models.

[0565] Step 3:

[0566] The text data is analyzed by sentiment analysis tools on the server. The input is the text data generated in step 2, and the output is information indicating the user's emotional state. In this process, the analysis engine evaluates the voice tone, speaking speed, and emphasis patterns to determine the user's emotions. For example, a machine learning model using an annotated dataset can predict emotions.

[0567] Step 4:

[0568] The server further analyzes the text data using analysis tools. The input is the sentiment information and text data from step 3, and the output is the detection results of keywords and phrases indicating signs of fraud or wrongdoing. This process uses natural language processing techniques to scan the data and perform dictionary matching and contextual analysis. Sentiment information improves contextual understanding.

[0569] Step 5:

[0570] The server generates appropriate responses to the user using warning and question-providing mechanisms. The input is the detection results and emotional state obtained in step 4, and the output is a customized warning message and question. This generative AI model enables communication in an emotionally sensitive tone. For example, it can provide voice messages in a gentle tone.

[0571] Step 6:

[0572] The terminal visually or audibly notifies the user of warning messages and suggested questions received from the server. Input is the generated response message, and output is the notification to the user. The terminal communicates information to the user through screen display and speech synthesis. Notifications are provided in an easy-to-understand format using a user interface.

[0573] Step 7:

[0574] Users provide feedback to the system through a feedback receiving mechanism. The input is the feedback information delivered by the user, and the output is feedback data that contributes to system improvement. The server analyzes this feedback along with sentiment analysis and uses the findings to improve the system's accuracy. The submitted feedback is stored in a database.

[0575] (Application Example 2)

[0576] 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."

[0577] Fraud and dishonest behavior are significant issues in daily life, and it is essential for users to be able to detect these behaviors during conversations and take appropriate action. However, conventional technologies do not provide warnings or instructions that take into account the user's emotional state, which can cause stress. Therefore, the present invention aims to provide a system that takes the user's emotional state into consideration and provides more accurate and appropriate warnings and support.

[0578] 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.

[0579] In this invention, the server includes voice data collection means, voice data analysis means, data analysis means, notification and question provision means, opinion receiving means, and sentiment analysis means. This makes it possible to detect fraudulent activity and provide warnings in an appropriate tone while taking into account the user's emotional state in real time.

[0580] "Voice data collection means" refers to means for acquiring the user's voice in real time and transmitting it to a processing device.

[0581] "Voice data analysis means" refers to means for converting acquired voice information into text data.

[0582] "Data analysis means" refers to means for analyzing converted text data to detect terms or phrases that may indicate fraudulent or dishonest activity.

[0583] "Notification and Question Provisioning Means" are means of alerting users and providing more detailed questions based on detected terms or phrases.

[0584] A "means for receiving opinions" refers to a means equipped with a user interaction interface that allows users to submit opinions.

[0585] "Emotion analysis means" refers to a means for detecting the user's emotional state and adjusting the tone of notification messages based on that state.

[0586] The system for realizing this invention consists of a voice data collection means, a voice data analysis means, a data analysis means, a notification and question provision means, an opinion receiving means, and an emotion analysis means. The operation of the system is as follows.

[0587] First, the voice spoken by the user is collected as audio data using the microphone built into the device, and the audio data collection means transmits it to the processing unit. The audio data analysis means converts this audio data into text data using speech recognition software such as the Google Speech-to-Text API. At this time, the server processes the converted text data using the data analysis means and detects terms and phrases that suggest the possibility of fraudulent or dishonest activity using natural language processing technology.

[0588] Next, the server uses sentiment analysis tools to analyze the user's emotional state using software such as Microsoft Azure's Text Analytics. This allows warning messages related to detected terms to be adjusted based on the emotional state. The adjusted messages are then visually and audibly communicated to the user through the terminal's display and speakers via notification and question-providing tools.

[0589] Users can provide feedback through various means of receiving it. This feedback will be used to improve the system's algorithms.

[0590] For example, if a user receives a solicitation call and the conversation contains suspicious language, the system will detect this and provide a message such as, "There may be some risks involved in what you're offering. Please consider it carefully." Furthermore, prompts such as, "Generate a reassuring message based on emotion data indicating whether the user is excited or nervous," will be used for the generative AI model.

[0591] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0592] Step 1:

[0593] The device collects the user's voice in real time via a microphone. The input is the user's voice data, which is converted into digital data as an audio signal. The output is digital audio data ready to be sent to the server.

[0594] Step 2:

[0595] The server receives audio data and uses an audio data analysis tool to execute speech recognition software such as the Google Speech-to-Text API. Here, the input is digital audio data, and this data is analyzed to output text data.

[0596] Step 3:

[0597] The server receives text data obtained by the voice data analysis system and uses data analysis tools to detect terms and phrases that suggest potential fraudulent or dishonest activities using natural language processing techniques. The input is text data, and the output is a list of detected keywords.

[0598] Step 4:

[0599] The server uses sentiment analysis tools to perform sentiment analysis on the obtained text data using software such as Microsoft Azure's Text Analytics. The input is text data, and the output is data indicating the user's emotional state.

[0600] Step 5:

[0601] The server integrates the information obtained in steps 3 and 4 and utilizes notification and question provision mechanisms to generate warnings and questions for the user. The input consists of detected keywords and sentiment state data, which are then output as warning messages adjusted to an appropriate tone.

[0602] Step 6:

[0603] The device receives the generated warning message and notifies the user visually and audibly using its display and speaker. The input is the warning message, and the output is the visual and audible notification to the user.

[0604] Step 7:

[0605] Users can provide feedback through a feedback reception system. This feedback is sent to the server and used to improve the system's algorithms. The input is user feedback data, and the output is data for system improvement.

[0606] 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.

[0607] 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.

[0608] 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.

[0609] [Fourth Embodiment]

[0610] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0611] 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.

[0612] 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).

[0613] 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.

[0614] 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.

[0615] 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).

[0616] 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.

[0617] 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.

[0618] 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.

[0619] 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.

[0620] 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.

[0621] 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.

[0622] 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".

[0623] The system of the present invention has the function of monitoring a user's voice conversation in real time and detecting and issuing warnings for fraudulent or inappropriate behavior. This system is realized through the following main components and their operation.

[0624] First, when a user begins speaking through the device, the device uses voice acquisition tools to collect voice data in real time. This voice data is immediately sent to the server.

[0625] The server accurately converts the audio data received through the speech recognition means into text, and the text information is processed by the analysis means. The analysis means uses natural language processing techniques to analyze the text data in detail and detect words and phrases that may contain fraudulent or malicious elements.

[0626] If detection occurs, the server will utilize warning and question provision mechanisms to display a warning message to the user and provide the user with specific questions that can be used to verify the veracity of the other party's statements.

[0627] The device uses this information to immediately notify the user of any potential risks. The user can then choose to continue the conversation or interrupt it by following the instructions received.

[0628] Furthermore, the system includes a means for receiving user feedback, and the feedback provided by users is used to improve the system's accuracy and functionality.

[0629] As a concrete example, consider a scenario where a user is told over the phone, "We need to verify your bank information, so please tell us your account number." In this case, the system detects keywords such as "bank" and "account number," issues a warning to the user, and further asks questions such as, "Which bank is calling from?" to help reduce the likelihood of becoming a victim of fraud.

[0630] In this way, the system of the present invention provides an effective means for protecting users from fraud and illegal activities.

[0631] The following describes the processing flow.

[0632] Step 1:

[0633] The device acquires the user's voice in real time through the microphone. This voice data is converted into packets at regular time intervals (for example, every few seconds) and sent to the server.

[0634] Step 2:

[0635] When the server receives audio data, it converts the audio into text data via speech recognition. This conversion process also removes background noise and corrects for differences in accent and intonation.

[0636] Step 3:

[0637] The server processes the converted text data using analysis tools. During this analysis process, natural language processing techniques are used to detect contextually relevant keywords and phrases, identifying elements that may be fraudulent or malicious.

[0638] Step 4:

[0639] If detection occurs, the server generates a warning message for the user using warning and question provision mechanisms. At the same time, it prepares to formulate and provide the user with specific and easy-to-understand questions.

[0640] Step 5:

[0641] The terminal receives warning messages and suggested questions sent from the server and immediately notifies the user through a visual or auditory interface.

[0642] Step 6:

[0643] Users can use the received warnings and suggested questions to decide whether to continue or discontinue the conversation with the other party. Users can also use the suggested questions to verify the other party's identity and intentions.

[0644] Step 7:

[0645] When a user provides feedback, the device collects that information and sends it to the server. The server then analyzes the received feedback and uses it to improve the system and increase its accuracy.

[0646] (Example 1)

[0647] 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".

[0648] In modern society, fraudulent activities and scams targeting users are on the rise, creating a need for real-time monitoring methods to effectively address them. However, current methods make it difficult to immediately warn users and provide concrete countermeasures. Furthermore, continuous system improvement utilizing feedback is not being adequately implemented. These challenges need to be addressed.

[0649] 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.

[0650] In this invention, the server includes voice acquisition means, voice recognition means, analysis means, warning and inquiry suggestion provision means, evaluation receiving means, and means for improving analysis accuracy and warning quality using a generated AI model. This enables real-time protection of users from fraud and misconduct, immediate provision of countermeasures, and continuous improvement of the system based on feedback.

[0651] "Voice acquisition means" refers to a device or method that acquires and transmits voice information from a user in real time.

[0652] "Speech recognition means" refers to a technology or process for converting speech information into symbolic information.

[0653] "Analysis means" refers to a technique or process that analyzes symbolic information to detect identifiers that indicate fraud or fraudulent activity.

[0654] "Warning and Inquiry Provision Methods" refers to technologies or processes that warn users of potential fraudulent activity and provide specific methods for making inquiries.

[0655] "Evaluation receiving means" refers to an interface or method for receiving evaluations from users.

[0656] A "generative AI model" is an artificial intelligence technology used to improve analysis accuracy and the quality of warnings based on feedback.

[0657] The system of this invention is designed to protect users from fraud and illegal activities by analyzing voice information in real time and providing immediate countermeasures. The specific implementation of the system is described below.

[0658] The terminal is equipped with a microphone and voice input device, and when the user begins to speak, voice data is collected in real time through these voice acquisition means. This voice data is transmitted to the server using a secure communication protocol. The communication means incorporate commonly used encryption technologies.

[0659] The server uses a speech recognition engine (e.g., a commercial speech recognition API) to convert the received audio data into text. At this stage, a high-precision language model is utilized to remove noise and support multiple languages. The converted text data is then analyzed in detail using natural language processing techniques. This process utilizes Python's NLTK and spaCy libraries to detect keywords and phrases that could potentially lead to fraud or misconduct.

[0660] If any malicious elements are detected, the server uses an AI model to generate a warning message. This message includes an immediate alert to the user and specific questions to help them respond in the conversation. The questions provided may include those that allow the user to verify the legitimacy of the other party, such as, "Which organization is contacting you?"

[0661] Based on the warning messages and suggested questions displayed on the device, users can choose the appropriate course of action to take with the other party. This information is communicated visually and audibly, allowing users to respond immediately.

[0662] In addition, the terminal is equipped with an interface for receiving user feedback. This feedback is sent to a server and used by an AI model to improve the system's accuracy. For example, prompts such as "Suggest improvements to the warning system based on recent fraud patterns" are input to the AI ​​to improve the system.

[0663] In this way, the system of the present invention provides a means to guide the user into a safe communication environment.

[0664] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0665] Step 1:

[0666] The terminal collects the user's speech in real time using a speech acquisition device. The input data is the user's voice, which the terminal converts into a digital audio signal and immediately sends to the server. The device used at this stage is the microphone built into the terminal. The output is digital audio data.

[0667] Step 2:

[0668] The server converts transmitted audio data into text data using a speech recognition engine. The input is digital audio data, and the server applies speech recognition technology to generate corresponding text from the audio signal through a language model. The output is the converted text data. This process involves background noise removal and accurate pronunciation recognition.

[0669] Step 3:

[0670] The server applies natural language processing techniques to analyze text data and detect specific keywords and phrases. The input is text data, and the data analysis process considers context and intent. It utilizes Python's NLTK and spaCy libraries to identify elements that indicate fraud or misconduct. The output consists of the detected fraudulent keywords and phrases.

[0671] Step 4:

[0672] The server generates warning messages and specific questions based on the detection results. The input is identified keywords or phrases, and a generative AI model is used to design appropriate warnings and questions. The warning message serves as a reminder to the user, and the questions help verify the legitimacy of the other party. The output is the warning message and questions to be displayed to the user.

[0673] Step 5:

[0674] The terminal receives warning messages and suggested questions from the server and notifies the user visually and audibly. Input is message data from the server, which the terminal presents to the user through its interface. Based on this, the user can select an appropriate response. Output is screen displays and audio notifications that the user can see.

[0675] Step 6:

[0676] Based on the information provided, the user decides whether to continue or interrupt the conversation and provides feedback to the system via the terminal. This feedback, containing information about the user's choices and actions, is sent to the server. The system receives this feedback and uses a generative AI model to improve the system's analysis accuracy and the quality of its warnings. The output represents the improved system performance.

[0677] (Application Example 1)

[0678] 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".

[0679] In modern society, fraud and dishonest practices are commonplace, increasing the risk of users becoming victims. In particular, detecting fraudulent behavior in advance and responding quickly is difficult in voice communication. Current technology is insufficient to detect fraudulent activity in real time and issue immediate warnings to users; therefore, an efficient solution is needed to ensure user safety.

[0680] 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.

[0681] In this invention, the server includes voice data acquisition means for receiving user voice in real time and transmitting it to the server, voice recognition means for converting the voice information transmitted from the voice data acquisition means into text, and analysis means for analyzing the converted text and detecting words or expressions that indicate fraud or misconduct. This enables early detection of fraudulent activity and rapid notification of warnings to the user.

[0682] A "user" is an individual or legal entity that uses the system to conduct voice communication.

[0683] "Voice data acquisition means" refers to devices or software that receive the user's voice in real time and transmit it to a server.

[0684] A "server" is a computer system used for data processing and communication.

[0685] "Speech recognition means" refers to a technology or device for converting speech data into text.

[0686] "Analysis means" refers to techniques or devices for analyzing converted text and detecting words or expressions that may indicate fraud or misconduct.

[0687] A "warning and question supply means" is a means of notifying the user of an alert based on detected words or expressions and providing specific examples of questions.

[0688] A "feedback receiving means" is a device or system equipped with an interface that allows users to provide comments.

[0689] A "cloud environment" is an environment that utilizes computing resources provided over the internet.

[0690] This invention is a system that protects users by detecting fraudulent or illegal behavior in real time and issuing warnings when users engage in voice communication using devices such as smartphones. Specific embodiments of this system are shown below.

[0691] When a user initiates a call via their smartphone, the audio is received in real time by an audio data acquisition system. The received audio is immediately transmitted to the server. At this stage, the smartphone uses a communication module to maintain a stable connection.

[0692] The server instantly converts audio data into text using speech recognition tools such as Google Cloud Speech-to-Text. This allows for real-time transcription of voice communication. The converted text is then analyzed using natural language processing techniques (e.g., SpaCy and BERT). In this analysis, the text is scrutinized for potentially fraudulent or deceptive words and phrases based on their context.

[0693] If the analysis detects fraudulent content, the server will issue visual and auditory warnings to the user's smartphone using warning and question-providing mechanisms. At the same time, it will provide the user with specific questions, such as "Which organization is calling from?", to help them identify suspicious points in the conversation.

[0694] Users can provide feedback to the system through feedback receiving mechanisms. This feedback is stored and analyzed on the server for continuous improvement of the system. The data for improvement is accumulated in a cloud environment and used to train the generated AI model.

[0695] For example, if a user calls customer support for an online shopping site and is asked for their credit card information, the system will detect this phrase and display a warning to the user saying, "We recommend using the official website." This helps protect the user from unintentional fraud.

[0696] An example of a prompt message is: "Identify the phrase indicating fraudulent activity from the following phone conversation: 'An error has occurred in our system. We need to verify your credit card information immediately.'"

[0697] In this way, this invention provides a means to prevent fraudulent activities in users' voice communications.

[0698] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0699] Step 1:

[0700] The terminal receives audio in real time using an audio data acquisition mechanism as soon as the user initiates a call. The input is an audio signal, and the output is digitized audio data. The audio data is immediately transmitted to the server.

[0701] Step 2:

[0702] The server converts received digital audio data into text using speech recognition technology. The input is digital audio data, and the output is text data. Here, the Google Cloud Speech-to-Text service is used to perform the audio-to-text conversion.

[0703] Step 3:

[0704] The server analyzes the converted text data using parsing tools. The input is text data, and the output is a flag indicating whether the text indicates fraudulent or illegal behavior. SpaCy and BERT are used for the analysis, and natural language processing techniques are employed to identify words and phrases.

[0705] Step 4:

[0706] Based on the analysis results, the server uses warning and question-providing mechanisms to send visual and auditory warnings to the terminal. The input is a flag indicating inappropriate behavior, and the output is a warning message and specific questions for the user. This results in a notification being displayed and played on the user's device.

[0707] Step 5:

[0708] The user chooses whether to continue or interrupt the conversation based on the provided warnings and suggested questions. The input consists of warning messages and suggested questions, while the output is the action chosen by the user. The user can ask additional questions to ensure safety.

[0709] Step 6:

[0710] The terminal sends feedback collected from users to a server via a feedback receiving mechanism. The input is feedback information provided by the user, and the output is the cloud storage of the feedback data. This data will be used for future system improvements.

[0711] Step 7:

[0712] The server analyzes feedback data to train and improve the AI ​​model generated in the cloud environment. The input is feedback data, and the output is an updated model for system improvement. The analysis results will be used to adjust the model and improve prompt messages in the future.

[0713] 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.

[0714] The system of the present invention not only monitors user conversations in real time and detects signs of fraud or misconduct, but also recognizes the user's emotional state and adjusts its response accordingly. This system is implemented with the following components:

[0715] First, when a user initiates a conversation via the device, the device uses voice acquisition means to collect the user's voice data in real time and sends the data to the server.

[0716] The server uses speech recognition to convert the audio data into text. During this process, an emotion engine analyzes the audio data and uses factors such as tone of voice, speaking speed, and emphasis patterns to identify the user's emotional state.

[0717] The converted text data is processed by an analysis tool to detect keywords or phrases indicating fraud or misleading expressions. This detection takes into account emotional state information from an emotion engine, enabling more contextually accurate analysis.

[0718] Based on the detection results, the server generates warnings and suggested questions for the user using warning and question provision mechanisms. At this time, based on information from the emotion engine, notifications are made in a tone and content that matches the user's current emotional state.

[0719] The device receives the generated warning messages and suggested questions and notifies the user visually or audibly. The user can then use this information to make informed decisions.

[0720] Furthermore, by using emotion engine data when users provide feedback, the emotional nuances of the feedback are also taken into consideration, which can be used to improve the system.

[0721] For example, if a user receives a suspicious phone call and is asked, "Would you like to receive a limited offer?", the system will detect the keyword "limited offer," which suggests potential fraud. If the emotion engine senses tension in the user's voice, it will respond by providing a gentle warning such as, "Please relax and think about it. Please check specifically what the offer is." In this way, the system can not only effectively protect users from fraud but also provide emotional support.

[0722] The following describes the processing flow.

[0723] Step 1:

[0724] The device acquires the user's voice in real time via the microphone and prepares to send this audio data to the server. During this process, noise reduction is also performed to maintain audio clarity.

[0725] Step 2:

[0726] The server converts the received audio data into text format using speech recognition technology. This conversion uses a language model and focuses on accurately transcribing the user's utterances into text.

[0727] Step 3:

[0728] The server analyzes the converted text data using analytical tools. This analysis detects keywords and phrases that indicate fraud or scams. Natural language processing techniques are used to consider the context and aim for higher accuracy in detection.

[0729] Step 4:

[0730] Simultaneously, the server activates an emotion engine, analyzing the tone, speed, and pacing of the audio data to determine the user's emotional state. The analysis results indicate the user's current emotions and influence the next steps.

[0731] Step 5:

[0732] The server generates warnings and suggested questions based on information from the emotion engine. This generation process takes into account the user's emotional state and includes reassuring language and suggestions to support the user.

[0733] Step 6:

[0734] The terminal receives warning messages and suggested questions sent from the server and presents them to the user through visual and auditory interfaces. For example, it may display text on the screen while simultaneously providing important information via audio.

[0735] Step 7:

[0736] Based on the information presented, users can decide whether to continue or end the conversation with the other party. They can also report their thoughts and suggestions for improvement through the system's feedback function as needed.

[0737] Step 8:

[0738] The server collects user feedback and uses it as data to improve the system's analytical capabilities and user interface. This feedback, in particular, contributes to improving the accuracy of the emotion engine.

[0739] (Example 2)

[0740] 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".

[0741] While it is crucial to detect signs of fraud and misconduct early in user conversations, traditional systems often issue warnings without considering the user's emotional state, potentially leading to misunderstandings and stress. Therefore, there is a need to more effectively protect users by providing appropriate warnings and feedback that are tailored to their emotional state.

[0742] 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.

[0743] In this invention, the server includes speech recognition means, emotion analysis means, analysis means, warning and question provision means, and feedback receiving means. This enables the provision of accurate warnings and questions that take into account the user's emotional state, thereby more effectively protecting the user and facilitating system improvement using feedback.

[0744] A "voice acquisition means" is a device that receives the user's voice in real time and transmits it to an information processing device.

[0745] "Speech recognition means" refers to a technology or device that analyzes received speech data and converts it into text data.

[0746] "Emotional analysis means" refers to a technology or device that analyzes voice data and uses voice tone, speaking speed, and emphasis patterns to identify the user's emotional state.

[0747] "Analysis means" refers to a technology or device that analyzes text data, detects keywords or phrases indicating fraud or scams, and performs contextual analysis.

[0748] "Warning and Question Provisioning Means" refers to a technology or device that notifies the user of a customized warning and provides specific questions based on detected keywords or phrases and emotional states.

[0749] "Feedback receiving means" refers to a technology or device that has an interface for users to provide feedback and to receive that feedback while taking emotional nuances into consideration.

[0750] The system of this invention not only analyzes user conversations from audio to detect signs of fraud or dishonest behavior, but also has the function of understanding the user's emotional state and providing appropriate responses. The specific configuration of the system includes a terminal, a server, and various analysis engines.

[0751] When a user begins a conversation, the device collects voice data in real time through its built-in microphone and voice acquisition software. This voice data is compressed and packetized as initial processing before being sent to the server. The server has powerful processing capabilities and utilizes common speech recognition APIs, particularly for speech recognition. Specific examples include Google Cloud Speech-to-Text and Amazon Transcribe.

[0752] The server converts the received audio data into text data, then uses an emotion analysis engine to analyze the tone, volume, and speaking speed of the voice to identify the user's emotional state. This analysis engine is specialized in extracting emotions from speech using machine learning models and specific algorithms.

[0753] The converted text data is analyzed using natural language processing (NLP) techniques to detect keywords and phrases that indicate tendencies toward fraudulent activity or deception. This analysis also considers emotional state information, enabling more accurate contextual analysis. Based on the analysis results, the server generates warning messages and suggested questions tailored to the user's emotional state. This process utilizes a generative AI model to generate user-specific responses.

[0754] The generated warning messages and proposed questions are sent to the terminal and notified to the user visually or audibly. The user can then make their own decisions based on this notification. Furthermore, user feedback is sent to the server via the terminal, where sentiment analysis is also performed. This feedback is used for the continuous improvement of the system.

[0755] For example, if a user receives a sales call and is asked, "Would you like to receive a limited offer?", the system will detect the keyword "limited offer." At the same time, if it detects tension in the user's voice, it will provide a warning such as, "Please relax and think carefully about what you're going to say. We recommend that you clearly confirm the details." In this way, the system ensures the user's safety while also providing emotional support.

[0756] An example of a prompt message would be, "How should the system generate a warning if a user receives a potentially fraudulent phone call?"

[0757] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0758] Step 1:

[0759] The user uses a device to initiate a conversation. The device collects the conversation in real time using a voice acquisition device. The input is the user's voice, and voice data is generated as output. This voice data undergoes noise reduction and echo cancellation through digital signal processing before being sent to the server.

[0760] Step 2:

[0761] The server passes the audio data received from the terminal to the speech recognition system. The input is audio data, which is converted into text data as output. During this conversion process, speech recognition algorithms are applied to recognize phonemes and words. Techniques used for the conversion include phoneme decomposition and hidden Markov models.

[0762] Step 3:

[0763] The text data is analyzed by sentiment analysis tools on the server. The input is the text data generated in step 2, and the output is information indicating the user's emotional state. In this process, the analysis engine evaluates the voice tone, speaking speed, and emphasis patterns to determine the user's emotions. For example, a machine learning model using an annotated dataset can predict emotions.

[0764] Step 4:

[0765] The server further analyzes the text data using analysis tools. The input is the sentiment information and text data from step 3, and the output is the detection results of keywords and phrases indicating signs of fraud or wrongdoing. This process uses natural language processing techniques to scan the data and perform dictionary matching and contextual analysis. Sentiment information improves contextual understanding.

[0766] Step 5:

[0767] The server generates appropriate responses to the user using warning and question-providing mechanisms. The input is the detection results and emotional state obtained in step 4, and the output is a customized warning message and question. This generative AI model enables communication in an emotionally sensitive tone. For example, it can provide voice messages in a gentle tone.

[0768] Step 6:

[0769] The terminal visually or audibly notifies the user of warning messages and suggested questions received from the server. Input is the generated response message, and output is the notification to the user. The terminal communicates information to the user through screen display and speech synthesis. Notifications are provided in an easy-to-understand format using a user interface.

[0770] Step 7:

[0771] Users provide feedback to the system through a feedback receiving mechanism. The input is the feedback information delivered by the user, and the output is feedback data that contributes to system improvement. The server analyzes this feedback along with sentiment analysis and uses the findings to improve the system's accuracy. The submitted feedback is stored in a database.

[0772] (Application Example 2)

[0773] 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".

[0774] Fraud and dishonest behavior are significant issues in daily life, and it is essential for users to be able to detect these behaviors during conversations and take appropriate action. However, conventional technologies do not provide warnings or instructions that take into account the user's emotional state, which can cause stress. Therefore, the present invention aims to provide a system that takes the user's emotional state into consideration and provides more accurate and appropriate warnings and support.

[0775] 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.

[0776] In this invention, the server includes voice data collection means, voice data analysis means, data analysis means, notification and question provision means, opinion receiving means, and sentiment analysis means. This makes it possible to detect fraudulent activity and provide warnings in an appropriate tone while taking into account the user's emotional state in real time.

[0777] "Voice data collection means" refers to means for acquiring the user's voice in real time and transmitting it to a processing device.

[0778] "Voice data analysis means" refers to means for converting acquired voice information into text data.

[0779] "Data analysis means" refers to means for analyzing converted text data to detect terms or phrases that may indicate fraudulent or dishonest activity.

[0780] "Notification and Question Provisioning Means" are means of alerting users and providing more detailed questions based on detected terms or phrases.

[0781] A "means for receiving opinions" refers to a means equipped with a user interaction interface that allows users to submit opinions.

[0782] "Emotion analysis means" refers to a means for detecting the user's emotional state and adjusting the tone of notification messages based on that state.

[0783] The system for realizing this invention consists of a voice data collection means, a voice data analysis means, a data analysis means, a notification and question provision means, an opinion receiving means, and an emotion analysis means. The operation of the system is as follows.

[0784] First, the voice spoken by the user is collected as audio data using the microphone built into the device, and the audio data collection means transmits it to the processing unit. The audio data analysis means converts this audio data into text data using speech recognition software such as the Google Speech-to-Text API. At this time, the server processes the converted text data using the data analysis means and detects terms and phrases that suggest the possibility of fraudulent or dishonest activity using natural language processing technology.

[0785] Next, the server uses sentiment analysis tools to analyze the user's emotional state using software such as Microsoft Azure's Text Analytics. This allows warning messages related to detected terms to be adjusted based on the emotional state. The adjusted messages are then visually and audibly communicated to the user through the terminal's display and speakers via notification and question-providing tools.

[0786] Users can provide feedback through various means of receiving it. This feedback will be used to improve the system's algorithms.

[0787] For example, if a user receives a solicitation call and the conversation contains suspicious language, the system will detect this and provide a message such as, "There may be some risks involved in what you're offering. Please consider it carefully." Furthermore, prompts such as, "Generate a reassuring message based on emotion data indicating whether the user is excited or nervous," will be used for the generative AI model.

[0788] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0789] Step 1:

[0790] The device collects the user's voice in real time via a microphone. The input is the user's voice data, which is converted into digital data as an audio signal. The output is digital audio data ready to be sent to the server.

[0791] Step 2:

[0792] The server receives audio data and uses an audio data analysis tool to execute speech recognition software such as the Google Speech-to-Text API. Here, the input is digital audio data, and this data is analyzed to output text data.

[0793] Step 3:

[0794] The server receives text data obtained by the voice data analysis system and uses data analysis tools to detect terms and phrases that suggest potential fraudulent or dishonest activities using natural language processing techniques. The input is text data, and the output is a list of detected keywords.

[0795] Step 4:

[0796] The server uses sentiment analysis tools to perform sentiment analysis on the obtained text data using software such as Microsoft Azure's Text Analytics. The input is text data, and the output is data indicating the user's emotional state.

[0797] Step 5:

[0798] The server integrates the information obtained in steps 3 and 4 and utilizes notification and question provision mechanisms to generate warnings and questions for the user. The input consists of detected keywords and sentiment state data, which are then output as warning messages adjusted to an appropriate tone.

[0799] Step 6:

[0800] The device receives the generated warning message and notifies the user visually and audibly using its display and speaker. The input is the warning message, and the output is the visual and audible notification to the user.

[0801] Step 7:

[0802] Users can provide feedback through a feedback reception system. This feedback is sent to the server and used to improve the system's algorithms. The input is user feedback data, and the output is data for system improvement.

[0803] 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.

[0804] 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.

[0805] 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.

[0806] 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.

[0807] 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.

[0808] 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.

[0809] 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.

[0810] 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.

[0811] 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."

[0812] 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.

[0813] 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.

[0814] 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.

[0815] 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.

[0816] 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.

[0817] 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.

[0818] 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.

[0819] 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.

[0820] 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.

[0821] 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.

[0822] 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.

[0823] 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.

[0824] The following is further disclosed regarding the embodiments described above.

[0825] (Claim 1)

[0826] A voice acquisition means that receives the user's voice in real time and sends it to a server,

[0827] A speech recognition means for converting speech data transmitted from the speech acquisition means into text,

[0828] An analytical means for analyzing the converted text and detecting keywords or phrases that indicate fraud or scams,

[0829] A warning and question provision means that notifies the user of a warning based on detected keywords or phrases and provides more specific question suggestions,

[0830] A feedback receiving means having an interface that allows users to provide feedback,

[0831] A system that includes this.

[0832] (Claim 2)

[0833] The system according to claim 1, wherein the analysis means performs context-aware analysis using natural language processing technology.

[0834] (Claim 3)

[0835] The system according to claim 1, wherein the warning and question provision means provides visual and auditory notification to the user.

[0836] "Example 1"

[0837] (Claim 1)

[0838] A voice acquisition means that receives the user's voice in real time and transmits it to an information processing device,

[0839] A speech recognition means for converting speech information transmitted from the speech acquisition means into symbolic information,

[0840] An analysis means for analyzing converted symbolic information and detecting identifiers that indicate fraud or fraudulent activity,

[0841] A means for issuing a warning and providing a suggested inquiry, which alerts the user based on the detected identifier and provides a more specific suggestion for inquiry.

[0842] An evaluation receiving means having an interface on which a user can provide an evaluation,

[0843] Based on the evaluation, we will use a generative AI model to improve the accuracy of the analysis and the quality of the warnings.

[0844] A system that includes this.

[0845] (Claim 2)

[0846] The system according to claim 1, wherein the analysis means performs context-aware analysis using natural language processing technology.

[0847] (Claim 3)

[0848] The system according to claim 1, wherein the aforementioned warning and inquiry provision means notifies the user visually and audibly.

[0849] "Application Example 1"

[0850] (Claim 1)

[0851] A means for acquiring voice data that receives the user's voice in real time and transmits it to a server,

[0852] A speech recognition means for converting speech information transmitted from the aforementioned speech data acquisition means into text,

[0853] An analytical means that analyzes the converted text and detects words or expressions that indicate fraud or scams,

[0854] A warning and question supply means that notifies the user of their attention based on detected words and expressions, and provides more specific examples of inquiries,

[0855] A feedback receiving means having an interface that allows users to provide comments,

[0856] A system that includes processing in a cloud environment.

[0857] (Claim 2)

[0858] The system according to claim 1, wherein the analysis means performs context-aware analysis using language processing technology.

[0859] (Claim 3)

[0860] The system according to claim 1, wherein the warning and question supply means provides visual and auditory notification to the user.

[0861] "Example 2 of combining an emotion engine"

[0862] (Claim 1)

[0863] A voice acquisition means that receives the user's voice in real time and transmits it to an information processing device,

[0864] A speech recognition means for converting speech data transmitted from the speech acquisition means into text data,

[0865] An emotion analysis method that analyzes voice data and identifies the user's emotional state using voice tone, speaking speed, and emphasis patterns,

[0866] An analysis means that analyzes converted character data, detects keywords and phrases indicating fraud or scams, and performs contextual analysis considering emotional state information,

[0867] A warning and question provision means that notifies the user of a customized warning based on detected keywords or phrases and emotional state, and further provides specific question suggestions,

[0868] A feedback receiving means having an interface that allows users to provide feedback and receives that feedback while taking emotional nuances into consideration,

[0869] A system that includes this.

[0870] (Claim 2)

[0871] The system according to claim 1, wherein the analysis means performs context-aware analysis using natural language processing technology and combines emotional state information to perform a more accurate analysis.

[0872] (Claim 3)

[0873] The system according to claim 1, wherein the warning and question provision means provides visual and auditory notifications in accordance with the user's emotional state.

[0874] "Application example 2 when combining with an emotional engine"

[0875] (Claim 1)

[0876] A voice data collection means that acquires the user's voice in real time and transmits it to a processing unit,

[0877] A voice data analysis means for converting voice information acquired by the voice data collection means into text data,

[0878] A data analysis means that analyzes converted text data and detects terms or phrases that indicate potential fraudulent or dishonest activity,

[0879] A notification and question provision mechanism that alerts the user based on detected terms and phrases and provides further detailed question suggestions,

[0880] A means for receiving opinions that includes a user interaction interface that allows users to submit opinions,

[0881] A sentiment analysis means that detects the emotional state and adjusts the tone of the notification message based on it,

[0882] A system that includes this.

[0883] (Claim 2)

[0884] The system according to claim 1, wherein the data analysis means performs context-aware analysis using natural language processing technology and further improves the accuracy of the analysis by utilizing emotional state information.

[0885] (Claim 3)

[0886] The system according to claim 1, wherein the notification and question provision means provides visual and auditory notifications to the user, and adjusts the content of the notifications based on information from the emotion analysis means. [Explanation of Symbols]

[0887] 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 voice acquisition means that receives the user's voice in real time and sends it to a server, A speech recognition means for converting speech data transmitted from the speech acquisition means into text, An analytical means for analyzing the converted text and detecting keywords or phrases that indicate fraud or scams, A warning and question provision means that notifies the user of a warning based on detected keywords or phrases and provides more specific question suggestions, A feedback receiving means having an interface that allows users to provide feedback, A system that includes this.

2. The system according to claim 1, wherein the analysis means performs context-aware analysis using natural language processing technology.

3. The system according to claim 1, wherein the warning and question provision means provides visual and auditory notification to the user.

Images