system

Abstract

This system performs voice recognition on phone calls, analyzes the content to determine whether action is needed, summarizes important information, and forwards it to the user. [Solution] A means for recognizing the content of an incoming call, A means for analyzing the content of a voice-recognized call using natural language processing, A means of determining whether action is necessary based on the analyzed call content, A means of generating and responding to an automated message for unwanted calls, A means of summarizing important call content and forwarding it to the user, 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, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In modern times, while the number of households and enterprises using landline telephones is decreasing, it is still a problem that many unwanted calls, such as sales calls, come in. Especially for the elderly and users with little technical knowledge, these unwanted calls can be a heavy burden. Also, it is difficult to efficiently handle important calls mixed in such a situation. There is a demand for a system that solves such problems and quickly and appropriately transfers only important calls for users.

Means for Solving the Problems

[0005] The present invention solves the above problems by providing a system that includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary from the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, and means for summarizing important call content and forwarding it to the user. Specifically, it implements an algorithm to determine whether an incoming call is a sales call, and while providing an appropriate automatic response for unnecessary calls, it is possible to summarize important call content and notify the user. As a result, the user can concentrate only on important calls and is not bothered by processing unnecessary calls. Furthermore, by recording the call content in real time and saving the content, it is easy to manage and review the call content at a later date.

[0006] "Received calls" refers to all calls that come into a landline phone.

[0007] "Means of speech recognition" refers to technical means that have the function of analyzing received voice data and converting it into text data.

[0008] "Means of analysis using natural language processing" refers to technical means for analyzing the semantic structure of language and understanding its content using text data obtained through speech recognition.

[0009] "Means for determining whether action is necessary" refers to an algorithm that uses natural language processing results to determine whether a call is important or unnecessary for the user.

[0010] "Means for generating and responding with automated response messages" refers to technical means for generating appropriate messages according to the type of call and transmitting those messages as voice.

[0011] "Means for summarizing important call content and forwarding it to the user" refers to technical means that have the function of summarizing the content of a call deemed important to the user and notifying the user of that summary on their device.

[0012] An "algorithm for determining whether a call is a sales call" refers to a set of computational rules used to determine whether a call is being made for sales purposes, based on text data obtained from speech recognition and natural language processing.

[0013] "Methods for recording call content in real time" refers to technical means for instantly recording audio data during a call in its original form.

[0014] "Means of saving call content" refers to data storage technology that allows recorded audio data to be stored for a long period of time. [Brief explanation of the drawing]

[0015] [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]Shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

Mode for Carrying Out the Invention

[0016] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0017] First, the terms used in the following description will be explained.

[0018] In the following embodiments, a processor with a reference numeral (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of a plurality of arithmetic units. Further, 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.

[0019] In the following embodiments, a RAM (Random Access Memory) with a reference numeral is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0020] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0021] 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).

[0022] 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."

[0023] [First Embodiment]

[0024] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0025] 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.

[0026] 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).

[0027] 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.

[0028] 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.

[0029] 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.

[0030] 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.

[0031] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] 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.

[0033] 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.

[0034] 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.

[0035] 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".

[0036] This invention relates to a system for efficiently managing incoming calls on a landline telephone, eliminating unwanted sales calls, and notifying the user only of important calls. Specific embodiments for implementing this invention are described below.

[0037] This system is primarily composed of three components: servers, terminals, and users.

[0038] server

[0039] 1. Answering incoming phone calls

[0040] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers.

[0041] 2. Speech Recognition

[0042] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This process makes the call content analyzable as text.

[0043] 3. Natural Language Processing (NLP)

[0044] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. This analysis determines whether the call is a sales call or not.

[0045] 4. Determining whether action is necessary.

[0046] Based on the results of natural language processing, the server determines whether the call is an unwanted sales call or an important call. If it is determined to be a sales call, it proceeds to the next step.

[0047] 5. Automated response

[0048] If the server determines that a call is a sales call, it automatically generates an appropriate response message and uses synthesized speech to answer. For example, it might respond with, "I'm sorry, but I can't help you right now."

[0049] 6. Summary of the call

[0050] If the server determines a call is important, it prepares to summarize the call and notify the user. The server extracts the main points of the call and generates a concise summary.

[0051] 7. Transfer of Summary

[0052] The server notifies the user's device of the generated summary. For example, it may send the summary to the user's mobile phone or email address.

[0053] terminal

[0054] 1. Receiving

[0055] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[0056] User

[0057] 1. Confirmation of notification

[0058] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0059] Specific example

[0060] 1. Power reception

[0061] When a relative calls the landline, the server receives the call and initiates the conversation.

[0062] 2. Speech Recognition

[0063] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0064] 3. Natural Language Processing

[0065] The server analyzes the converted text and understands that it is a call from a relative checking on the person's well-being.

[0066] 4. Determining whether action is necessary.

[0067] The server will appropriately determine that this call is important.

[0068] 5. Automated response

[0069] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[0070] 6. Summary of the call

[0071] After the call ends, the server summarizes the content and generates a summary such as, "You received a call from a relative. They called to check on your recent health."

[0072] 7. Transfer of Summary

[0073] The server notifies the user's mobile phone of the generated summary.

[0074] This processing flow allows users to efficiently receive important call information and avoid being bothered by unnecessary sales calls.

[0075] The following describes the processing flow.

[0076] Step 1: Answering incoming calls

[0077] The server receives incoming calls to the landline phone. The server detects that a call has been made through the landline phone's interface, activates the automated answering module, and initiates the call.

[0078] Step 2: Speech Recognition

[0079] The server records the call in real time. The server uses a speech recognition engine to convert the recorded audio data into text. This makes the call content analyzable as text data.

[0080] Step 3: Natural Language Processing

[0081] The server passes the converted text data to a natural language processing (NLP) algorithm. The server uses the NLP module to analyze the text content and extract important information.

[0082] Step 4: Determine whether action is necessary.

[0083] The server uses natural language processing results to determine whether the call is important or an unwanted call, such as a sales call. If the call is important, it proceeds to the next step.

[0084] Step 5: Automated response

[0085] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. Using synthesized speech, it will respond with phrases such as, "I'm sorry, but I can't assist you at this time."

[0086] Step 6: Summarize the call

[0087] If the server determines a call is important, it prepares to summarize the call. After the call ends, the server uses NLP to extract the key points and generates a summary to notify the user.

[0088] Step 7: Transfer the summary

[0089] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative has contacted you. They called to check on your recent health."

[0090] Step 8: Final check

[0091] Users can view summaries through their devices. This allows users to efficiently identify only important calls and be freed from unnecessary sales calls.

[0092] (Example 1)

[0093] 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."

[0094] Traditional landline telephone systems lack mechanisms to efficiently filter out unwanted sales calls and notify users only of important calls. As a result, users are often bothered by numerous unwanted calls and may miss important ones. A system is needed to solve this problem and optimize the user's communication experience.

[0095] 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.

[0096] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary from the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, means for using a speech recognition engine when converting the call content to text, means for using a generative AI model to generate the summary, and means for using a communication device to forward the summary notification. As a result, the user can quickly and efficiently obtain only important call information and communicate without being bothered by unnecessary sales calls.

[0097] "Receiving a call" refers to receiving a call on a landline telephone.

[0098] "Speech recognition" is a technology that converts speech data into text data.

[0099] "Natural language processing" is a technology that analyzes text data to understand its meaning.

[0100] "Determining whether action is necessary" refers to determining whether a call is important based on the analyzed content.

[0101] An "automatic response message" is a message that the system automatically generates and plays for the person on the other end of the call.

[0102] A "summary" is a document that extracts the main points of a phone call and presents them concisely.

[0103] A "speech recognition engine" refers to software or algorithms used to convert speech data into text data.

[0104] A "generative AI model" is an artificial intelligence model that performs summarization and other processing on given text data.

[0105] A "communication device" refers to hardware or software used to transfer generated information to a specific device or system.

[0106] Modes for carrying out the invention

[0107] This invention relates to a system that efficiently manages incoming calls to a landline telephone, eliminates unnecessary sales calls, and notifies the user only of important calls. The following describes a specific embodiment for implementing this invention. This system mainly consists of three components: a server, a terminal, and a user.

[0108] server

[0109] The server receives incoming calls to a landline phone and initiates the call. Upon initiating the call, the server plays a pre-configured automated response message. This can be done using, for example, Python libraries or communication APIs.

[0110] The content of the call is recorded in real time, and the recorded audio data is converted to text using a speech recognition engine such as Google® Cloud Speech-to-Text. In this process, the audio can be recorded using the Pydub library, uploaded to Google Cloud Storage, and then converted to text by calling an API.

[0111] The transformed text data is passed to a natural language processing (NLP) algorithm, such as the Hugging Face BERT model, to analyze the text. This then uses the Python Transformers library to determine whether the call is a sales call.

[0112] Based on the analysis results, the system determines whether the call is an unnecessary sales call or an important call. If it is determined to be a sales call, an automated response message can be generated and played back using a text-to-speech engine (e.g., Google Cloud Text-to-Speech). The text-to-speech engine uses an API to generate messages such as, "I'm sorry, but I can't help you right now."

[0113] If a call is deemed important, the server extracts the main points of the call and generates a summary using, for example, OpenAI's GPT-3®. The summary is compiled into a concise document and ready to be sent to the user. The generated summary is transferred to the user's device via SMS or email.

[0114] terminal

[0115] The terminal receives a summary sent from the server. Typically, this is received on the user's mobile phone or computer and can be viewed through a notification interface. On mobile phones, it's received as an SMS message, and on computers, it's displayed in an email client.

[0116] User

[0117] Users review summaries sent from the server via their mobile phones or computers. This allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0118] Specific examples of operation

[0119] The following shows a specific example of how this system works.

[0120] 1. When a relative calls the landline, the server receives the call and initiates the conversation.

[0121] 2. When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text using Google Cloud Speech-to-Text.

[0122] 3. The server analyzes the converted text using a BERT model and recognizes that it is a call from a relative checking on the recipient's well-being.

[0123] 4. The server determines that the call is important and responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[0124] 5. After the call ends, the server summarizes the call and generates a summary such as, "You received a call from a relative. They called to check on your recent health."

[0125] 6. The server sends the generated summary to the user's mobile phone via SMS.

[0126] Example of a prompt

[0127] "The following text is a transcript of a phone call received on a landline. Please summarize the main points of this call. Call content: 'How are you? How has your health been lately?'"

[0128] By inputting this prompt into the AI ​​generation model, important information can be extracted from the call content and notified to the user in an easy-to-understand format.

[0129] As described above, the present invention provides the effect of preventing users from missing important calls and freeing them from unwanted sales calls.

[0130] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0131] Step 1:

[0132] Answering incoming calls

[0133] The server receives incoming calls to a landline phone. The input is the landline phone's ringing signal, which the server detects and initiates the call. Upon initiating the call, the server plays a pre-configured automated response message. Specifically, it uses various Python libraries and communication APIs to detect the ringing signal and play the response message. The output is a call session in a waiting state.

[0134] Step 2:

[0135] Speech recognition

[0136] The server records incoming calls in real time and converts the audio data to text using a speech recognition engine such as Google Cloud Speech-to-Text. The input is the audio from the call, which is recorded using libraries such as Pydub. Once the recording is complete, the audio data is uploaded to Google Cloud Storage, and the Google Cloud Speech-to-Text API is called to convert the audio to text. The output is text data.

[0137] Step 3:

[0138] natural language analysis

[0139] The server passes the converted text data to a natural language processing algorithm for analysis. The input is the text data generated in step 2, which is analyzed using the Hugging Face BERT model. Specifically, the Python Transformers library is used to analyze the text using prompt phrases such as "Is this a sales call?". The output is the result of the analysis.

[0140] Step 4:

[0141] Determining whether action is necessary

[0142] The server determines whether a call is an important call or an unnecessary sales call based on the results of natural language processing. The input is the analysis result from step 3, and an algorithm is applied to determine whether it is an unnecessary sales call. The output is the result of the determination of whether the call is an important call or a sales call.

[0143] Step 5:

[0144] Automated response

[0145] If the server determines it is a sales call, it generates an automated response message and plays it back using a speech synthesis engine (e.g., Amazon Polly). The input is the result of the determination in step 4, which is specifically applied if it is a sales call. Specifically, it generates a message such as "I'm sorry, but I can't help you right now," synthesizes it into speech using the Amazon Polly API, and plays it back using the Twilio API. The output is the voice response to the sales call.

[0146] Step 6:

[0147] Summary of the call

[0148] The server prepares to summarize the call content and notify the user. The input is the result of the judgment in step 4, using the text that was determined to be from an important call as a prompt. Specifically, a generative AI model such as OpenAI's GPT-3 is used to input a prompt such as, "The following text is the content of a call received on a landline. Please summarize the main points of this call," and generates a summary. The output is the summarized text.

[0149] Step 7:

[0150] Summary transfer

[0151] The server notifies the user's device of the generated summary. The input is the summary text generated in step 6, which is formatted appropriately and sent to the user's mobile phone or email address. Specifically, this is done by sending an SMS using the Twilio API or by sending an email via an SMTP server. The output is the summary notification sent to the user's device.

[0152] (Application Example 1)

[0153] 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."

[0154] Traditional landline phones are often plagued by unwanted calls such as sales calls and scam calls, creating a risk of missing important calls. Furthermore, it's difficult to understand the content of important calls when the user is away or during busy periods. Therefore, there is a need for a system that efficiently manages incoming landline calls, automatically filters out unwanted sales and scam calls, and notifies the user only of important calls.

[0155] 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.

[0156] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, and means for notifying the smartphone of the call content summary. This enables efficient management of calls received on a landline, allowing the user to quickly grasp important call content without being bothered by unwanted sales calls or scam calls.

[0157] "Speech recognition" is a technology that analyzes speech data and converts it into text data.

[0158] "Natural language processing" is a technology that analyzes text obtained through speech recognition and understands and classifies its content.

[0159] An "automated response message" is a message that is automatically generated based on specific conditions and responds to the caller in voice.

[0160] A "summary" is information that extracts the main points of an important phone call and presents them concisely.

[0161] "Means of forwarding to the user" refers to the function of sending summarized call content to the user's device.

[0162] A "smartphone" is a portable information terminal that receives summaries of call content and notifies the user.

[0163] A "generative AI model" is an artificial intelligence algorithm that generates new information based on training data.

[0164] A "prompt statement" is an instruction given to a generative AI model and is used to obtain a specific output.

[0165] Modes for carrying out the invention

[0166] server

[0167] The server has key functions for efficiently managing incoming calls to landlines. This is done using the following hardware and software:

[0168] Speech Recognition: The server uses a speech recognition library (e.g., the speech_recognition library) to record incoming calls on the landline in real time and convert the content to text.

[0169] Natural Language Processing: The server analyzes the transcribed call content using a natural language processing library (e.g., the transformers library). This analysis determines whether the call is a sales call or a scam call.

[0170] Generative AI Model: The server uses a generative AI model to create automated response messages and summaries of important calls based on the analysis results. The generative AI model uses a pre-trained algorithm.

[0171] Summary Notification: If the server determines a call is important, it generates a summary and notifies the user's smartphone. The summarized call content is concisely presented to reduce the user's burden.

[0172] For example, if a relative calls to check on your well-being, the server will use speech recognition to process the message, such as "Hello, how have you been feeling lately?", and then analyze it as text. If the analysis determines that the call is important, it will generate a summary and notify your smartphone, "A relative has contacted you. They called to check on your recent health."

[0173] terminal

[0174] The terminal receives summaries sent from the server and notifies the user. This typically includes the user's smartphone or computer.

[0175] Notification Reception: The device receives a summarized call content and displays it to the user. Notifications are often displayed on the smartphone as push notifications.

[0176] User

[0177] Users can review a summary of the call content through their device and take necessary actions.

[0178] Notification Confirmation: Users can check notifications on their devices to understand the content of important calls. This allows them to respond only to important calls without being bothered by unwanted sales calls or scam calls.

[0179] As supplementary information, here are some examples of prompt statements that are input to the generative AI model:

[0180] "Please convert the following audio call into text and use natural language processing to determine if it is a sales call. If it is determined to be a sales call, please enter an automated response message."

[0181] This embodiment of the invention allows users to efficiently manage incoming calls on their landline and reliably identify only important calls.

[0182] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0183] Step 1:

[0184] The server receives incoming calls to a landline phone. The input is audio data from the landline. The server records this audio data in real time. The output is a recorded audio file. Specifically, the server uses an audio recording device to save the call.

[0185] Step 2:

[0186] The server uses a speech recognition library (e.g., speech_recognition) to convert recorded audio files into text. The input is a recorded audio file. The server uses a speech recognition engine to convert the audio data into text data. The output is the converted text. Specifically, the server initializes the speech recognition engine and analyzes and converts the audio data.

[0187] Step 3:

[0188] The server uses a natural language processing library (e.g., transformers) to analyze the transformed text. The input is the text obtained by speech recognition. The server uses a natural language processing engine to analyze the call content to determine whether it is a sales call or a scam call. The output is the analysis result. Specifically, the server uses a natural language processing model to classify the text data.

[0189] Step 4:

[0190] The server uses a generative AI model to generate automated response messages based on the analysis results. The input is the analysis results of natural language processing. The server inputs prompt sentences into the generative AI model to create automated response messages. The output is the generated automated response message. Specifically, the server provides prompt sentences to the generative AI model and generates an appropriate response.

[0191] Step 5:

[0192] The server summarizes the call content and prepares it for transfer to the smartphone. The input consists of analysis results and the text data of the call. The server uses a generative AI model to summarize the key points of the call and condense them into a concise format. The output is the summarized call content. Specifically, the server uses the generative AI model to extract the main points of the call and create a summary.

[0193] Step 6:

[0194] The server notifies the user's smartphone with a summary. The summarized call content is provided as input. The server uses push notifications to send a notification to the smartphone. The output is a summary notification displayed on the smartphone. Specifically, the server calls a notification API and sends the generated summary to the smartphone.

[0195] Step 7:

[0196] The user checks the notification on their smartphone and takes action as needed. The input is a summary of the notification. The user reviews the received summary to understand the important call content. The output is the user's action (e.g., returning the call). Specifically, the user opens the smartphone notification, reviews the call content, and decides on a course of action.

[0197] 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.

[0198] This invention relates to a system that efficiently manages incoming calls to a landline, eliminates unnecessary sales calls, and notifies the user only of important calls. Furthermore, by combining this with an emotion engine, the invention provides a system that can recognize the user's emotional state and respond accordingly.

[0199] server

[0200] 1. Answering incoming phone calls

[0201] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers.

[0202] 2. Speech Recognition

[0203] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This process makes the call content analyzable as text.

[0204] 3. Natural Language Processing (NLP)

[0205] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. This analysis determines whether the call is a sales call or not.

[0206] 4. Emotion recognition

[0207] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data and stores that information.

[0208] 5. Determining whether action is necessary.

[0209] The server determines whether the call is important or unwanted, such as a sales call, based on the results of natural language processing and sentiment recognition. If the call is important, it proceeds to the next step.

[0210] 6. Automated response

[0211] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. This message will use synthesized speech to say something like, "I'm sorry, but I can't assist you right now." It can also generate different messages depending on the user's emotional state.

[0212] 7. Summary of the call

[0213] If the server determines a call is important, it prepares to summarize the call content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user. The results of sentiment recognition can also be included in the summary.

[0214] 8. Transfer of Summary

[0215] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative has contacted you. They called to check on your recent health."

[0216] terminal

[0217] 1. Receiving

[0218] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[0219] User

[0220] 1. Confirmation of notification

[0221] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0222] Specific example

[0223] 1. Power reception

[0224] When a relative calls the landline, the server receives the call and initiates the conversation.

[0225] 2. Speech Recognition

[0226] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0227] 3. Natural Language Processing

[0228] The server analyzes the converted text data and understands that it is a phone call from a relative checking on the person's well-being.

[0229] 4. Emotion recognition

[0230] The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (for example, worry or relief).

[0231] 5. Determining whether action is necessary.

[0232] The server will appropriately determine that this call is important.

[0233] 6. Automated response

[0234] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message." It then generates a more appropriate response message depending on the relative's emotional state.

[0235] 7. Summary of the call

[0236] After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative contacted you. They called to check on your recent health. The person who contacted you was worried."

[0237] 8. Transfer of Summary

[0238] The server notifies the user's mobile phone of the generated summary.

[0239] This processing flow allows users to efficiently receive important call information and the caller's emotional state, freeing them from unnecessary sales calls.

[0240] The following describes the processing flow.

[0241] Step 1: Answering incoming calls

[0242] The server receives incoming calls to the landline phone. The server detects that a call has been made through the landline phone's interface, activates the automated answering module, and initiates the call.

[0243] Step 2: Speech Recognition

[0244] The server records the call in real time. The server uses a speech recognition engine to convert the recorded audio data into text. This makes the call content analyzable as text data.

[0245] Step 3: Natural Language Processing (NLP)

[0246] The server passes the converted text data to a natural language processing (NLP) algorithm. The server uses the NLP module to analyze the text content and extract important information and keywords.

[0247] Step 4: Emotion Recognition

[0248] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data and stores that information. For example, it identifies emotions such as "worried" or "angry."

[0249] Step 5: Determine whether action is necessary.

[0250] The server determines whether the call is important or unwanted, such as a sales call, based on the results of natural language processing and sentiment recognition. If the call is important, it proceeds to the next step.

[0251] Step 6: Automated response

[0252] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. This message will use synthesized speech to say something like, "I'm sorry, but I can't assist you right now." It can also generate different messages depending on the user's emotional state.

[0253] Step 7: Summarize the call

[0254] If the server determines a call is important, it prepares to summarize the call content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user. The results of sentiment recognition can also be included in the summary.

[0255] Step 8: Transferring the summary

[0256] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative contacted you. They called to check on your recent health. They were worried about you."

[0257] Step 9: Final Check

[0258] The user reviews the summary via their device. This summary includes the main points of the call and the caller's emotional state. This allows the user to efficiently grasp the important points of the call and the caller's emotional state.

[0259] (Example 2)

[0260] 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".

[0261] The challenge lies in efficiently distinguishing between unwanted sales calls and important calls on landlines, thereby saving users time and effort. Furthermore, it is necessary to analyze the emotional state of the caller to provide more appropriate responses.

[0262] 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.

[0263] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, means for analyzing the emotional state during the call, and means for generating summary information including the emotion analysis results. This makes it possible to eliminate unnecessary sales calls and efficiently grasp important call content. Furthermore, it is possible to respond based on the emotions of the caller, thereby improving user satisfaction.

[0264] "Received call" refers to the audio signal that the server receives when a phone call comes in.

[0265] "Speech recognition" is a technology that converts received audio signals into text data.

[0266] "Natural language processing" is a technology that analyzes text data obtained through speech recognition to understand its content.

[0267] "Determining whether action is necessary" means deciding whether the analyzed call content is important or unnecessary.

[0268] An "automatic response message" is a message that a server automatically generates and sends to the caller in response to an unnecessary call.

[0269] A "summary" is information that extracts the main points of an important phone call and presents them concisely.

[0270] "Transferring to the user" means sending the generated summary information to the user's device.

[0271] "Analyzing emotional state" means analyzing the emotions of the person on the other end of a call from their voice or text and determining their emotional state.

[0272] The "sentiment analysis result" refers to the information obtained as a result of analyzing the emotional state.

[0273] The "summary information" refers to a short explanatory text that includes the main points of the call content and the result of the sentiment analysis.

[0274] This invention relates to a system that efficiently manages calls coming in to a landline phone, eliminates unnecessary sales calls, and notifies the user only of important calls. Furthermore, it provides a system that analyzes the emotional state of the call partner and makes responses according to the user's emotional state.

[0275] System Configuration

[0276] Hardware and Software

[0277] The server plays a central role in the system and uses the following main components:

[0278] 1. Landline phone receiver: Receives calls from the landline phone.

[0279] 2. Speech recognition engine: Such as Google Cloud Speech-to-Text.

[0280] 3. Natural language processing engine: An algorithm based on TENSORFLOW (registered trademark).

[0281] 4. Sentiment recognition engine: Such as IBM Watson (registered trademark).

[0282] 5. Automatic response function: Synthetic voice engines such as Asterisk and Amazon Polly.

[0283] 6. Database: Such as MongoDB, for storing call content and sentiment analysis results.

[0284] The terminal is a device for receiving notifications from the server and includes the following components:

[0285] 1. Mobile phone or computer: A device for the user to check the summary.

[0286] 2. Notification system: An API such as Firebase Cloud Messaging.

[0287] Program processing

[0288] When the server receives a call from a landline phone, it starts the automatic answering software such as Asterisk and starts the call. The call content is converted into text data in real time using Google Cloud Speech-to-Text. The converted text data is passed to a natural language processing engine based on TensorFlow and analyzed.

[0289] At the same time as determining whether the call is a sales call by natural language processing, the emotional state of the call recipient is also analyzed using an emotion recognition engine such as IBM Watson. Based on these results, the server determines whether the call is important or unnecessary.

[0290] If the call is a sales call, an appropriate automatic response message is generated using the synthetic voice engine of Amazon Polly and notified to the call recipient. In the case of an important call, a summary of the call content is generated and notified to the user including the emotional information.

[0291] The terminal receives the summary from the server using, for example, Firebase Cloud Messaging and makes it available for the user to check. For example, the mobile phone screen is notified with "There was a call to check on the health of a relative."

[0292] The user checks the summary notified through the terminal and takes follow-up actions as needed. With this system, the user is freed from unnecessary calls and can efficiently manage only important calls.

[0293] Specific example

[0294] 1. Receiving a call: When a relative calls the landline, the server receives the call and initiates the conversation.

[0295] 2. Voice Recognition: When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0296] 3. Natural Language Processing: The server analyzes the converted text data and understands that it is a call from a relative checking on the user's well-being.

[0297] 4. Emotion Recognition: The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (e.g., worry or relief).

[0298] 5. Determination of whether action is needed: The server appropriately determines that this call is important.

[0299] 6. Automated response: The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[0300] 7. Summary of Call Content: After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative called to check on your recent health. The person who called was worried."

[0301] 8. Summary Transfer: The server notifies the user's mobile phone of the generated summary.

[0302] Examples of prompts for generative AI models

[0303] Example prompt: "Describe a system that converts audio data into text data in real time, analyzes it, and then summarizes and notifies the user of the important content of the call."

[0304] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0305] Step 1:

[0306] Receiving call response

[0307] Input: Incoming call signal to a landline phone.

[0308] Specific operation: The server receives an incoming call to the landline phone.

[0309] Data processing / data calculation: Start automatic response software such as Asterisk and initiate a call.

[0310] Output: The automatic response message is sent to the call recipient.

[0311] Step 2:

[0312] Speech recognition

[0313] Input: Voice data collected in real-time during a call.

[0314] Specific operation: The server records the content of the call in real-time.

[0315] Data processing / data calculation: Use Google Cloud Speech-to-Text to convert voice data into text data.

[0316] Output: Call content data in string format.

[0317] Step 3:

[0318] Natural language processing (NLP)

[0319] Input: Text data recognized by speech.

[0320] Specific operation: The server passes the text data to a TensorFlow-based NLP algorithm.

[0321] Data processing / data calculation: Analyze the content of the call and apply an algorithm to determine whether it is a sales call.

[0322] Output: A flag indicating whether the call is a sales call (True / False).

[0323] Step 4:

[0324] emotion recognition

[0325] Input: Speech-recognized text data and audio data.

[0326] Specific operation: The server passes text data and audio data to IBM Watson's emotion recognition engine.

[0327] Data processing / data calculation: Analyze the emotional state of the person on the other end of the phone call.

[0328] Output: The emotional state of the person on the other end of the call (e.g., worried, relieved).

[0329] Step 5:

[0330] Determining whether action is necessary

[0331] Input: Results of natural language processing and emotion recognition.

[0332] Specific operation: The server determines the importance of the call based on the analysis results.

[0333] Data processing / data calculation: Apply logic to determine whether a call is important or unnecessary.

[0334] Output: A flag indicating the importance of the call (important / unnecessary).

[0335] Step 6:

[0336] Automated response

[0337] Input: If the call is deemed unnecessary.

[0338] Specific operation: The server uses Amazon Polly to generate an automated response message.

[0339] Data processing / data calculation: Outputs the generated response message using synthesized speech.

[0340] Output: Automated response to the caller.

[0341] Step 7:

[0342] Summary of the call

[0343] Input: Text data if the call is deemed important.

[0344] Specific operation: The server uses an NLP library such as SpaCy to extract the main points of the call.

[0345] Data processing / calculation: Summarize call content, incorporating emotional information into the final product.

[0346] Output: Summary data including call content and sentiment information.

[0347] Step 8:

[0348] Summary transfer

[0349] Input: Summary data.

[0350] Specific operation: The server uses the Firebase Cloud Messaging API to send a summary to the user's device.

[0351] Data processing / data calculation: Convert summarized data into an appropriate format and send it.

[0352] Output: A summary message displayed on the user's terminal.

[0353] Examples of prompts for generative AI models

[0354] Example prompt: "Describe a system that converts audio data into text data in real time, analyzes it, and then summarizes and notifies the user of the important content of the call."

[0355] (Application Example 2)

[0356] 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".

[0357] This invention relates to a system that efficiently manages incoming calls on landlines, eliminates unnecessary sales calls, and notifies the user only of important calls. It also aims to enable food delivery drivers to streamline customer interactions and quickly convey important information.

[0358] 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.

[0359] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for converting the content of a customer call from voice data to text data, means for analyzing the converted text data using sentiment recognition and generating a necessary response, and means for summarizing important call content and forwarding it to the user. This enables the management of incoming calls to landlines and improves the work efficiency of food delivery drivers.

[0360] "Content of received calls" refers to the audio information of calls received through the communication system.

[0361] "Means of speech recognition" refers to technologies for converting speech data into text data.

[0362] "Natural language processing" is a set of techniques for analyzing text data and understanding its content and meaning.

[0363] "Methods for determining whether action is necessary based on analyzed call content" refers to a method of determining the importance of the call content based on speech-recognized text.

[0364] "Means for generating and responding with automated response messages" refers to technology that automatically creates an appropriate response message based on analysis results and transmits that message to the caller.

[0365] "A means of summarizing important call content and forwarding it to the user" refers to a method of concisely summarizing the content of a call and notifying the user of that summary.

[0366] "Customer calls" refer to voice conversations between food delivery drivers and service users.

[0367] "Methods for converting audio data to text data" refers to technologies for converting the content of a phone call into text information.

[0368] "Emotion recognition" is a technology that determines a speaker's emotions from audio or text data.

[0369] "Means for generating necessary responses" refers to a method that automatically creates appropriate responses based on the results of emotion recognition and text analysis.

[0370] This invention is a system that efficiently manages incoming calls on landlines, eliminates unnecessary sales calls, and notifies the user only of important calls. It also enables food delivery drivers to streamline customer communication and quickly convey important information.

[0371] Server operation

[0372] 1. Answering incoming phone calls

[0373] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers. The hardware used includes server equipment and a communication interface. The software used is an automated voice response system.

[0374] 2. Speech Recognition

[0375] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This utilizes the speech_recognition library.

[0376] 3. Natural Language Processing (NLP)

[0377] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. The transformers library is used for this purpose.

[0378] 4. Emotion recognition

[0379] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data.

[0380] 5. Determining whether action is necessary.

[0381] Based on the results of natural language processing and emotion recognition, the server determines whether the call is important or unwanted, such as a sales call.

[0382] 6. Automated response

[0383] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. Using synthesized speech, it will respond with phrases such as, "I'm sorry, but I can't assist you right now." It is also possible to generate different messages depending on the user's emotional state.

[0384] 7. Summary of the call

[0385] If a call is deemed important, preparations are made to summarize its content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user.

[0386] 8. Transfer of Summary

[0387] The server notifies the user's device of the generated summary. Specifically, it sends a summary message to the user's mobile phone or email address.

[0388] Terminal operation

[0389] 1. Receiving

[0390] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[0391] User actions

[0392] 1. Confirmation of notification

[0393] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0394] Specific example

[0395] 1. Power reception

[0396] When a relative calls the landline, the server receives the call and initiates the conversation.

[0397] 2. Speech Recognition

[0398] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0399] 3. Natural Language Processing

[0400] The server analyzes the converted text data and understands that it is a phone call from a relative checking on the person's well-being.

[0401] 4. Emotion recognition

[0402] The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (for example, worry or relief).

[0403] 5. Determining whether action is necessary.

[0404] The server will appropriately determine that this call is important.

[0405] 6. Automated response

[0406] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message." It then generates a more appropriate response message depending on the relative's emotional state.

[0407] 7. Summary of the call

[0408] After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative contacted you. They called to check on your recent health. The person who contacted you was worried."

[0409] 8. Transfer of Summary

[0410] The server notifies the user's mobile phone of the generated summary.

[0411] Examples of prompt statements used

[0412] "Please upload the audio recording of your conversation with the customer."

[0413] Through this invention, it is possible to improve the efficiency of call management for landline telephones and the work of delivery personnel for food delivery services.

[0414] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0415] Step 1:

[0416] The server receives a call from a landline and initiates the call. This opens the call and prepares it for the subsequent steps. The input is the call signal from the telephone line, and the output is the call connection status.

[0417] Step 2:

[0418] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. The speech_recognition library is used for this process. The input is the audio data of the call, and the output is text data.

[0419] Step 3:

[0420] The server passes the speech-recognized text data to a natural language processing algorithm to analyze the call content. The transformers library is used here. The input is the text data obtained from speech recognition, and the output is the analysis result (semantic information of the call content).

[0421] Step 4:

[0422] The server uses an emotion engine to analyze the emotions of the person on the call. Specifically, it analyzes the speaker's emotional state from voice and text data. The input is text data and voice parameters, and the output is identified emotion information.

[0423] Step 5:

[0424] The server determines whether or not to respond to a call based on the results of natural language processing and emotion recognition. The input is the analysis results and emotion information, and the output is a response-required flag.

[0425] Step 6:

[0426] If the server determines from the call content that it is a sales call, it automatically generates an appropriate response message. This uses pre-synthesized voice. The input is a flag indicating whether a response is needed and text data, and the output is the response message.

[0427] Step 7:

[0428] If a call is deemed important, preparations are made to summarize the call content. Speech recognition and NLP tools are used to extract key points and generate a summary. Input is parsed text data and sentiment information, and output is the summarized text.

[0429] Step 8:

[0430] The server notifies the user's device of the generated summary. Specifically, it sends a summary message to the user's mobile phone or email. The input is the summarized text, and the output is a message that can be viewed on the user's receiving device.

[0431] Specific example

[0432] For example, the process flow when a relative calls and says, "Hi, how have you been feeling lately?" is as follows:

[0433] Step 1: The server receives the call and initiates the conversation.

[0434] Step 2: Record the conversation "Hello, how have you been feeling lately?" in real time and convert the audio data into text data.

[0435] Step 3: The converted text data is analyzed using a natural language processing algorithm to understand that it is a phone call from a relative checking on the user's well-being.

[0436] Step 4: The server analyzes the relative's tone of voice and speech patterns to determine if they are worried.

[0437] Step 5: Determine that this call is important and generate an appropriate response message that includes the relative's emotional state.

[0438] Step 6: Generate the answering message "We are currently unable to answer your call, but we will take your message" and send it to the caller.

[0439] Step 7: After the call ends, summarize the content and generate a summary such as, "A relative wanted to check on your recent health. The person who contacted them was worried."

[0440] Step 8: Notify the user of the summary on their mobile phone so they can efficiently grasp important call information.

[0441] In this way, the system of the present invention enables call management for landline telephones and improves the work efficiency of food delivery personnel.

[0442] 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.

[0443] 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.

[0444] 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.

[0445] [Second Embodiment]

[0446] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0447] 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.

[0448] 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).

[0449] 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.

[0450] 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.

[0451] 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).

[0452] 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.

[0453] 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.

[0454] 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.

[0455] 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.

[0456] 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.

[0457] 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".

[0458] This invention relates to a system for efficiently managing incoming calls on a landline telephone, eliminating unwanted sales calls, and notifying the user only of important calls. Specific embodiments for implementing this invention are described below.

[0459] This system is primarily composed of three components: servers, terminals, and users.

[0460] server

[0461] 1. Answering incoming phone calls

[0462] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers.

[0463] 2. Speech Recognition

[0464] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This process makes the call content analyzable as text.

[0465] 3. Natural Language Processing (NLP)

[0466] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. This analysis determines whether the call is a sales call or not.

[0467] 4. Determining whether action is necessary.

[0468] Based on the results of natural language processing, the server determines whether the call is an unwanted sales call or an important call. If it is determined to be a sales call, it proceeds to the next step.

[0469] 5. Automated response

[0470] If the server determines that a call is a sales call, it automatically generates an appropriate response message and uses synthesized speech to answer. For example, it might respond with, "I'm sorry, but I can't help you right now."

[0471] 6. Summary of the call

[0472] If the server determines a call is important, it prepares to summarize the call and notify the user. The server extracts the main points of the call and generates a concise summary.

[0473] 7. Transfer of Summary

[0474] The server notifies the user's device of the generated summary. For example, it may send the summary to the user's mobile phone or email address.

[0475] terminal

[0476] 1. Receiving

[0477] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[0478] User

[0479] 1. Confirmation of notification

[0480] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0481] Specific example

[0482] 1. Power reception

[0483] When a relative calls the landline, the server receives the call and initiates the conversation.

[0484] 2. Speech Recognition

[0485] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0486] 3. Natural Language Processing

[0487] The server analyzes the converted text and understands that it is a call from a relative checking on the person's well-being.

[0488] 4. Determining whether action is necessary.

[0489] The server will appropriately determine that this call is important.

[0490] 5. Automated response

[0491] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[0492] 6. Summary of the call

[0493] After the call ends, the server summarizes the content and generates a summary such as, "You received a call from a relative. They called to check on your recent health."

[0494] 7. Transfer of Summary

[0495] The server notifies the user's mobile phone of the generated summary.

[0496] This processing flow allows users to efficiently receive important call information and avoid being bothered by unnecessary sales calls.

[0497] The following describes the processing flow.

[0498] Step 1: Answering incoming calls

[0499] The server receives incoming calls to the landline phone. The server detects that a call has been made through the landline phone's interface, activates the automated answering module, and initiates the call.

[0500] Step 2: Speech Recognition

[0501] The server records the call in real time. The server uses a speech recognition engine to convert the recorded audio data into text. This makes the call content analyzable as text data.

[0502] Step 3: Natural Language Processing

[0503] The server passes the converted text data to a natural language processing (NLP) algorithm. The server uses the NLP module to analyze the text content and extract important information.

[0504] Step 4: Determine whether action is necessary.

[0505] The server uses natural language processing results to determine whether the call is important or an unwanted call, such as a sales call. If the call is important, it proceeds to the next step.

[0506] Step 5: Automated response

[0507] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. Using synthesized speech, it will respond with phrases such as, "I'm sorry, but I can't assist you at this time."

[0508] Step 6: Summarize the call

[0509] If the server determines a call is important, it prepares to summarize the call. After the call ends, the server uses NLP to extract the key points and generates a summary to notify the user.

[0510] Step 7: Transfer the summary

[0511] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative has contacted you. They called to check on your recent health."

[0512] Step 8: Final check

[0513] Users can view summaries through their devices. This allows users to efficiently identify only important calls and be freed from unnecessary sales calls.

[0514] (Example 1)

[0515] 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".

[0516] Traditional landline telephone systems lack mechanisms to efficiently filter out unwanted sales calls and notify users only of important calls. As a result, users are often bothered by numerous unwanted calls and may miss important ones. A system is needed to solve this problem and optimize the user's communication experience.

[0517] 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.

[0518] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary from the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, means for using a speech recognition engine when converting the call content to text, means for using a generative AI model to generate the summary, and means for using a communication device to forward the summary notification. As a result, the user can quickly and efficiently obtain only important call information and communicate without being bothered by unnecessary sales calls.

[0519] "Receiving a call" refers to receiving a call on a landline telephone.

[0520] "Speech recognition" is a technology that converts speech data into text data.

[0521] "Natural language processing" is a technology that analyzes text data to understand its meaning.

[0522] "Determining whether action is necessary" refers to determining whether a call is important based on the analyzed content.

[0523] An "automatic response message" is a message that the system automatically generates and plays for the person on the other end of the call.

[0524] A "summary" is a document that extracts the main points of a phone call and presents them concisely.

[0525] A "speech recognition engine" refers to software or algorithms used to convert speech data into text data.

[0526] A "generative AI model" is an artificial intelligence model that performs summarization and other processing on given text data.

[0527] A "communication device" refers to hardware or software used to transfer generated information to a specific device or system.

[0528] Modes for carrying out the invention

[0529] This invention relates to a system that efficiently manages incoming calls to a landline telephone, eliminates unnecessary sales calls, and notifies the user only of important calls. The following describes a specific embodiment for implementing this invention. This system mainly consists of three components: a server, a terminal, and a user.

[0530] server

[0531] The server receives incoming calls to a landline phone and initiates the call. Upon initiating the call, the server plays a pre-configured automated response message. This can be done using, for example, Python libraries or communication APIs.

[0532] The call content is recorded in real time, and the recorded audio data is converted to text using a speech recognition engine such as Google Cloud Speech-to-Text. In this process, the audio can be recorded using the Pydub library, uploaded to Google Cloud Storage, and then converted to text by calling an API.

[0533] The transformed text data is passed to a natural language processing (NLP) algorithm, such as the Hugging Face BERT model, to analyze the text. This then uses the Python Transformers library to determine whether the call is a sales call.

[0534] Based on the analysis results, the system determines whether the call is an unnecessary sales call or an important call. If it is determined to be a sales call, an automated response message can be generated and played back using a text-to-speech engine (e.g., Google Cloud Text-to-Speech). The text-to-speech engine uses an API to generate messages such as, "I'm sorry, but I can't help you right now."

[0535] If a call is deemed important, the server extracts the main points of the call and generates a summary using a method such as OpenAI's GPT-3. The summary is compiled into a concise document and ready to be sent to the user. The generated summary is transferred to the user's device via SMS or email.

[0536] terminal

[0537] The terminal receives a summary sent from the server. Typically, this is received on the user's mobile phone or computer and can be viewed through a notification interface. On mobile phones, it's received as an SMS message, and on computers, it's displayed in an email client.

[0538] User

[0539] Users review summaries sent from the server via their mobile phones or computers. This allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0540] Specific examples of operation

[0541] The following shows a specific example of how this system works.

[0542] 1. When a relative calls the landline, the server receives the call and initiates the conversation.

[0543] 2. When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text using Google Cloud Speech-to-Text.

[0544] 3. The server analyzes the converted text using a BERT model and recognizes that it is a call from a relative checking on the recipient's well-being.

[0545] 4. The server determines that the call is important and responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[0546] 5. After the call ends, the server summarizes the call and generates a summary such as, "You received a call from a relative. They called to check on your recent health."

[0547] 6. The server sends the generated summary to the user's mobile phone via SMS.

[0548] Example of a prompt

[0549] "The following text is a transcript of a phone call received on a landline. Please summarize the main points of this call. Call content: 'How are you? How has your health been lately?'"

[0550] By inputting this prompt into the AI ​​generation model, important information can be extracted from the call content and notified to the user in an easy-to-understand format.

[0551] As described above, the present invention provides the effect of preventing users from missing important calls and freeing them from unwanted sales calls.

[0552] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0553] Step 1:

[0554] Answering incoming calls

[0555] The server receives incoming calls to a landline phone. The input is the landline phone's ringing signal, which the server detects and initiates the call. Upon initiating the call, the server plays a pre-configured automated response message. Specifically, it uses various Python libraries and communication APIs to detect the ringing signal and play the response message. The output is a call session in a waiting state.

[0556] Step 2:

[0557] Speech recognition

[0558] The server records incoming calls in real time and converts the audio data to text using a speech recognition engine such as Google Cloud Speech-to-Text. The input is the audio from the call, which is recorded using libraries such as Pydub. Once the recording is complete, the audio data is uploaded to Google Cloud Storage, and the Google Cloud Speech-to-Text API is called to convert the audio to text. The output is text data.

[0559] Step 3:

[0560] natural language analysis

[0561] The server passes the converted text data to a natural language processing algorithm for analysis. The input is the text data generated in step 2, which is analyzed using the Hugging Face BERT model. Specifically, the Python Transformers library is used to analyze the text using prompt phrases such as "Is this a sales call?". The output is the result of the analysis.

[0562] Step 4:

[0563] Determining whether action is necessary

[0564] The server determines whether a call is an important call or an unnecessary sales call based on the results of natural language processing. The input is the analysis result from step 3, and an algorithm is applied to determine whether it is an unnecessary sales call. The output is the result of the determination of whether the call is an important call or a sales call.

[0565] Step 5:

[0566] Automated response

[0567] If the server determines it is a sales call, it generates an automated response message and plays it back using a speech synthesis engine (e.g., Amazon Polly). The input is the result of the determination in step 4, which is specifically applied if it is a sales call. Specifically, it generates a message such as "I'm sorry, but I can't help you right now," synthesizes it into speech using the Amazon Polly API, and plays it back using the Twilio API. The output is the voice response to the sales call.

[0568] Step 6:

[0569] Summary of the call

[0570] The server prepares to summarize the call content and notify the user. The input is the result of the judgment in step 4, using the text that was determined to be from an important call as a prompt. Specifically, a generative AI model such as OpenAI's GPT-3 is used to input a prompt such as, "The following text is the content of a call received on a landline. Please summarize the main points of this call," and generates a summary. The output is the summarized text.

[0571] Step 7:

[0572] Summary transfer

[0573] The server notifies the user's device of the generated summary. The input is the summary text generated in step 6, which is formatted appropriately and sent to the user's mobile phone or email address. Specifically, this is done by sending an SMS using the Twilio API or by sending an email via an SMTP server. The output is the summary notification sent to the user's device.

[0574] (Application Example 1)

[0575] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0576] Traditional landline phones are often plagued by unwanted calls such as sales calls and scam calls, creating a risk of missing important calls. Furthermore, it's difficult to understand the content of important calls when the user is away or during busy periods. Therefore, there is a need for a system that efficiently manages incoming landline calls, automatically filters out unwanted sales and scam calls, and notifies the user only of important calls.

[0577] 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.

[0578] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, and means for notifying the smartphone of the call content summary. This enables efficient management of calls received on a landline, allowing the user to quickly grasp important call content without being bothered by unwanted sales calls or scam calls.

[0579] "Speech recognition" is a technology that analyzes speech data and converts it into text data.

[0580] "Natural language processing" is a technology that analyzes text obtained through speech recognition and understands and classifies its content.

[0581] An "automated response message" is a message that is automatically generated based on specific conditions and responds to the caller in voice.

[0582] A "summary" is information that extracts the main points of an important phone call and presents them concisely.

[0583] "Means of forwarding to the user" refers to the function of sending summarized call content to the user's device.

[0584] A "smartphone" is a portable information terminal that receives summaries of call content and notifies the user.

[0585] A "generative AI model" is an artificial intelligence algorithm that generates new information based on training data.

[0586] A "prompt statement" is an instruction given to a generative AI model and is used to obtain a specific output.

[0587] Modes for carrying out the invention

[0588] server

[0589] The server has key functions for efficiently managing incoming calls to landlines. This is done using the following hardware and software:

[0590] Speech Recognition: The server uses a speech recognition library (e.g., the speech_recognition library) to record incoming calls on the landline in real time and convert the content to text.

[0591] Natural Language Processing: The server analyzes the transcribed call content using a natural language processing library (e.g., the transformers library). This analysis determines whether the call is a sales call or a scam call.

[0592] Generative AI Model: The server uses a generative AI model to create automated response messages and summaries of important calls based on the analysis results. The generative AI model uses a pre-trained algorithm.

[0593] Summary Notification: If the server determines a call is important, it generates a summary and notifies the user's smartphone. The summarized call content is concisely presented to reduce the user's burden.

[0594] For example, if a relative calls to check on your well-being, the server will use speech recognition to process the message, such as "Hello, how have you been feeling lately?", and then analyze it as text. If the analysis determines that the call is important, it will generate a summary and notify your smartphone, "A relative has contacted you. They called to check on your recent health."

[0595] terminal

[0596] The terminal receives summaries sent from the server and notifies the user. This typically includes the user's smartphone or computer.

[0597] Notification Reception: The device receives a summarized call content and displays it to the user. Notifications are often displayed on the smartphone as push notifications.

[0598] User

[0599] Users can review a summary of the call content through their device and take necessary actions.

[0600] Notification Confirmation: Users can check notifications on their devices to understand the content of important calls. This allows them to respond only to important calls without being bothered by unwanted sales calls or scam calls.

[0601] As supplementary information, here are some examples of prompt statements that are input to the generative AI model:

[0602] "Please convert the following audio call into text and use natural language processing to determine if it is a sales call. If it is determined to be a sales call, please enter an automated response message."

[0603] This embodiment of the invention allows users to efficiently manage incoming calls on their landline and reliably identify only important calls.

[0604] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0605] Step 1:

[0606] The server receives incoming calls to a landline phone. The input is audio data from the landline. The server records this audio data in real time. The output is a recorded audio file. Specifically, the server uses an audio recording device to save the call.

[0607] Step 2:

[0608] The server uses a speech recognition library (e.g., speech_recognition) to convert recorded audio files into text. The input is a recorded audio file. The server uses a speech recognition engine to convert the audio data into text data. The output is the converted text. Specifically, the server initializes the speech recognition engine and analyzes and converts the audio data.

[0609] Step 3:

[0610] The server uses a natural language processing library (e.g., transformers) to analyze the transformed text. The input is the text obtained by speech recognition. The server uses a natural language processing engine to analyze the call content to determine whether it is a sales call or a scam call. The output is the analysis result. Specifically, the server uses a natural language processing model to classify the text data.

[0611] Step 4:

[0612] The server uses a generative AI model to generate automated response messages based on the analysis results. The input is the analysis results of natural language processing. The server inputs prompt sentences into the generative AI model to create automated response messages. The output is the generated automated response message. Specifically, the server provides prompt sentences to the generative AI model and generates an appropriate response.

[0613] Step 5:

[0614] The server summarizes the call content and prepares it for transfer to the smartphone. The input consists of analysis results and the text data of the call. The server uses a generative AI model to summarize the key points of the call and condense them into a concise format. The output is the summarized call content. Specifically, the server uses the generative AI model to extract the main points of the call and create a summary.

[0615] Step 6:

[0616] The server notifies the user's smartphone with a summary. The summarized call content is provided as input. The server uses push notifications to send a notification to the smartphone. The output is a summary notification displayed on the smartphone. Specifically, the server calls a notification API and sends the generated summary to the smartphone.

[0617] Step 7:

[0618] The user checks the notification on their smartphone and takes action as needed. The input is a summary of the notification. The user reviews the received summary to understand the important call content. The output is the user's action (e.g., returning the call). Specifically, the user opens the smartphone notification, reviews the call content, and decides on a course of action.

[0619] 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.

[0620] This invention relates to a system that efficiently manages incoming calls to a landline, eliminates unnecessary sales calls, and notifies the user only of important calls. Furthermore, by combining this with an emotion engine, the invention provides a system that can recognize the user's emotional state and respond accordingly.

[0621] server

[0622] 1. Answering incoming phone calls

[0623] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers.

[0624] 2. Speech Recognition

[0625] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This process makes the call content analyzable as text.

[0626] 3. Natural Language Processing (NLP)

[0627] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. This analysis determines whether the call is a sales call or not.

[0628] 4. Emotion recognition

[0629] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data and stores that information.

[0630] 5. Determining whether action is necessary.

[0631] The server determines whether the call is important or unwanted, such as a sales call, based on the results of natural language processing and sentiment recognition. If the call is important, it proceeds to the next step.

[0632] 6. Automated response

[0633] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. This message will use synthesized speech to say something like, "I'm sorry, but I can't assist you right now." It can also generate different messages depending on the user's emotional state.

[0634] 7. Summary of the call

[0635] If the server determines a call is important, it prepares to summarize the call content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user. The results of sentiment recognition can also be included in the summary.

[0636] 8. Transfer of Summary

[0637] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative has contacted you. They called to check on your recent health."

[0638] terminal

[0639] 1. Receiving

[0640] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[0641] User

[0642] 1. Confirmation of notification

[0643] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0644] Specific example

[0645] 1. Power reception

[0646] When a relative calls the landline, the server receives the call and initiates the conversation.

[0647] 2. Speech Recognition

[0648] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0649] 3. Natural Language Processing

[0650] The server analyzes the converted text data and understands that it is a phone call from a relative checking on the person's well-being.

[0651] 4. Emotion recognition

[0652] The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (for example, worry or relief).

[0653] 5. Determining whether action is necessary.

[0654] The server will appropriately determine that this call is important.

[0655] 6. Automated response

[0656] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message." It then generates a more appropriate response message depending on the relative's emotional state.

[0657] 7. Summary of the call

[0658] After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative contacted you. They called to check on your recent health. The person who contacted you was worried."

[0659] 8. Transfer of Summary

[0660] The server notifies the user's mobile phone of the generated summary.

[0661] This processing flow allows users to efficiently receive important call information and the caller's emotional state, freeing them from unnecessary sales calls.

[0662] The following describes the processing flow.

[0663] Step 1: Answering incoming calls

[0664] The server receives incoming calls to the landline phone. The server detects that a call has been made through the landline phone's interface, activates the automated answering module, and initiates the call.

[0665] Step 2: Speech Recognition

[0666] The server records the call in real time. The server uses a speech recognition engine to convert the recorded audio data into text. This makes the call content analyzable as text data.

[0667] Step 3: Natural Language Processing (NLP)

[0668] The server passes the converted text data to a natural language processing (NLP) algorithm. The server uses the NLP module to analyze the text content and extract important information and keywords.

[0669] Step 4: Emotion Recognition

[0670] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data and stores that information. For example, it identifies emotions such as "worried" or "angry."

[0671] Step 5: Determine whether action is necessary.

[0672] The server determines whether the call is important or unwanted, such as a sales call, based on the results of natural language processing and sentiment recognition. If the call is important, it proceeds to the next step.

[0673] Step 6: Automated response

[0674] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. This message will use synthesized speech to say something like, "I'm sorry, but I can't assist you right now." It can also generate different messages depending on the user's emotional state.

[0675] Step 7: Summarize the call

[0676] If the server determines a call is important, it prepares to summarize the call content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user. The results of sentiment recognition can also be included in the summary.

[0677] Step 8: Transferring the summary

[0678] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative contacted you. They called to check on your recent health. They were worried about you."

[0679] Step 9: Final Check

[0680] The user reviews the summary via their device. This summary includes the main points of the call and the caller's emotional state. This allows the user to efficiently grasp the important points of the call and the caller's emotional state.

[0681] (Example 2)

[0682] 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".

[0683] The challenge lies in efficiently distinguishing between unwanted sales calls and important calls on landlines, thereby saving users time and effort. Furthermore, it is necessary to analyze the emotional state of the caller to provide more appropriate responses.

[0684] 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.

[0685] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, means for analyzing the emotional state during the call, and means for generating summary information including the emotion analysis results. This makes it possible to eliminate unnecessary sales calls and efficiently grasp important call content. Furthermore, it is possible to respond based on the emotions of the caller, thereby improving user satisfaction.

[0686] "Received call" refers to the audio signal that the server receives when a phone call comes in.

[0687] "Speech recognition" is a technology that converts received audio signals into text data.

[0688] "Natural language processing" is a technology that analyzes text data obtained through speech recognition to understand its content.

[0689] "Determining whether action is necessary" means deciding whether the analyzed call content is important or unnecessary.

[0690] An "automatic response message" is a message that a server automatically generates and sends to the caller in response to an unnecessary call.

[0691] A "summary" is information that extracts the main points of an important phone call and presents them concisely.

[0692] "Transferring to the user" means sending the generated summary information to the user's device.

[0693] "Analyzing emotional state" means analyzing the emotions of the person on the other end of a call from their voice or text and determining their emotional state.

[0694] "Emotional analysis results" refer to information obtained as a result of analyzing an emotional state.

[0695] "Summary information" refers to a short explanatory text that includes the main points of the call and the results of sentiment analysis.

[0696] This invention relates to a system that efficiently manages incoming calls on a landline, eliminates unnecessary sales calls, and notifies the user only of important calls. Furthermore, it provides a system that analyzes the emotional state of the caller and responds accordingly to the user's emotional state.

[0697] System Configuration

[0698] Hardware and software

[0699] The server plays a central role in the system and uses the following main components:

[0700] 1. Landline telephone receiving device: Receives calls from landlines.

[0701] 2. Speech recognition engine: Google Cloud Speech-to-Text, etc.

[0702] 3. Natural Language Processing Engine: TensorFlow-based algorithm.

[0703] 4. Emotion recognition engine: IBM Watson, etc.

[0704] 5. Automated response function: Synthetic speech engines such as Asterisk and Amazon Polly.

[0705] 6. Database: MongoDB or similar database to store call content and sentiment analysis results.

[0706] The terminal is a device for receiving notifications from the server and includes the following components:

[0707] 1. Mobile phone or computer: A device for the user to view the summary.

[0708] 2. Notification system: APIs such as Firebase Cloud Messaging.

[0709] Program processing

[0710] When the server receives a call from a landline, it launches automated answering software such as Asterisk and initiates the call. The call content is converted into text data in real time using Google Cloud Speech-to-Text. The converted text data is then passed to a TensorFlow-based natural language processing engine for analysis.

[0711] Natural language processing is used to determine whether a call is a sales call, and at the same time, an emotion recognition engine such as IBM Watson is used to analyze the emotional state of the caller. Based on these results, the server determines whether the call is important or unnecessary.

[0712] If the call is a sales call, Amazon Polly's synthesized speech engine is used to generate an appropriate automated response message and notify the caller. For important calls, a summary of the call content is generated and notified to the user, including sentiment information.

[0713] The device receives summaries from the server using Firebase Cloud Messaging or similar methods, allowing the user to review them. For example, a notification might appear on the mobile phone screen stating, "A relative called to check on your health."

[0714] Users review summaries notified via their devices and take follow-up actions as needed. This system frees users from unnecessary calls and allows them to efficiently manage only important calls.

[0715] Specific example

[0716] 1. Receiving a call: When a relative calls the landline, the server receives the call and initiates the conversation.

[0717] 2. Voice Recognition: When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0718] 3. Natural Language Processing: The server analyzes the converted text data and understands that it is a call from a relative checking on the user's well-being.

[0719] 4. Emotion Recognition: The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (e.g., worry or relief).

[0720] 5. Determination of whether action is needed: The server appropriately determines that this call is important.

[0721] 6. Automated response: The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[0722] 7. Summary of Call Content: After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative called to check on your recent health. The person who called was worried."

[0723] 8. Summary Transfer: The server notifies the user's mobile phone of the generated summary.

[0724] Examples of prompts for generative AI models

[0725] Example prompt: "Describe a system that converts audio data into text data in real time, analyzes it, and then summarizes and notifies the user of the important content of the call."

[0726] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0727] Step 1:

[0728] Answering incoming calls

[0729] Input: A ringing signal received on a landline phone.

[0730] Specific operation: The server receives incoming calls to a landline phone.

[0731] Data processing / data calculation: Launch automated answering software such as Asterisk and initiate a call.

[0732] Output: An automated response message is played to the caller.

[0733] Step 2:

[0734] Speech recognition

[0735] Input: Voice data collected in real time during a call.

[0736] Specific operation: The server records the call content in real time.

[0737] Data processing / data calculation: Convert audio data to text data using Google Cloud Speech-to-Text.

[0738] Output: Call content data in string format.

[0739] Step 3:

[0740] Natural Language Processing (NLP)

[0741] Input: Speech-recognized text data.

[0742] Specific operation: The server passes text data to a TensorFlow-based NLP algorithm.

[0743] Data processing / data calculation: Analyze the content of the call and apply an algorithm to determine whether it is a sales call.

[0744] Output: A flag indicating whether the call is a sales call (True / False).

[0745] Step 4:

[0746] emotion recognition

[0747] Input: Speech-recognized text data and audio data.

[0748] Specific operation: The server passes text data and audio data to IBM Watson's emotion recognition engine.

[0749] Data processing / data calculation: Analyze the emotional state of the person on the other end of the phone call.

[0750] Output: The emotional state of the person on the other end of the call (e.g., worried, relieved).

[0751] Step 5:

[0752] Determining whether action is necessary

[0753] Input: Results of natural language processing and emotion recognition.

[0754] Specific operation: The server determines the importance of the call based on the analysis results.

[0755] Data processing / data calculation: Apply logic to determine whether a call is important or unnecessary.

[0756] Output: A flag indicating the importance of the call (important / unnecessary).

[0757] Step 6:

[0758] Automated response

[0759] Input: If the call is deemed unnecessary.

[0760] Specific operation: The server uses Amazon Polly to generate an automated response message.

[0761] Data processing / data calculation: Outputs the generated response message using synthesized speech.

[0762] Output: Automated response to the caller.

[0763] Step 7:

[0764] Summary of the call

[0765] Input: Text data if the call is deemed important.

[0766] Specific operation: The server uses an NLP library such as SpaCy to extract the main points of the call.

[0767] Data processing / calculation: Summarize call content, incorporating emotional information into the final product.

[0768] Output: Summary data including call content and sentiment information.

[0769] Step 8:

[0770] Summary transfer

[0771] Input: Summary data.

[0772] Specific operation: The server uses the Firebase Cloud Messaging API to send a summary to the user's device.

[0773] Data processing / data calculation: Convert summarized data into an appropriate format and send it.

[0774] Output: A summary message displayed on the user's terminal.

[0775] Examples of prompts for generative AI models

[0776] Example prompt: "Describe a system that converts audio data into text data in real time, analyzes it, and then summarizes and notifies the user of the important content of the call."

[0777] (Application Example 2)

[0778] 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."

[0779] This invention relates to a system that efficiently manages incoming calls on landlines, eliminates unnecessary sales calls, and notifies the user only of important calls. It also aims to enable food delivery drivers to streamline customer interactions and quickly convey important information.

[0780] 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.

[0781] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for converting the content of a customer call from voice data to text data, means for analyzing the converted text data using sentiment recognition and generating a necessary response, and means for summarizing important call content and forwarding it to the user. This enables the management of incoming calls to landlines and improves the work efficiency of food delivery drivers.

[0782] "Content of received calls" refers to the audio information of calls received through the communication system.

[0783] "Means of speech recognition" refers to technologies for converting speech data into text data.

[0784] "Natural language processing" is a set of techniques for analyzing text data and understanding its content and meaning.

[0785] "Methods for determining whether action is necessary based on analyzed call content" refers to a method of determining the importance of the call content based on speech-recognized text.

[0786] "Means for generating and responding with automated response messages" refers to technology that automatically creates an appropriate response message based on analysis results and transmits that message to the caller.

[0787] "A means of summarizing important call content and forwarding it to the user" refers to a method of concisely summarizing the content of a call and notifying the user of that summary.

[0788] "Customer calls" refer to voice conversations between food delivery drivers and service users.

[0789] "Methods for converting audio data to text data" refers to technologies for converting the content of a phone call into text information.

[0790] "Emotion recognition" is a technology that determines a speaker's emotions from audio or text data.

[0791] "Means for generating necessary responses" refers to a method that automatically creates appropriate responses based on the results of emotion recognition and text analysis.

[0792] This invention is a system that efficiently manages incoming calls on landlines, eliminates unnecessary sales calls, and notifies the user only of important calls. It also enables food delivery drivers to streamline customer communication and quickly convey important information.

[0793] Server operation

[0794] 1. Answering incoming phone calls

[0795] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers. The hardware used includes server equipment and a communication interface. The software used is an automated voice response system.

[0796] 2. Speech Recognition

[0797] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This utilizes the speech_recognition library.

[0798] 3. Natural Language Processing (NLP)

[0799] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. The transformers library is used for this purpose.

[0800] 4. Emotion recognition

[0801] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data.

[0802] 5. Determining whether action is necessary.

[0803] Based on the results of natural language processing and emotion recognition, the server determines whether the call is important or unwanted, such as a sales call.

[0804] 6. Automated response

[0805] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. Using synthesized speech, it will respond with phrases such as, "I'm sorry, but I can't assist you right now." It is also possible to generate different messages depending on the user's emotional state.

[0806] 7. Summary of the call

[0807] If a call is deemed important, preparations are made to summarize its content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user.

[0808] 8. Transfer of Summary

[0809] The server notifies the user's device of the generated summary. Specifically, it sends a summary message to the user's mobile phone or email address.

[0810] Terminal operation

[0811] 1. Receiving

[0812] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[0813] User actions

[0814] 1. Confirmation of notification

[0815] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0816] Specific example

[0817] 1. Power reception

[0818] When a relative calls the landline, the server receives the call and initiates the conversation.

[0819] 2. Speech Recognition

[0820] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0821] 3. Natural Language Processing

[0822] The server analyzes the converted text data and understands that it is a phone call from a relative checking on the person's well-being.

[0823] 4. Emotion recognition

[0824] The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (for example, worry or relief).

[0825] 5. Determining whether action is necessary.

[0826] The server will appropriately determine that this call is important.

[0827] 6. Automated response

[0828] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message." It then generates a more appropriate response message depending on the relative's emotional state.

[0829] 7. Summary of the call

[0830] After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative contacted you. They called to check on your recent health. The person who contacted you was worried."

[0831] 8. Transfer of Summary

[0832] The server notifies the user's mobile phone of the generated summary.

[0833] Examples of prompt statements used

[0834] "Please upload the audio recording of your conversation with the customer."

[0835] Through this invention, it is possible to improve the efficiency of call management for landline telephones and the work of delivery personnel for food delivery services.

[0836] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0837] Step 1:

[0838] The server receives a call from a landline and initiates the call. This opens the call and prepares it for the subsequent steps. The input is the call signal from the telephone line, and the output is the call connection status.

[0839] Step 2:

[0840] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. The speech_recognition library is used for this process. The input is the audio data of the call, and the output is text data.

[0841] Step 3:

[0842] The server passes the speech-recognized text data to a natural language processing algorithm to analyze the call content. The transformers library is used here. The input is the text data obtained from speech recognition, and the output is the analysis result (semantic information of the call content).

[0843] Step 4:

[0844] The server uses an emotion engine to analyze the emotions of the person on the call. Specifically, it analyzes the speaker's emotional state from voice and text data. The input is text data and voice parameters, and the output is identified emotion information.

[0845] Step 5:

[0846] The server determines whether or not to respond to a call based on the results of natural language processing and emotion recognition. The input is the analysis results and emotion information, and the output is a response-required flag.

[0847] Step 6:

[0848] If the server determines from the call content that it is a sales call, it automatically generates an appropriate response message. This uses pre-synthesized voice. The input is a flag indicating whether a response is needed and text data, and the output is the response message.

[0849] Step 7:

[0850] If a call is deemed important, preparations are made to summarize the call content. Speech recognition and NLP tools are used to extract key points and generate a summary. Input is parsed text data and sentiment information, and output is the summarized text.

[0851] Step 8:

[0852] The server notifies the user's device of the generated summary. Specifically, it sends a summary message to the user's mobile phone or email. The input is the summarized text, and the output is a message that can be viewed on the user's receiving device.

[0853] Specific example

[0854] For example, the process flow when a relative calls and says, "Hi, how have you been feeling lately?" is as follows:

[0855] Step 1: The server receives the call and initiates the conversation.

[0856] Step 2: Record the conversation "Hello, how have you been feeling lately?" in real time and convert the audio data into text data.

[0857] Step 3: The converted text data is analyzed using a natural language processing algorithm to understand that it is a phone call from a relative checking on the user's well-being.

[0858] Step 4: The server analyzes the relative's tone of voice and speech patterns to determine if they are worried.

[0859] Step 5: Determine that this call is important and generate an appropriate response message that includes the relative's emotional state.

[0860] Step 6: Generate the answering message "We are currently unable to answer your call, but we will take your message" and send it to the caller.

[0861] Step 7: After the call ends, summarize the content and generate a summary such as, "A relative wanted to check on your recent health. The person who contacted them was worried."

[0862] Step 8: Notify the user of the summary on their mobile phone so they can efficiently grasp important call information.

[0863] In this way, the system of the present invention enables call management for landline telephones and improves the work efficiency of food delivery personnel.

[0864] 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.

[0865] 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.

[0866] 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.

[0867] [Third Embodiment]

[0868] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0869] 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.

[0870] 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).

[0871] 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.

[0872] 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.

[0873] 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).

[0874] 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.

[0875] 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.

[0876] 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.

[0877] 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.

[0878] 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.

[0879] 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".

[0880] This invention relates to a system for efficiently managing incoming calls on a landline telephone, eliminating unwanted sales calls, and notifying the user only of important calls. Specific embodiments for implementing this invention are described below.

[0881] This system is primarily composed of three components: servers, terminals, and users.

[0882] server

[0883] 1. Answering incoming phone calls

[0884] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers.

[0885] 2. Speech Recognition

[0886] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This process makes the call content analyzable as text.

[0887] 3. Natural Language Processing (NLP)

[0888] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. This analysis determines whether the call is a sales call or not.

[0889] 4. Determining whether action is necessary.

[0890] Based on the results of natural language processing, the server determines whether the call is an unwanted sales call or an important call. If it is determined to be a sales call, it proceeds to the next step.

[0891] 5. Automated response

[0892] If the server determines that a call is a sales call, it automatically generates an appropriate response message and uses synthesized speech to answer. For example, it might respond with, "I'm sorry, but I can't help you right now."

[0893] 6. Summary of the call

[0894] If the server determines a call is important, it prepares to summarize the call and notify the user. The server extracts the main points of the call and generates a concise summary.

[0895] 7. Transfer of Summary

[0896] The server notifies the user's device of the generated summary. For example, it may send the summary to the user's mobile phone or email address.

[0897] terminal

[0898] 1. Receiving

[0899] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[0900] User

[0901] 1. Confirmation of notification

[0902] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0903] Specific example

[0904] 1. Power reception

[0905] When a relative calls the landline, the server receives the call and initiates the conversation.

[0906] 2. Speech Recognition

[0907] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[0908] 3. Natural Language Processing

[0909] The server analyzes the converted text and understands that it is a call from a relative checking on the person's well-being.

[0910] 4. Determining whether action is necessary.

[0911] The server will appropriately determine that this call is important.

[0912] 5. Automated response

[0913] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[0914] 6. Summary of the call

[0915] After the call ends, the server summarizes the content and generates a summary such as, "You received a call from a relative. They called to check on your recent health."

[0916] 7. Transfer of Summary

[0917] The server notifies the user's mobile phone of the generated summary.

[0918] This processing flow allows users to efficiently receive important call information and avoid being bothered by unnecessary sales calls.

[0919] The following describes the processing flow.

[0920] Step 1: Answering incoming calls

[0921] The server receives incoming calls to the landline phone. The server detects that a call has been made through the landline phone's interface, activates the automated answering module, and initiates the call.

[0922] Step 2: Speech Recognition

[0923] The server records the call in real time. The server uses a speech recognition engine to convert the recorded audio data into text. This makes the call content analyzable as text data.

[0924] Step 3: Natural Language Processing

[0925] The server passes the converted text data to a natural language processing (NLP) algorithm. The server uses the NLP module to analyze the text content and extract important information.

[0926] Step 4: Determine whether action is necessary.

[0927] The server uses natural language processing results to determine whether the call is important or an unwanted call, such as a sales call. If the call is important, it proceeds to the next step.

[0928] Step 5: Automated response

[0929] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. Using synthesized speech, it will respond with phrases such as, "I'm sorry, but I can't assist you at this time."

[0930] Step 6: Summarize the call

[0931] If the server determines a call is important, it prepares to summarize the call. After the call ends, the server uses NLP to extract the key points and generates a summary to notify the user.

[0932] Step 7: Transfer the summary

[0933] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative has contacted you. They called to check on your recent health."

[0934] Step 8: Final check

[0935] Users can view summaries through their devices. This allows users to efficiently identify only important calls and be freed from unnecessary sales calls.

[0936] (Example 1)

[0937] 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."

[0938] Traditional landline telephone systems lack mechanisms to efficiently filter out unwanted sales calls and notify users only of important calls. As a result, users are often bothered by numerous unwanted calls and may miss important ones. A system is needed to solve this problem and optimize the user's communication experience.

[0939] 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.

[0940] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary from the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, means for using a speech recognition engine when converting the call content to text, means for using a generative AI model to generate the summary, and means for using a communication device to forward the summary notification. As a result, the user can quickly and efficiently obtain only important call information and communicate without being bothered by unnecessary sales calls.

[0941] "Receiving a call" refers to receiving a call on a landline telephone.

[0942] "Speech recognition" is a technology that converts speech data into text data.

[0943] "Natural language processing" is a technology that analyzes text data to understand its meaning.

[0944] "Determining whether action is necessary" refers to determining whether a call is important based on the analyzed content.

[0945] An "automatic response message" is a message that the system automatically generates and plays for the person on the other end of the call.

[0946] A "summary" is a document that extracts the main points of a phone call and presents them concisely.

[0947] A "speech recognition engine" refers to software or algorithms used to convert speech data into text data.

[0948] A "generative AI model" is an artificial intelligence model that performs summarization and other processing on given text data.

[0949] A "communication device" refers to hardware or software used to transfer generated information to a specific device or system.

[0950] Modes for carrying out the invention

[0951] This invention relates to a system that efficiently manages incoming calls to a landline telephone, eliminates unnecessary sales calls, and notifies the user only of important calls. The following describes a specific embodiment for implementing this invention. This system mainly consists of three components: a server, a terminal, and a user.

[0952] server

[0953] The server receives incoming calls to a landline phone and initiates the call. Upon initiating the call, the server plays a pre-configured automated response message. This can be done using, for example, Python libraries or communication APIs.

[0954] The call content is recorded in real time, and the recorded audio data is converted to text using a speech recognition engine such as Google Cloud Speech-to-Text. In this process, the audio can be recorded using the Pydub library, uploaded to Google Cloud Storage, and then converted to text by calling an API.

[0955] The transformed text data is passed to a natural language processing (NLP) algorithm, such as the Hugging Face BERT model, to analyze the text. This then uses the Python Transformers library to determine whether the call is a sales call.

[0956] Based on the analysis results, the system determines whether the call is an unnecessary sales call or an important call. If it is determined to be a sales call, an automated response message can be generated and played back using a text-to-speech engine (e.g., Google Cloud Text-to-Speech). The text-to-speech engine uses an API to generate messages such as, "I'm sorry, but I can't help you right now."

[0957] If a call is deemed important, the server extracts the main points of the call and generates a summary using a method such as OpenAI's GPT-3. The summary is compiled into a concise document and ready to be sent to the user. The generated summary is transferred to the user's device via SMS or email.

[0958] terminal

[0959] The terminal receives a summary sent from the server. Typically, this is received on the user's mobile phone or computer and can be viewed through a notification interface. On mobile phones, it's received as an SMS message, and on computers, it's displayed in an email client.

[0960] User

[0961] Users review summaries sent from the server via their mobile phones or computers. This allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[0962] Specific examples of operation

[0963] The following shows a specific example of how this system works.

[0964] 1. When a relative calls the landline, the server receives the call and initiates the conversation.

[0965] 2. When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text using Google Cloud Speech-to-Text.

[0966] 3. The server analyzes the converted text using a BERT model and recognizes that it is a call from a relative checking on the recipient's well-being.

[0967] 4. The server determines that the call is important and responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[0968] 5. After the call ends, the server summarizes the call and generates a summary such as, "You received a call from a relative. They called to check on your recent health."

[0969] 6. The server sends the generated summary to the user's mobile phone via SMS.

[0970] Example of a prompt

[0971] "The following text is a transcript of a phone call received on a landline. Please summarize the main points of this call. Call content: 'How are you? How has your health been lately?'"

[0972] By inputting this prompt into the AI ​​generation model, important information can be extracted from the call content and notified to the user in an easy-to-understand format.

[0973] As described above, the present invention provides the effect of preventing users from missing important calls and freeing them from unwanted sales calls.

[0974] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0975] Step 1:

[0976] Answering incoming calls

[0977] The server receives incoming calls to a landline phone. The input is the landline phone's ringing signal, which the server detects and initiates the call. Upon initiating the call, the server plays a pre-configured automated response message. Specifically, it uses various Python libraries and communication APIs to detect the ringing signal and play the response message. The output is a call session in a waiting state.

[0978] Step 2:

[0979] Speech recognition

[0980] The server records incoming calls in real time and converts the audio data to text using a speech recognition engine such as Google Cloud Speech-to-Text. The input is the audio from the call, which is recorded using libraries such as Pydub. Once the recording is complete, the audio data is uploaded to Google Cloud Storage, and the Google Cloud Speech-to-Text API is called to convert the audio to text. The output is text data.

[0981] Step 3:

[0982] natural language analysis

[0983] The server passes the converted text data to a natural language processing algorithm for analysis. The input is the text data generated in step 2, which is analyzed using the Hugging Face BERT model. Specifically, the Python Transformers library is used to analyze the text using prompt phrases such as "Is this a sales call?". The output is the result of the analysis.

[0984] Step 4:

[0985] Determining whether action is necessary

[0986] The server determines whether a call is an important call or an unnecessary sales call based on the results of natural language processing. The input is the analysis result from step 3, and an algorithm is applied to determine whether it is an unnecessary sales call. The output is the result of the determination of whether the call is an important call or a sales call.

[0987] Step 5:

[0988] Automated response

[0989] If the server determines it is a sales call, it generates an automated response message and plays it back using a speech synthesis engine (e.g., Amazon Polly). The input is the result of the determination in step 4, which is specifically applied if it is a sales call. Specifically, it generates a message such as "I'm sorry, but I can't help you right now," synthesizes it into speech using the Amazon Polly API, and plays it back using the Twilio API. The output is the voice response to the sales call.

[0990] Step 6:

[0991] Summary of the call

[0992] The server prepares to summarize the call content and notify the user. The input is the result of the judgment in step 4, using the text that was determined to be from an important call as a prompt. Specifically, a generative AI model such as OpenAI's GPT-3 is used to input a prompt such as, "The following text is the content of a call received on a landline. Please summarize the main points of this call," and generates a summary. The output is the summarized text.

[0993] Step 7:

[0994] Summary transfer

[0995] The server notifies the user's device of the generated summary. The input is the summary text generated in step 6, which is formatted appropriately and sent to the user's mobile phone or email address. Specifically, this is done by sending an SMS using the Twilio API or by sending an email via an SMTP server. The output is the summary notification sent to the user's device.

[0996] (Application Example 1)

[0997] 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."

[0998] Traditional landline phones are often plagued by unwanted calls such as sales calls and scam calls, creating a risk of missing important calls. Furthermore, it's difficult to understand the content of important calls when the user is away or during busy periods. Therefore, there is a need for a system that efficiently manages incoming landline calls, automatically filters out unwanted sales and scam calls, and notifies the user only of important calls.

[0999] 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.

[1000] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, and means for notifying the smartphone of the call content summary. This enables efficient management of calls received on a landline, allowing the user to quickly grasp important call content without being bothered by unwanted sales calls or scam calls.

[1001] "Speech recognition" is a technology that analyzes speech data and converts it into text data.

[1002] "Natural language processing" is a technology that analyzes text obtained through speech recognition and understands and classifies its content.

[1003] An "automated response message" is a message that is automatically generated based on specific conditions and responds to the caller in voice.

[1004] A "summary" is information that extracts the main points of an important phone call and presents them concisely.

[1005] "Means of forwarding to the user" refers to the function of sending summarized call content to the user's device.

[1006] A "smartphone" is a portable information terminal that receives summaries of call content and notifies the user.

[1007] A "generative AI model" is an artificial intelligence algorithm that generates new information based on training data.

[1008] A "prompt statement" is an instruction given to a generative AI model and is used to obtain a specific output.

[1009] Modes for carrying out the invention

[1010] server

[1011] The server has key functions for efficiently managing incoming calls to landlines. This is done using the following hardware and software:

[1012] Speech Recognition: The server uses a speech recognition library (e.g., the speech_recognition library) to record incoming calls on the landline in real time and convert the content to text.

[1013] Natural Language Processing: The server analyzes the transcribed call content using a natural language processing library (e.g., the transformers library). This analysis determines whether the call is a sales call or a scam call.

[1014] Generative AI Model: The server uses a generative AI model to create automated response messages and summaries of important calls based on the analysis results. The generative AI model uses a pre-trained algorithm.

[1015] Summary Notification: If the server determines a call is important, it generates a summary and notifies the user's smartphone. The summarized call content is concisely presented to reduce the user's burden.

[1016] For example, if a relative calls to check on your well-being, the server will use speech recognition to process the message, such as "Hello, how have you been feeling lately?", and then analyze it as text. If the analysis determines that the call is important, it will generate a summary and notify your smartphone, "A relative has contacted you. They called to check on your recent health."

[1017] terminal

[1018] The terminal receives summaries sent from the server and notifies the user. This typically includes the user's smartphone or computer.

[1019] Notification Reception: The device receives a summarized call content and displays it to the user. Notifications are often displayed on the smartphone as push notifications.

[1020] User

[1021] Users can review a summary of the call content through their device and take necessary actions.

[1022] Notification Confirmation: Users can check notifications on their devices to understand the content of important calls. This allows them to respond only to important calls without being bothered by unwanted sales calls or scam calls.

[1023] As supplementary information, here are some examples of prompt statements that are input to the generative AI model:

[1024] "Please convert the following audio call into text and use natural language processing to determine if it is a sales call. If it is determined to be a sales call, please enter an automated response message."

[1025] This embodiment of the invention allows users to efficiently manage incoming calls on their landline and reliably identify only important calls.

[1026] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1027] Step 1:

[1028] The server receives incoming calls to a landline phone. The input is audio data from the landline. The server records this audio data in real time. The output is a recorded audio file. Specifically, the server uses an audio recording device to save the call.

[1029] Step 2:

[1030] The server uses a speech recognition library (e.g., speech_recognition) to convert recorded audio files into text. The input is a recorded audio file. The server uses a speech recognition engine to convert the audio data into text data. The output is the converted text. Specifically, the server initializes the speech recognition engine and analyzes and converts the audio data.

[1031] Step 3:

[1032] The server uses a natural language processing library (e.g., transformers) to analyze the transformed text. The input is the text obtained by speech recognition. The server uses a natural language processing engine to analyze the call content to determine whether it is a sales call or a scam call. The output is the analysis result. Specifically, the server uses a natural language processing model to classify the text data.

[1033] Step 4:

[1034] The server uses a generative AI model to generate automated response messages based on the analysis results. The input is the analysis results of natural language processing. The server inputs prompt sentences into the generative AI model to create automated response messages. The output is the generated automated response message. Specifically, the server provides prompt sentences to the generative AI model and generates an appropriate response.

[1035] Step 5:

[1036] The server summarizes the call content and prepares it for transfer to the smartphone. The input consists of analysis results and the text data of the call. The server uses a generative AI model to summarize the key points of the call and condense them into a concise format. The output is the summarized call content. Specifically, the server uses the generative AI model to extract the main points of the call and create a summary.

[1037] Step 6:

[1038] The server notifies the user's smartphone with a summary. The summarized call content is provided as input. The server uses push notifications to send a notification to the smartphone. The output is a summary notification displayed on the smartphone. Specifically, the server calls a notification API and sends the generated summary to the smartphone.

[1039] Step 7:

[1040] The user checks the notification on their smartphone and takes action as needed. The input is a summary of the notification. The user reviews the received summary to understand the important call content. The output is the user's action (e.g., returning the call). Specifically, the user opens the smartphone notification, reviews the call content, and decides on a course of action.

[1041] 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.

[1042] This invention relates to a system that efficiently manages incoming calls to a landline, eliminates unnecessary sales calls, and notifies the user only of important calls. Furthermore, by combining this with an emotion engine, the invention provides a system that can recognize the user's emotional state and respond accordingly.

[1043] server

[1044] 1. Answering incoming phone calls

[1045] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers.

[1046] 2. Speech Recognition

[1047] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This process makes the call content analyzable as text.

[1048] 3. Natural Language Processing (NLP)

[1049] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. This analysis determines whether the call is a sales call or not.

[1050] 4. Emotion recognition

[1051] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data and stores that information.

[1052] 5. Determining whether action is necessary.

[1053] The server determines whether the call is important or unwanted, such as a sales call, based on the results of natural language processing and sentiment recognition. If the call is important, it proceeds to the next step.

[1054] 6. Automated response

[1055] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. This message will use synthesized speech to say something like, "I'm sorry, but I can't assist you right now." It can also generate different messages depending on the user's emotional state.

[1056] 7. Summary of the call

[1057] If the server determines a call is important, it prepares to summarize the call content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user. The results of sentiment recognition can also be included in the summary.

[1058] 8. Transfer of Summary

[1059] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative has contacted you. They called to check on your recent health."

[1060] terminal

[1061] 1. Receiving

[1062] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[1063] User

[1064] 1. Confirmation of notification

[1065] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[1066] Specific example

[1067] 1. Power reception

[1068] When a relative calls the landline, the server receives the call and initiates the conversation.

[1069] 2. Speech Recognition

[1070] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[1071] 3. Natural Language Processing

[1072] The server analyzes the converted text data and understands that it is a phone call from a relative checking on the person's well-being.

[1073] 4. Emotion recognition

[1074] The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (for example, worry or relief).

[1075] 5. Determining whether action is necessary.

[1076] The server will appropriately determine that this call is important.

[1077] 6. Automated response

[1078] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message." It then generates a more appropriate response message depending on the relative's emotional state.

[1079] 7. Summary of the call

[1080] After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative contacted you. They called to check on your recent health. The person who contacted you was worried."

[1081] 8. Transfer of Summary

[1082] The server notifies the user's mobile phone of the generated summary.

[1083] This processing flow allows users to efficiently receive important call information and the caller's emotional state, freeing them from unnecessary sales calls.

[1084] The following describes the processing flow.

[1085] Step 1: Answering incoming calls

[1086] The server receives incoming calls to the landline phone. The server detects that a call has been made through the landline phone's interface, activates the automated answering module, and initiates the call.

[1087] Step 2: Speech Recognition

[1088] The server records the call in real time. The server uses a speech recognition engine to convert the recorded audio data into text. This makes the call content analyzable as text data.

[1089] Step 3: Natural Language Processing (NLP)

[1090] The server passes the converted text data to a natural language processing (NLP) algorithm. The server uses the NLP module to analyze the text content and extract important information and keywords.

[1091] Step 4: Emotion Recognition

[1092] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data and stores that information. For example, it identifies emotions such as "worried" or "angry."

[1093] Step 5: Determine whether action is necessary.

[1094] The server determines whether the call is important or unwanted, such as a sales call, based on the results of natural language processing and sentiment recognition. If the call is important, it proceeds to the next step.

[1095] Step 6: Automated response

[1096] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. This message will use synthesized speech to say something like, "I'm sorry, but I can't assist you right now." It can also generate different messages depending on the user's emotional state.

[1097] Step 7: Summarize the call

[1098] If the server determines a call is important, it prepares to summarize the call content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user. The results of sentiment recognition can also be included in the summary.

[1099] Step 8: Transferring the summary

[1100] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative contacted you. They called to check on your recent health. They were worried about you."

[1101] Step 9: Final Check

[1102] The user reviews the summary via their device. This summary includes the main points of the call and the caller's emotional state. This allows the user to efficiently grasp the important points of the call and the caller's emotional state.

[1103] (Example 2)

[1104] 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."

[1105] The challenge lies in efficiently distinguishing between unwanted sales calls and important calls on landlines, thereby saving users time and effort. Furthermore, it is necessary to analyze the emotional state of the caller to provide more appropriate responses.

[1106] 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.

[1107] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, means for analyzing the emotional state during the call, and means for generating summary information including the emotion analysis results. This makes it possible to eliminate unnecessary sales calls and efficiently grasp important call content. Furthermore, it is possible to respond based on the emotions of the caller, thereby improving user satisfaction.

[1108] "Received call" refers to the audio signal that the server receives when a phone call comes in.

[1109] "Speech recognition" is a technology that converts received audio signals into text data.

[1110] "Natural language processing" is a technology that analyzes text data obtained through speech recognition to understand its content.

[1111] "Determining whether action is necessary" means deciding whether the analyzed call content is important or unnecessary.

[1112] An "automatic response message" is a message that a server automatically generates and sends to the caller in response to an unnecessary call.

[1113] A "summary" is information that extracts the main points of an important phone call and presents them concisely.

[1114] "Transferring to the user" means sending the generated summary information to the user's device.

[1115] "Analyzing emotional state" means analyzing the emotions of the person on the other end of a call from their voice or text and determining their emotional state.

[1116] "Emotional analysis results" refer to information obtained as a result of analyzing an emotional state.

[1117] "Summary information" refers to a short explanatory text that includes the main points of the call and the results of sentiment analysis.

[1118] This invention relates to a system that efficiently manages incoming calls on a landline, eliminates unnecessary sales calls, and notifies the user only of important calls. Furthermore, it provides a system that analyzes the emotional state of the caller and responds accordingly to the user's emotional state.

[1119] System Configuration

[1120] Hardware and software

[1121] The server plays a central role in the system and uses the following main components:

[1122] 1. Landline telephone receiving device: Receives calls from landlines.

[1123] 2. Speech recognition engine: Google Cloud Speech-to-Text, etc.

[1124] 3. Natural Language Processing Engine: TensorFlow-based algorithm.

[1125] 4. Emotion recognition engine: IBM Watson, etc.

[1126] 5. Automated response function: Synthetic speech engines such as Asterisk and Amazon Polly.

[1127] 6. Database: MongoDB or similar database to store call content and sentiment analysis results.

[1128] The terminal is a device for receiving notifications from the server and includes the following components:

[1129] 1. Mobile phone or computer: A device for the user to view the summary.

[1130] 2. Notification system: APIs such as Firebase Cloud Messaging.

[1131] Program processing

[1132] When the server receives a call from a landline, it launches automated answering software such as Asterisk and initiates the call. The call content is converted into text data in real time using Google Cloud Speech-to-Text. The converted text data is then passed to a TensorFlow-based natural language processing engine for analysis.

[1133] Natural language processing is used to determine whether a call is a sales call, and at the same time, an emotion recognition engine such as IBM Watson is used to analyze the emotional state of the caller. Based on these results, the server determines whether the call is important or unnecessary.

[1134] If the call is a sales call, Amazon Polly's synthesized speech engine is used to generate an appropriate automated response message and notify the caller. For important calls, a summary of the call content is generated and notified to the user, including sentiment information.

[1135] The device receives summaries from the server using Firebase Cloud Messaging or similar methods, allowing the user to review them. For example, a notification might appear on the mobile phone screen stating, "A relative called to check on your health."

[1136] Users review summaries notified via their devices and take follow-up actions as needed. This system frees users from unnecessary calls and allows them to efficiently manage only important calls.

[1137] Specific example

[1138] 1. Receiving a call: When a relative calls the landline, the server receives the call and initiates the conversation.

[1139] 2. Voice Recognition: When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[1140] 3. Natural Language Processing: The server analyzes the converted text data and understands that it is a call from a relative checking on the user's well-being.

[1141] 4. Emotion Recognition: The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (e.g., worry or relief).

[1142] 5. Determination of whether action is needed: The server appropriately determines that this call is important.

[1143] 6. Automated response: The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[1144] 7. Summary of Call Content: After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative called to check on your recent health. The person who called was worried."

[1145] 8. Summary Transfer: The server notifies the user's mobile phone of the generated summary.

[1146] Examples of prompts for generative AI models

[1147] Example prompt: "Describe a system that converts audio data into text data in real time, analyzes it, and then summarizes and notifies the user of the important content of the call."

[1148] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1149] Step 1:

[1150] Answering incoming calls

[1151] Input: A ringing signal received on a landline phone.

[1152] Specific operation: The server receives incoming calls to a landline phone.

[1153] Data processing / data calculation: Launch automated answering software such as Asterisk and initiate a call.

[1154] Output: An automated response message is played to the caller.

[1155] Step 2:

[1156] Speech recognition

[1157] Input: Voice data collected in real time during a call.

[1158] Specific operation: The server records the call content in real time.

[1159] Data processing / data calculation: Convert audio data to text data using Google Cloud Speech-to-Text.

[1160] Output: Call content data in string format.

[1161] Step 3:

[1162] Natural Language Processing (NLP)

[1163] Input: Speech-recognized text data.

[1164] Specific operation: The server passes text data to a TensorFlow-based NLP algorithm.

[1165] Data processing / data calculation: Analyze the content of the call and apply an algorithm to determine whether it is a sales call.

[1166] Output: A flag indicating whether the call is a sales call (True / False).

[1167] Step 4:

[1168] emotion recognition

[1169] Input: Speech-recognized text data and audio data.

[1170] Specific operation: The server passes text data and audio data to IBM Watson's emotion recognition engine.

[1171] Data processing / data calculation: Analyze the emotional state of the person on the other end of the phone call.

[1172] Output: The emotional state of the person on the other end of the call (e.g., worried, relieved).

[1173] Step 5:

[1174] Determining whether action is necessary

[1175] Input: Results of natural language processing and emotion recognition.

[1176] Specific operation: The server determines the importance of the call based on the analysis results.

[1177] Data processing / data calculation: Apply logic to determine whether a call is important or unnecessary.

[1178] Output: A flag indicating the importance of the call (important / unnecessary).

[1179] Step 6:

[1180] Automated response

[1181] Input: If the call is deemed unnecessary.

[1182] Specific operation: The server uses Amazon Polly to generate an automated response message.

[1183] Data processing / data calculation: Outputs the generated response message using synthesized speech.

[1184] Output: Automated response to the caller.

[1185] Step 7:

[1186] Summary of the call

[1187] Input: Text data if the call is deemed important.

[1188] Specific operation: The server uses an NLP library such as SpaCy to extract the main points of the call.

[1189] Data processing / calculation: Summarize call content, incorporating emotional information into the final product.

[1190] Output: Summary data including call content and sentiment information.

[1191] Step 8:

[1192] Summary transfer

[1193] Input: Summary data.

[1194] Specific operation: The server uses the Firebase Cloud Messaging API to send a summary to the user's device.

[1195] Data processing / data calculation: Convert summarized data into an appropriate format and send it.

[1196] Output: A summary message displayed on the user's terminal.

[1197] Examples of prompts for generative AI models

[1198] Example prompt: "Describe a system that converts audio data into text data in real time, analyzes it, and then summarizes and notifies the user of the important content of the call."

[1199] (Application Example 2)

[1200] 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."

[1201] This invention relates to a system that efficiently manages incoming calls on landlines, eliminates unnecessary sales calls, and notifies the user only of important calls. It also aims to enable food delivery drivers to streamline customer interactions and quickly convey important information.

[1202] 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.

[1203] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for converting the content of a customer call from voice data to text data, means for analyzing the converted text data using sentiment recognition and generating a necessary response, and means for summarizing important call content and forwarding it to the user. This enables the management of incoming calls to landlines and improves the work efficiency of food delivery drivers.

[1204] "Content of received calls" refers to the audio information of calls received through the communication system.

[1205] "Means of speech recognition" refers to technologies for converting speech data into text data.

[1206] "Natural language processing" is a set of techniques for analyzing text data and understanding its content and meaning.

[1207] "Methods for determining whether action is necessary based on analyzed call content" refers to a method of determining the importance of the call content based on speech-recognized text.

[1208] "Means for generating and responding with automated response messages" refers to technology that automatically creates an appropriate response message based on analysis results and transmits that message to the caller.

[1209] "A means of summarizing important call content and forwarding it to the user" refers to a method of concisely summarizing the content of a call and notifying the user of that summary.

[1210] "Customer calls" refer to voice conversations between food delivery drivers and service users.

[1211] "Methods for converting audio data to text data" refers to technologies for converting the content of a phone call into text information.

[1212] "Emotion recognition" is a technology that determines a speaker's emotions from audio or text data.

[1213] "Means for generating necessary responses" refers to a method that automatically creates appropriate responses based on the results of emotion recognition and text analysis.

[1214] This invention is a system that efficiently manages incoming calls on landlines, eliminates unnecessary sales calls, and notifies the user only of important calls. It also enables food delivery drivers to streamline customer communication and quickly convey important information.

[1215] Server operation

[1216] 1. Answering incoming phone calls

[1217] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers. The hardware used includes server equipment and a communication interface. The software used is an automated voice response system.

[1218] 2. Speech Recognition

[1219] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This utilizes the speech_recognition library.

[1220] 3. Natural Language Processing (NLP)

[1221] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. The transformers library is used for this purpose.

[1222] 4. Emotion recognition

[1223] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data.

[1224] 5. Determining whether action is necessary.

[1225] Based on the results of natural language processing and emotion recognition, the server determines whether the call is important or unwanted, such as a sales call.

[1226] 6. Automated response

[1227] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. Using synthesized speech, it will respond with phrases such as, "I'm sorry, but I can't assist you right now." It is also possible to generate different messages depending on the user's emotional state.

[1228] 7. Summary of the call

[1229] If a call is deemed important, preparations are made to summarize its content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user.

[1230] 8. Transfer of Summary

[1231] The server notifies the user's device of the generated summary. Specifically, it sends a summary message to the user's mobile phone or email address.

[1232] Terminal operation

[1233] 1. Receiving

[1234] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[1235] User actions

[1236] 1. Confirmation of notification

[1237] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[1238] Specific example

[1239] 1. Power reception

[1240] When a relative calls the landline, the server receives the call and initiates the conversation.

[1241] 2. Speech Recognition

[1242] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[1243] 3. Natural Language Processing

[1244] The server analyzes the converted text data and understands that it is a phone call from a relative checking on the person's well-being.

[1245] 4. Emotion recognition

[1246] The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (for example, worry or relief).

[1247] 5. Determining whether action is necessary.

[1248] The server will appropriately determine that this call is important.

[1249] 6. Automated response

[1250] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message." It then generates a more appropriate response message depending on the relative's emotional state.

[1251] 7. Summary of the call

[1252] After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative contacted you. They called to check on your recent health. The person who contacted you was worried."

[1253] 8. Transfer of Summary

[1254] The server notifies the user's mobile phone of the generated summary.

[1255] Examples of prompt statements used

[1256] "Please upload the audio recording of your conversation with the customer."

[1257] Through this invention, it is possible to improve the efficiency of call management for landline telephones and the work of delivery personnel for food delivery services.

[1258] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1259] Step 1:

[1260] The server receives a call from a landline and initiates the call. This opens the call and prepares it for the subsequent steps. The input is the call signal from the telephone line, and the output is the call connection status.

[1261] Step 2:

[1262] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. The speech_recognition library is used for this process. The input is the audio data of the call, and the output is text data.

[1263] Step 3:

[1264] The server passes the speech-recognized text data to a natural language processing algorithm to analyze the call content. The transformers library is used here. The input is the text data obtained from speech recognition, and the output is the analysis result (semantic information of the call content).

[1265] Step 4:

[1266] The server uses an emotion engine to analyze the emotions of the person on the call. Specifically, it analyzes the speaker's emotional state from voice and text data. The input is text data and voice parameters, and the output is identified emotion information.

[1267] Step 5:

[1268] The server determines whether or not to respond to a call based on the results of natural language processing and emotion recognition. The input is the analysis results and emotion information, and the output is a response-required flag.

[1269] Step 6:

[1270] If the server determines from the call content that it is a sales call, it automatically generates an appropriate response message. This uses pre-synthesized voice. The input is a flag indicating whether a response is needed and text data, and the output is the response message.

[1271] Step 7:

[1272] If a call is deemed important, preparations are made to summarize the call content. Speech recognition and NLP tools are used to extract key points and generate a summary. Input is parsed text data and sentiment information, and output is the summarized text.

[1273] Step 8:

[1274] The server notifies the user's device of the generated summary. Specifically, it sends a summary message to the user's mobile phone or email. The input is the summarized text, and the output is a message that can be viewed on the user's receiving device.

[1275] Specific example

[1276] For example, the process flow when a relative calls and says, "Hi, how have you been feeling lately?" is as follows:

[1277] Step 1: The server receives the call and initiates the conversation.

[1278] Step 2: Record the conversation "Hello, how have you been feeling lately?" in real time and convert the audio data into text data.

[1279] Step 3: The converted text data is analyzed using a natural language processing algorithm to understand that it is a phone call from a relative checking on the user's well-being.

[1280] Step 4: The server analyzes the relative's tone of voice and speech patterns to determine if they are worried.

[1281] Step 5: Determine that this call is important and generate an appropriate response message that includes the relative's emotional state.

[1282] Step 6: Generate the answering message "We are currently unable to answer your call, but we will take your message" and send it to the caller.

[1283] Step 7: After the call ends, summarize the content and generate a summary such as, "A relative wanted to check on your recent health. The person who contacted them was worried."

[1284] Step 8: Notify the user of the summary on their mobile phone so they can efficiently grasp important call information.

[1285] In this way, the system of the present invention enables call management for landline telephones and improves the work efficiency of food delivery personnel.

[1286] 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.

[1287] 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.

[1288] 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.

[1289] [Fourth Embodiment]

[1290] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1291] 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.

[1292] 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).

[1293] 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.

[1294] 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.

[1295] 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).

[1296] 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.

[1297] 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.

[1298] 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.

[1299] 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.

[1300] 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.

[1301] 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.

[1302] 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".

[1303] This invention relates to a system for efficiently managing incoming calls on a landline telephone, eliminating unwanted sales calls, and notifying the user only of important calls. Specific embodiments for implementing this invention are described below.

[1304] This system is primarily composed of three components: servers, terminals, and users.

[1305] server

[1306] 1. Answering incoming phone calls

[1307] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers.

[1308] 2. Speech Recognition

[1309] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This process makes the call content analyzable as text.

[1310] 3. Natural Language Processing (NLP)

[1311] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. This analysis determines whether the call is a sales call or not.

[1312] 4. Determining whether action is necessary.

[1313] Based on the results of natural language processing, the server determines whether the call is an unwanted sales call or an important call. If it is determined to be a sales call, it proceeds to the next step.

[1314] 5. Automated response

[1315] If the server determines that a call is a sales call, it automatically generates an appropriate response message and uses synthesized speech to answer. For example, it might respond with, "I'm sorry, but I can't help you right now."

[1316] 6. Summary of the call

[1317] If the server determines a call is important, it prepares to summarize the call and notify the user. The server extracts the main points of the call and generates a concise summary.

[1318] 7. Transfer of Summary

[1319] The server notifies the user's device of the generated summary. For example, it may send the summary to the user's mobile phone or email address.

[1320] terminal

[1321] 1. Receiving

[1322] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[1323] User

[1324] 1. Confirmation of notification

[1325] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[1326] Specific example

[1327] 1. Power reception

[1328] When a relative calls the landline, the server receives the call and initiates the conversation.

[1329] 2. Speech Recognition

[1330] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[1331] 3. Natural Language Processing

[1332] The server analyzes the converted text and understands that it is a call from a relative checking on the person's well-being.

[1333] 4. Determining whether action is necessary.

[1334] The server will appropriately determine that this call is important.

[1335] 5. Automated response

[1336] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[1337] 6. Summary of the call

[1338] After the call ends, the server summarizes the content and generates a summary such as, "You received a call from a relative. They called to check on your recent health."

[1339] 7. Transfer of Summary

[1340] The server notifies the user's mobile phone of the generated summary.

[1341] This processing flow allows users to efficiently receive important call information and avoid being bothered by unnecessary sales calls.

[1342] The following describes the processing flow.

[1343] Step 1: Answering incoming calls

[1344] The server receives incoming calls to the landline phone. The server detects that a call has been made through the landline phone's interface, activates the automated answering module, and initiates the call.

[1345] Step 2: Speech Recognition

[1346] The server records the call in real time. The server uses a speech recognition engine to convert the recorded audio data into text. This makes the call content analyzable as text data.

[1347] Step 3: Natural Language Processing

[1348] The server passes the converted text data to a natural language processing (NLP) algorithm. The server uses the NLP module to analyze the text content and extract important information.

[1349] Step 4: Determine whether action is necessary.

[1350] The server uses natural language processing results to determine whether the call is important or an unwanted call, such as a sales call. If the call is important, it proceeds to the next step.

[1351] Step 5: Automated response

[1352] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. Using synthesized speech, it will respond with phrases such as, "I'm sorry, but I can't assist you at this time."

[1353] Step 6: Summarize the call

[1354] If the server determines a call is important, it prepares to summarize the call. After the call ends, the server uses NLP to extract the key points and generates a summary to notify the user.

[1355] Step 7: Transfer the summary

[1356] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative has contacted you. They called to check on your recent health."

[1357] Step 8: Final check

[1358] Users can view summaries through their devices. This allows users to efficiently identify only important calls and be freed from unnecessary sales calls.

[1359] (Example 1)

[1360] 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".

[1361] Traditional landline telephone systems lack mechanisms to efficiently filter out unwanted sales calls and notify users only of important calls. As a result, users are often bothered by numerous unwanted calls and may miss important ones. A system is needed to solve this problem and optimize the user's communication experience.

[1362] 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.

[1363] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary from the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, means for using a speech recognition engine when converting the call content to text, means for using a generative AI model to generate the summary, and means for using a communication device to forward the summary notification. As a result, the user can quickly and efficiently obtain only important call information and communicate without being bothered by unnecessary sales calls.

[1364] "Receiving a call" refers to receiving a call on a landline telephone.

[1365] "Speech recognition" is a technology that converts speech data into text data.

[1366] "Natural language processing" is a technology that analyzes text data to understand its meaning.

[1367] "Determining whether action is necessary" refers to determining whether a call is important based on the analyzed content.

[1368] An "automatic response message" is a message that the system automatically generates and plays for the person on the other end of the call.

[1369] A "summary" is a document that extracts the main points of a phone call and presents them concisely.

[1370] A "speech recognition engine" refers to software or algorithms used to convert speech data into text data.

[1371] A "generative AI model" is an artificial intelligence model that performs summarization and other processing on given text data.

[1372] A "communication device" refers to hardware or software used to transfer generated information to a specific device or system.

[1373] Modes for carrying out the invention

[1374] This invention relates to a system that efficiently manages incoming calls to a landline telephone, eliminates unnecessary sales calls, and notifies the user only of important calls. The following describes a specific embodiment for implementing this invention. This system mainly consists of three components: a server, a terminal, and a user.

[1375] server

[1376] The server receives incoming calls to a landline phone and initiates the call. Upon initiating the call, the server plays a pre-configured automated response message. This can be done using, for example, Python libraries or communication APIs.

[1377] The call content is recorded in real time, and the recorded audio data is converted to text using a speech recognition engine such as Google Cloud Speech-to-Text. In this process, the audio can be recorded using the Pydub library, uploaded to Google Cloud Storage, and then converted to text by calling an API.

[1378] The transformed text data is passed to a natural language processing (NLP) algorithm, such as the Hugging Face BERT model, to analyze the text. This then uses the Python Transformers library to determine whether the call is a sales call.

[1379] Based on the analysis results, the system determines whether the call is an unnecessary sales call or an important call. If it is determined to be a sales call, an automated response message can be generated and played back using a text-to-speech engine (e.g., Google Cloud Text-to-Speech). The text-to-speech engine uses an API to generate messages such as, "I'm sorry, but I can't help you right now."

[1380] If a call is deemed important, the server extracts the main points of the call and generates a summary using a method such as OpenAI's GPT-3. The summary is compiled into a concise document and ready to be sent to the user. The generated summary is transferred to the user's device via SMS or email.

[1381] terminal

[1382] The terminal receives a summary sent from the server. Typically, this is received on the user's mobile phone or computer and can be viewed through a notification interface. On mobile phones, it's received as an SMS message, and on computers, it's displayed in an email client.

[1383] User

[1384] Users review summaries sent from the server via their mobile phones or computers. This allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[1385] Specific examples of operation

[1386] The following shows a specific example of how this system works.

[1387] 1. When a relative calls the landline, the server receives the call and initiates the conversation.

[1388] 2. When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text using Google Cloud Speech-to-Text.

[1389] 3. The server analyzes the converted text using a BERT model and recognizes that it is a call from a relative checking on the recipient's well-being.

[1390] 4. The server determines that the call is important and responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[1391] 5. After the call ends, the server summarizes the call and generates a summary such as, "You received a call from a relative. They called to check on your recent health."

[1392] 6. The server sends the generated summary to the user's mobile phone via SMS.

[1393] Example of a prompt

[1394] "The following text is a transcript of a phone call received on a landline. Please summarize the main points of this call. Call content: 'How are you? How has your health been lately?'"

[1395] By inputting this prompt into the AI ​​generation model, important information can be extracted from the call content and notified to the user in an easy-to-understand format.

[1396] As described above, the present invention provides the effect of preventing users from missing important calls and freeing them from unwanted sales calls.

[1397] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1398] Step 1:

[1399] Answering incoming calls

[1400] The server receives incoming calls to a landline phone. The input is the landline phone's ringing signal, which the server detects and initiates the call. Upon initiating the call, the server plays a pre-configured automated response message. Specifically, it uses various Python libraries and communication APIs to detect the ringing signal and play the response message. The output is a call session in a waiting state.

[1401] Step 2:

[1402] Speech recognition

[1403] The server records incoming calls in real time and converts the audio data to text using a speech recognition engine such as Google Cloud Speech-to-Text. The input is the audio from the call, which is recorded using libraries such as Pydub. Once the recording is complete, the audio data is uploaded to Google Cloud Storage, and the Google Cloud Speech-to-Text API is called to convert the audio to text. The output is text data.

[1404] Step 3:

[1405] natural language analysis

[1406] The server passes the converted text data to a natural language processing algorithm for analysis. The input is the text data generated in step 2, which is analyzed using the Hugging Face BERT model. Specifically, the Python Transformers library is used to analyze the text using prompt phrases such as "Is this a sales call?". The output is the result of the analysis.

[1407] Step 4:

[1408] Determining whether action is necessary

[1409] The server determines whether a call is an important call or an unnecessary sales call based on the results of natural language processing. The input is the analysis result from step 3, and an algorithm is applied to determine whether it is an unnecessary sales call. The output is the result of the determination of whether the call is an important call or a sales call.

[1410] Step 5:

[1411] Automated response

[1412] If the server determines it is a sales call, it generates an automated response message and plays it back using a speech synthesis engine (e.g., Amazon Polly). The input is the result of the determination in step 4, which is specifically applied if it is a sales call. Specifically, it generates a message such as "I'm sorry, but I can't help you right now," synthesizes it into speech using the Amazon Polly API, and plays it back using the Twilio API. The output is the voice response to the sales call.

[1413] Step 6:

[1414] Summary of the call

[1415] The server prepares to summarize the call content and notify the user. The input is the result of the judgment in step 4, using the text that was determined to be from an important call as a prompt. Specifically, a generative AI model such as OpenAI's GPT-3 is used to input a prompt such as, "The following text is the content of a call received on a landline. Please summarize the main points of this call," and generates a summary. The output is the summarized text.

[1416] Step 7:

[1417] Summary transfer

[1418] The server notifies the user's device of the generated summary. The input is the summary text generated in step 6, which is formatted appropriately and sent to the user's mobile phone or email address. Specifically, this is done by sending an SMS using the Twilio API or by sending an email via an SMTP server. The output is the summary notification sent to the user's device.

[1419] (Application Example 1)

[1420] 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".

[1421] Traditional landline phones are often plagued by unwanted calls such as sales calls and scam calls, creating a risk of missing important calls. Furthermore, it's difficult to understand the content of important calls when the user is away or during busy periods. Therefore, there is a need for a system that efficiently manages incoming landline calls, automatically filters out unwanted sales and scam calls, and notifies the user only of important calls.

[1422] 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.

[1423] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, and means for notifying the smartphone of the call content summary. This enables efficient management of calls received on a landline, allowing the user to quickly grasp important call content without being bothered by unwanted sales calls or scam calls.

[1424] "Speech recognition" is a technology that analyzes speech data and converts it into text data.

[1425] "Natural language processing" is a technology that analyzes text obtained through speech recognition and understands and classifies its content.

[1426] An "automated response message" is a message that is automatically generated based on specific conditions and responds to the caller in voice.

[1427] A "summary" is information that extracts the main points of an important phone call and presents them concisely.

[1428] "Means of forwarding to the user" refers to the function of sending summarized call content to the user's device.

[1429] A "smartphone" is a portable information terminal that receives summaries of call content and notifies the user.

[1430] A "generative AI model" is an artificial intelligence algorithm that generates new information based on training data.

[1431] A "prompt statement" is an instruction given to a generative AI model and is used to obtain a specific output.

[1432] Modes for carrying out the invention

[1433] server

[1434] The server has key functions for efficiently managing incoming calls to landlines. This is done using the following hardware and software:

[1435] Speech Recognition: The server uses a speech recognition library (e.g., the speech_recognition library) to record incoming calls on the landline in real time and convert the content to text.

[1436] Natural Language Processing: The server analyzes the transcribed call content using a natural language processing library (e.g., the transformers library). This analysis determines whether the call is a sales call or a scam call.

[1437] Generative AI Model: The server uses a generative AI model to create automated response messages and summaries of important calls based on the analysis results. The generative AI model uses a pre-trained algorithm.

[1438] Summary Notification: If the server determines a call is important, it generates a summary and notifies the user's smartphone. The summarized call content is concisely presented to reduce the user's burden.

[1439] For example, if a relative calls to check on your well-being, the server will use speech recognition to process the message, such as "Hello, how have you been feeling lately?", and then analyze it as text. If the analysis determines that the call is important, it will generate a summary and notify your smartphone, "A relative has contacted you. They called to check on your recent health."

[1440] terminal

[1441] The terminal receives summaries sent from the server and notifies the user. This typically includes the user's smartphone or computer.

[1442] Notification Reception: The device receives a summarized call content and displays it to the user. Notifications are often displayed on the smartphone as push notifications.

[1443] User

[1444] Users can review a summary of the call content through their device and take necessary actions.

[1445] Notification Confirmation: Users can check notifications on their devices to understand the content of important calls. This allows them to respond only to important calls without being bothered by unwanted sales calls or scam calls.

[1446] As supplementary information, here are some examples of prompt statements that are input to the generative AI model:

[1447] "Please convert the following audio call into text and use natural language processing to determine if it is a sales call. If it is determined to be a sales call, please enter an automated response message."

[1448] This embodiment of the invention allows users to efficiently manage incoming calls on their landline and reliably identify only important calls.

[1449] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1450] Step 1:

[1451] The server receives incoming calls to a landline phone. The input is audio data from the landline. The server records this audio data in real time. The output is a recorded audio file. Specifically, the server uses an audio recording device to save the call.

[1452] Step 2:

[1453] The server uses a speech recognition library (e.g., speech_recognition) to convert recorded audio files into text. The input is a recorded audio file. The server uses a speech recognition engine to convert the audio data into text data. The output is the converted text. Specifically, the server initializes the speech recognition engine and analyzes and converts the audio data.

[1454] Step 3:

[1455] The server uses a natural language processing library (e.g., transformers) to analyze the transformed text. The input is the text obtained by speech recognition. The server uses a natural language processing engine to analyze the call content to determine whether it is a sales call or a scam call. The output is the analysis result. Specifically, the server uses a natural language processing model to classify the text data.

[1456] Step 4:

[1457] The server uses a generative AI model to generate automated response messages based on the analysis results. The input is the analysis results of natural language processing. The server inputs prompt sentences into the generative AI model to create automated response messages. The output is the generated automated response message. Specifically, the server provides prompt sentences to the generative AI model and generates an appropriate response.

[1458] Step 5:

[1459] The server summarizes the call content and prepares it for transfer to the smartphone. The input consists of analysis results and the text data of the call. The server uses a generative AI model to summarize the key points of the call and condense them into a concise format. The output is the summarized call content. Specifically, the server uses the generative AI model to extract the main points of the call and create a summary.

[1460] Step 6:

[1461] The server notifies the user's smartphone with a summary. The summarized call content is provided as input. The server uses push notifications to send a notification to the smartphone. The output is a summary notification displayed on the smartphone. Specifically, the server calls a notification API and sends the generated summary to the smartphone.

[1462] Step 7:

[1463] The user checks the notification on their smartphone and takes action as needed. The input is a summary of the notification. The user reviews the received summary to understand the important call content. The output is the user's action (e.g., returning the call). Specifically, the user opens the smartphone notification, reviews the call content, and decides on a course of action.

[1464] 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.

[1465] This invention relates to a system that efficiently manages incoming calls to a landline, eliminates unnecessary sales calls, and notifies the user only of important calls. Furthermore, by combining this with an emotion engine, the invention provides a system that can recognize the user's emotional state and respond accordingly.

[1466] server

[1467] 1. Answering incoming phone calls

[1468] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers.

[1469] 2. Speech Recognition

[1470] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This process makes the call content analyzable as text.

[1471] 3. Natural Language Processing (NLP)

[1472] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. This analysis determines whether the call is a sales call or not.

[1473] 4. Emotion recognition

[1474] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data and stores that information.

[1475] 5. Determining whether action is necessary.

[1476] The server determines whether the call is important or unwanted, such as a sales call, based on the results of natural language processing and sentiment recognition. If the call is important, it proceeds to the next step.

[1477] 6. Automated response

[1478] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. This message will use synthesized speech to say something like, "I'm sorry, but I can't assist you right now." It can also generate different messages depending on the user's emotional state.

[1479] 7. Summary of the call

[1480] If the server determines a call is important, it prepares to summarize the call content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user. The results of sentiment recognition can also be included in the summary.

[1481] 8. Transfer of Summary

[1482] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative has contacted you. They called to check on your recent health."

[1483] terminal

[1484] 1. Receiving

[1485] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[1486] User

[1487] 1. Confirmation of notification

[1488] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[1489] Specific example

[1490] 1. Power reception

[1491] When a relative calls the landline, the server receives the call and initiates the conversation.

[1492] 2. Speech Recognition

[1493] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[1494] 3. Natural Language Processing

[1495] The server analyzes the converted text data and understands that it is a phone call from a relative checking on the person's well-being.

[1496] 4. Emotion recognition

[1497] The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (for example, worry or relief).

[1498] 5. Determining whether action is necessary.

[1499] The server will appropriately determine that this call is important.

[1500] 6. Automated response

[1501] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message." It then generates a more appropriate response message depending on the relative's emotional state.

[1502] 7. Summary of the call

[1503] After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative contacted you. They called to check on your recent health. The person who contacted you was worried."

[1504] 8. Transfer of Summary

[1505] The server notifies the user's mobile phone of the generated summary.

[1506] This processing flow allows users to efficiently receive important call information and the caller's emotional state, freeing them from unnecessary sales calls.

[1507] The following describes the processing flow.

[1508] Step 1: Answering incoming calls

[1509] The server receives incoming calls to the landline phone. The server detects that a call has been made through the landline phone's interface, activates the automated answering module, and initiates the call.

[1510] Step 2: Speech Recognition

[1511] The server records the call in real time. The server uses a speech recognition engine to convert the recorded audio data into text. This makes the call content analyzable as text data.

[1512] Step 3: Natural Language Processing (NLP)

[1513] The server passes the converted text data to a natural language processing (NLP) algorithm. The server uses the NLP module to analyze the text content and extract important information and keywords.

[1514] Step 4: Emotion Recognition

[1515] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data and stores that information. For example, it identifies emotions such as "worried" or "angry."

[1516] Step 5: Determine whether action is necessary.

[1517] The server determines whether the call is important or unwanted, such as a sales call, based on the results of natural language processing and sentiment recognition. If the call is important, it proceeds to the next step.

[1518] Step 6: Automated response

[1519] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. This message will use synthesized speech to say something like, "I'm sorry, but I can't assist you right now." It can also generate different messages depending on the user's emotional state.

[1520] Step 7: Summarize the call

[1521] If the server determines a call is important, it prepares to summarize the call content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user. The results of sentiment recognition can also be included in the summary.

[1522] Step 8: Transferring the summary

[1523] The server generates a summary and notifies the user's device. Specifically, it sends a summary message to the user's mobile phone or email. For example, the notification might say, "A relative contacted you. They called to check on your recent health. They were worried about you."

[1524] Step 9: Final Check

[1525] The user reviews the summary via their device. This summary includes the main points of the call and the caller's emotional state. This allows the user to efficiently grasp the important points of the call and the caller's emotional state.

[1526] (Example 2)

[1527] 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".

[1528] The challenge lies in efficiently distinguishing between unwanted sales calls and important calls on landlines, thereby saving users time and effort. Furthermore, it is necessary to analyze the emotional state of the caller to provide more appropriate responses.

[1529] 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.

[1530] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for summarizing important call content and forwarding it to the user, means for analyzing the emotional state during the call, and means for generating summary information including the emotion analysis results. This makes it possible to eliminate unnecessary sales calls and efficiently grasp important call content. Furthermore, it is possible to respond based on the emotions of the caller, thereby improving user satisfaction.

[1531] "Received call" refers to the audio signal that the server receives when a phone call comes in.

[1532] "Speech recognition" is a technology that converts received audio signals into text data.

[1533] "Natural language processing" is a technology that analyzes text data obtained through speech recognition to understand its content.

[1534] "Determining whether action is necessary" means deciding whether the analyzed call content is important or unnecessary.

[1535] An "automatic response message" is a message that a server automatically generates and sends to the caller in response to an unnecessary call.

[1536] A "summary" is information that extracts the main points of an important phone call and presents them concisely.

[1537] "Transferring to the user" means sending the generated summary information to the user's device.

[1538] "Analyzing emotional state" means analyzing the emotions of the person on the other end of a call from their voice or text and determining their emotional state.

[1539] "Emotional analysis results" refer to information obtained as a result of analyzing an emotional state.

[1540] "Summary information" refers to a short explanatory text that includes the main points of the call and the results of sentiment analysis.

[1541] This invention relates to a system that efficiently manages incoming calls on a landline, eliminates unnecessary sales calls, and notifies the user only of important calls. Furthermore, it provides a system that analyzes the emotional state of the caller and responds accordingly to the user's emotional state.

[1542] System Configuration

[1543] Hardware and software

[1544] The server plays a central role in the system and uses the following main components:

[1545] 1. Landline telephone receiving device: Receives calls from landlines.

[1546] 2. Speech recognition engine: Google Cloud Speech-to-Text, etc.

[1547] 3. Natural Language Processing Engine: TensorFlow-based algorithm.

[1548] 4. Emotion recognition engine: IBM Watson, etc.

[1549] 5. Automated response function: Synthetic speech engines such as Asterisk and Amazon Polly.

[1550] 6. Database: MongoDB or similar database to store call content and sentiment analysis results.

[1551] The terminal is a device for receiving notifications from the server and includes the following components:

[1552] 1. Mobile phone or computer: A device for the user to view the summary.

[1553] 2. Notification system: APIs such as Firebase Cloud Messaging.

[1554] Program processing

[1555] When the server receives a call from a landline, it launches automated answering software such as Asterisk and initiates the call. The call content is converted into text data in real time using Google Cloud Speech-to-Text. The converted text data is then passed to a TensorFlow-based natural language processing engine for analysis.

[1556] Natural language processing is used to determine whether a call is a sales call, and at the same time, an emotion recognition engine such as IBM Watson is used to analyze the emotional state of the caller. Based on these results, the server determines whether the call is important or unnecessary.

[1557] If the call is a sales call, Amazon Polly's synthesized speech engine is used to generate an appropriate automated response message and notify the caller. For important calls, a summary of the call content is generated and notified to the user, including sentiment information.

[1558] The device receives summaries from the server using Firebase Cloud Messaging or similar methods, allowing the user to review them. For example, a notification might appear on the mobile phone screen stating, "A relative called to check on your health."

[1559] Users review summaries notified via their devices and take follow-up actions as needed. This system frees users from unnecessary calls and allows them to efficiently manage only important calls.

[1560] Specific example

[1561] 1. Receiving a call: When a relative calls the landline, the server receives the call and initiates the conversation.

[1562] 2. Voice Recognition: When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[1563] 3. Natural Language Processing: The server analyzes the converted text data and understands that it is a call from a relative checking on the user's well-being.

[1564] 4. Emotion Recognition: The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (e.g., worry or relief).

[1565] 5. Determination of whether action is needed: The server appropriately determines that this call is important.

[1566] 6. Automated response: The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message."

[1567] 7. Summary of Call Content: After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative called to check on your recent health. The person who called was worried."

[1568] 8. Summary Transfer: The server notifies the user's mobile phone of the generated summary.

[1569] Examples of prompts for generative AI models

[1570] Example prompt: "Describe a system that converts audio data into text data in real time, analyzes it, and then summarizes and notifies the user of the important content of the call."

[1571] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1572] Step 1:

[1573] Answering incoming calls

[1574] Input: A ringing signal received on a landline phone.

[1575] Specific operation: The server receives incoming calls to a landline phone.

[1576] Data processing / data calculation: Launch automated answering software such as Asterisk and initiate a call.

[1577] Output: An automated response message is played to the caller.

[1578] Step 2:

[1579] Speech recognition

[1580] Input: Voice data collected in real time during a call.

[1581] Specific operation: The server records the call content in real time.

[1582] Data processing / data calculation: Convert audio data to text data using Google Cloud Speech-to-Text.

[1583] Output: Call content data in string format.

[1584] Step 3:

[1585] Natural Language Processing (NLP)

[1586] Input: Speech-recognized text data.

[1587] Specific operation: The server passes text data to a TensorFlow-based NLP algorithm.

[1588] Data processing / data calculation: Analyze the content of the call and apply an algorithm to determine whether it is a sales call.

[1589] Output: A flag indicating whether the call is a sales call (True / False).

[1590] Step 4:

[1591] emotion recognition

[1592] Input: Speech-recognized text data and audio data.

[1593] Specific operation: The server passes text data and audio data to IBM Watson's emotion recognition engine.

[1594] Data processing / data calculation: Analyze the emotional state of the person on the other end of the phone call.

[1595] Output: The emotional state of the person on the other end of the call (e.g., worried, relieved).

[1596] Step 5:

[1597] Determining whether action is necessary

[1598] Input: Results of natural language processing and emotion recognition.

[1599] Specific operation: The server determines the importance of the call based on the analysis results.

[1600] Data processing / data calculation: Apply logic to determine whether a call is important or unnecessary.

[1601] Output: A flag indicating the importance of the call (important / unnecessary).

[1602] Step 6:

[1603] Automated response

[1604] Input: If the call is deemed unnecessary.

[1605] Specific operation: The server uses Amazon Polly to generate an automated response message.

[1606] Data processing / data calculation: Outputs the generated response message using synthesized speech.

[1607] Output: Automated response to the caller.

[1608] Step 7:

[1609] Summary of the call

[1610] Input: Text data if the call is deemed important.

[1611] Specific operation: The server uses an NLP library such as SpaCy to extract the main points of the call.

[1612] Data processing / calculation: Summarize call content, incorporating emotional information into the final product.

[1613] Output: Summary data including call content and sentiment information.

[1614] Step 8:

[1615] Summary transfer

[1616] Input: Summary data.

[1617] Specific operation: The server uses the Firebase Cloud Messaging API to send a summary to the user's device.

[1618] Data processing / data calculation: Convert summarized data into an appropriate format and send it.

[1619] Output: A summary message displayed on the user's terminal.

[1620] Examples of prompts for generative AI models

[1621] Example prompt: "Describe a system that converts audio data into text data in real time, analyzes it, and then summarizes and notifies the user of the important content of the call."

[1622] (Application Example 2)

[1623] 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".

[1624] This invention relates to a system that efficiently manages incoming calls on landlines, eliminates unnecessary sales calls, and notifies the user only of important calls. It also aims to enable food delivery drivers to streamline customer interactions and quickly convey important information.

[1625] 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.

[1626] In this invention, the server includes means for speech recognition of the content of an incoming call, means for analyzing the speech-recognized call content using natural language processing, means for determining whether a response is necessary based on the analyzed call content, means for generating and responding to an automatic response message for unnecessary calls, means for converting the content of a customer call from voice data to text data, means for analyzing the converted text data using sentiment recognition and generating a necessary response, and means for summarizing important call content and forwarding it to the user. This enables the management of incoming calls to landlines and improves the work efficiency of food delivery drivers.

[1627] "Content of received calls" refers to the audio information of calls received through the communication system.

[1628] "Means of speech recognition" refers to technologies for converting speech data into text data.

[1629] "Natural language processing" is a set of techniques for analyzing text data and understanding its content and meaning.

[1630] "Methods for determining whether action is necessary based on analyzed call content" refers to a method of determining the importance of the call content based on speech-recognized text.

[1631] "Means for generating and responding with automated response messages" refers to technology that automatically creates an appropriate response message based on analysis results and transmits that message to the caller.

[1632] "A means of summarizing important call content and forwarding it to the user" refers to a method of concisely summarizing the content of a call and notifying the user of that summary.

[1633] "Customer calls" refer to voice conversations between food delivery drivers and service users.

[1634] "Methods for converting audio data to text data" refers to technologies for converting the content of a phone call into text information.

[1635] "Emotion recognition" is a technology that determines a speaker's emotions from audio or text data.

[1636] "Means for generating necessary responses" refers to a method that automatically creates appropriate responses based on the results of emotion recognition and text analysis.

[1637] This invention is a system that efficiently manages incoming calls on landlines, eliminates unnecessary sales calls, and notifies the user only of important calls. It also enables food delivery drivers to streamline customer communication and quickly convey important information.

[1638] Server operation

[1639] 1. Answering incoming phone calls

[1640] The server receives incoming calls to a landline phone and initiates the call. Once the call is initiated, the server automatically answers. The hardware used includes server equipment and a communication interface. The software used is an automated voice response system.

[1641] 2. Speech Recognition

[1642] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. This utilizes the speech_recognition library.

[1643] 3. Natural Language Processing (NLP)

[1644] The server passes the speech-recognized text to a natural language processing algorithm to analyze the call content. The transformers library is used for this purpose.

[1645] 4. Emotion recognition

[1646] The server uses an emotion engine to analyze the emotions of the person on the call. It determines the user's emotional state from voice and text data.

[1647] 5. Determining whether action is necessary.

[1648] Based on the results of natural language processing and emotion recognition, the server determines whether the call is important or unwanted, such as a sales call.

[1649] 6. Automated response

[1650] If the server determines from the call content that it is a sales call, it will automatically generate an appropriate response message. Using synthesized speech, it will respond with phrases such as, "I'm sorry, but I can't assist you right now." It is also possible to generate different messages depending on the user's emotional state.

[1651] 7. Summary of the call

[1652] If a call is deemed important, preparations are made to summarize its content. After the call ends, the server uses NLP to extract key points and generates a summary to notify the user.

[1653] 8. Transfer of Summary

[1654] The server notifies the user's device of the generated summary. Specifically, it sends a summary message to the user's mobile phone or email address.

[1655] Terminal operation

[1656] 1. Receiving

[1657] The terminal receives a summary sent from the server. This terminal is usually the user's mobile phone or computer, and they can view the notified summary.

[1658] User actions

[1659] 1. Confirmation of notification

[1660] Users review summaries sent from the server via their terminals. This system allows users to filter out unwanted sales calls and efficiently understand only the important content of conversations.

[1661] Specific example

[1662] 1. Power reception

[1663] When a relative calls the landline, the server receives the call and initiates the conversation.

[1664] 2. Speech Recognition

[1665] When a relative says, "Hello, how have you been feeling lately?", the server records the conversation and immediately converts it to text.

[1666] 3. Natural Language Processing

[1667] The server analyzes the converted text data and understands that it is a phone call from a relative checking on the person's well-being.

[1668] 4. Emotion recognition

[1669] The server analyzes the relative's tone of voice and speech patterns to determine their emotional state (for example, worry or relief).

[1670] 5. Determining whether action is necessary.

[1671] The server will appropriately determine that this call is important.

[1672] 6. Automated response

[1673] The server responds to the relative with an AI voice saying, "We are unable to answer your call at this time, but we will take your message." It then generates a more appropriate response message depending on the relative's emotional state.

[1674] 7. Summary of the call

[1675] After the call ends, the server summarizes the content and generates a summary that includes emotional information, such as, "A relative contacted you. They called to check on your recent health. The person who contacted you was worried."

[1676] 8. Transfer of Summary

[1677] The server notifies the user's mobile phone of the generated summary.

[1678] Examples of prompt statements used

[1679] "Please upload the audio recording of your conversation with the customer."

[1680] Through this invention, it is possible to improve the efficiency of call management for landline telephones and the work of delivery personnel for food delivery services.

[1681] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1682] Step 1:

[1683] The server receives a call from a landline and initiates the call. This opens the call and prepares it for the subsequent steps. The input is the call signal from the telephone line, and the output is the call connection status.

[1684] Step 2:

[1685] The server records the call in real time and uses a speech recognition engine to convert the recorded audio data into text. The speech_recognition library is used for this process. The input is the audio data of the call, and the output is text data.

[1686] Step 3:

[1687] The server passes the speech-recognized text data to a natural language processing algorithm to analyze the call content. The transformers library is used here. The input is the text data obtained from speech recognition, and the output is the analysis result (semantic information of the call content).

[1688] Step 4:

[1689] The server uses an emotion engine to analyze the emotions of the person on the call. Specifically, it analyzes the speaker's emotional state from voice and text data. The input is text data and voice parameters, and the output is identified emotion information.

[1690] Step 5:

[1691] The server determines whether or not to respond to a call based on the results of natural language processing and emotion recognition. The input is the analysis results and emotion information, and the output is a response-required flag.

[1692] Step 6:

[1693] If the server determines from the call content that it is a sales call, it automatically generates an appropriate response message. This uses pre-synthesized voice. The input is a flag indicating whether a response is needed and text data, and the output is the response message.

[1694] Step 7:

[1695] If a call is deemed important, preparations are made to summarize the call content. Speech recognition and NLP tools are used to extract key points and generate a summary. Input is parsed text data and sentiment information, and output is the summarized text.

[1696] Step 8:

[1697] The server notifies the user's device of the generated summary. Specifically, it sends a summary message to the user's mobile phone or email. The input is the summarized text, and the output is a message that can be viewed on the user's receiving device.

[1698] Specific example

[1699] For example, the process flow when a relative calls and says, "Hi, how have you been feeling lately?" is as follows:

[1700] Step 1: The server receives the call and initiates the conversation.

[1701] Step 2: Record the conversation "Hello, how have you been feeling lately?" in real time and convert the audio data into text data.

[1702] Step 3: The converted text data is analyzed using a natural language processing algorithm to understand that it is a phone call from a relative checking on the user's well-being.

[1703] Step 4: The server analyzes the relative's tone of voice and speech patterns to determine if they are worried.

[1704] Step 5: Determine that this call is important and generate an appropriate response message that includes the relative's emotional state.

[1705] Step 6: Generate the answering message "We are currently unable to answer your call, but we will take your message" and send it to the caller.

[1706] Step 7: After the call ends, summarize the content and generate a summary such as, "A relative wanted to check on your recent health. The person who contacted them was worried."

[1707] Step 8: Notify the user of the summary on their mobile phone so they can efficiently grasp important call information.

[1708] In this way, the system of the present invention enables call management for landline telephones and improves the work efficiency of food delivery personnel.

[1709] 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.

[1710] 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.

[1711] 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.

[1712] 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.

[1713] 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.

[1714] 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.

[1715] 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.

[1716] 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.

[1717] 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."

[1718] 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.

[1719] 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.

[1720] 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.

[1721] 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.

[1722] 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.

[1723] 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.

[1724] 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.

[1725] 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.

[1726] 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.

[1727] 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.

[1728] 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.

[1729] 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.

[1730] The following is further disclosed regarding the embodiments described above.

[1731] (Claim 1)

[1732] A means of recognizing the content of an incoming call,

[1733] A means for analyzing the content of a voice-recognized call using natural language processing,

[1734] A means of determining whether action is necessary based on the analyzed call content,

[1735] A means of generating and responding to an automated message for unwanted calls,

[1736] A means of summarizing important call content and forwarding it to the user,

[1737] A system that includes this.

[1738] (Claim 2)

[1739] The system according to claim 1, comprising an algorithm for determining whether an incoming call is a sales call.

[1740] (Claim 3)

[1741] The system according to claim 1, comprising means for recording the content of a phone call in real time and saving the content thereof.

[1742] "Example 1"

[1743] (Claim 1)

[1744] A means of recognizing the content of an incoming call,

[1745] A means for analyzing the content of a voice-recognized call using natural language processing,

[1746] A means of determining whether action is necessary based on the analyzed call content,

[1747] A means of generating and responding to an automated message for unwanted calls,

[1748] A means of summarizing important call content and forwarding it to the user,

[1749] A method of using a speech recognition engine to convert call content into text,

[1750] A means of using a generative AI model to generate summaries,

[1751] Means of using a communication device for forwarding summary notifications,

[1752] A system that includes this.

[1753] (Claim 2)

[1754] The system according to claim 1, comprising an algorithm for determining whether an incoming call is a sales call.

[1755] (Claim 3)

[1756] The system according to claim 1, comprising means for recording the content of a phone call in real time and saving the content thereof.

[1757] "Application Example 1"

[1758] (Claim 1)

[1759] A means of recognizing the content of an incoming call,

[1760] A means for analyzing the content of a voice-recognized call using natural language processing,

[1761] A means of determining whether action is necessary based on the analyzed call content,

[1762] A means of generating and responding to an automated message for unwanted calls,

[1763] A means of summarizing important call content and forwarding it to the user,

[1764] A means of notifying a smartphone of a summary of the call content,

[1765] A system that includes this.

[1766] (Claim 2)

[1767] The system according to claim 1, comprising an algorithm for determining whether an incoming call is a sales call or a scam call.

[1768] (Claim 3)

[1769] The system according to claim 1, comprising means for recording the content of a call in real time and saving the content, and means for analyzing the speech-recognized text using an AI model to generate prompt sentences.

[1770] "Example 2 of combining an emotion engine"

[1771] (Claim 1)

[1772] A means of recognizing the content of an incoming call,

[1773] A means for analyzing the content of a voice-recognized call using natural language processing,

[1774] A means of determining whether action is necessary based on the analyzed call content,

[1775] A means of generating and responding to an automated message for unwanted calls,

[1776] A means of summarizing important call content and forwarding it to the user,

[1777] A means of analyzing emotional states during a phone call,

[1778] A means for generating summary information including sentiment analysis results,

[1779] A system that includes this.

[1780] (Claim 2)

[1781] The system according to claim 1, comprising an algorithm for determining whether an incoming call is a sales call.

[1782] (Claim 3)

[1783] The system according to claim 1, comprising means for recording the content of a phone call in real time and saving the content thereof.

[1784] "Application example 2 when combining with an emotional engine"

[1785] (Claim 1)

[1786] A means of recognizing the content of an incoming call,

[1787] A means for analyzing the content of a voice-recognized call using natural language processing,

[1788] A means of determining whether action is necessary based on the analyzed call content,

[1789] A means of generating and responding to an automated message for unwanted calls,

[1790] A means of summarizing important call content and forwarding it to the user,

[1791] A means of converting the content of a call between the moderator and the customer from audio data to text data,

[1792] A means for analyzing converted text data using emotion recognition and generating the necessary response,

[1793] A system that includes this.

[1794] (Claim 2)

[1795] The system according to claim 1, comprising an algorithm for determining whether an incoming call is a sales call.

[1796] (Claim 3)

[1797] The system according to claim 1, comprising means for recording the content of a phone call in real time and saving the content thereof. [Explanation of Symbols]

[1798] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of recognizing the content of an incoming call, A means for analyzing the content of a voice-recognized call using natural language processing, A means of determining whether action is necessary based on the analyzed call content, A means of generating and responding to an automated message for unwanted calls, A means of summarizing important call content and forwarding it to the user, A system that includes this.

2. The system according to claim 1, comprising an algorithm for determining whether an incoming call is a sales call.

3. The system according to claim 1, comprising means for recording the content of a phone call in real time and saving the content thereof.

Images