system

Abstract

The system according to this embodiment aims to reduce the response delay of a voice dialogue system and enable smoother conversations. [Solution] The system according to the embodiment comprises a prediction unit, a generation unit, and a provision unit. The prediction unit predicts the end and final sentence of a conversation. The generation unit generates a response based on the final sentence predicted by the prediction unit. The provision unit provides the response generated by the generation unit.

Description

Technical Field

[0006] , , , , ,

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a 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 the conventional technology, there is a problem that response delay occurs in the voice dialogue system, making smooth conversation difficult.

[0005] The system according to the embodiment aims to shorten the response delay of the voice dialogue system and realize smoother conversation.

Means for Solving the Problems

[0006] The system according to the embodiment includes a prediction unit, a generation unit, and a provision unit. The prediction unit predicts the end of the conversation and the last sentence. The generation unit generates a response based on the last sentence predicted by the prediction unit. The provision unit provides the response generated by the generation unit.

Effects of the Invention

[0007] The system according to this embodiment can reduce the response delay of a voice dialogue system and enable smoother conversations. [Brief explanation of the drawing]

[0008] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Modes for carrying out the invention]

[0009] Hereinafter, an example of an embodiment of the system relating to the technology of this disclosure will be described with reference to the attached drawings.

[0010] First, let's explain the terminology used in the following explanation.

[0011] In the following embodiments, the signed processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Furthermore, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), or TPU (Tensor Processing Unit).

[0012] In the following embodiments, signed RAM (Random Access Memory) is a memory that temporarily stores information and is used as work memory by the processor.

[0013] 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.

[0014] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna. 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).

[0015] 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 only A, only B, or a combination of A and B. Also, in this specification, when expressing three or more matters connected by "and / or", the same concept as "A and / or B" applies.

[0016] [First Embodiment] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0017] As shown in FIG. 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.

[0018] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. Also, the database 24 and the communication I / F 26 are connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0019] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. Also, the reception device 38, the output device 40, and the camera 42 are connected to the bus 52.

[0020] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, and accepts user input. The touch panel 38A accepts user input via touch by detecting contact with an object (e.g., a pen or finger). The microphone 38B accepts user input via voice 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 unit 12. In the data processing unit 12, the specific processing unit 290 (see Figure 2) acquires the data indicating the user input.

[0021] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user by outputting the data in a form perceptible to the user (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.

[0022] 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.

[0023] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0024] 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.

[0025] 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. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0026] In the smart device 14, specific processing is performed by the processor 46. The storage 50 stores a specific processing program 60. The specific processing program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the specific processing program 60 from the storage 50 and executes the read specific processing program 60 on the RAM 48. The specific processing is realized by the processor 46 operating as a control unit 46A according to the specific processing program 60 executed on the RAM 48. The smart device 14 also has a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0027] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device having the data generation model 58. The data processing device 12 may also be a server device or a terminal device owned by a user (e.g., a mobile phone, robot, home appliance, etc.). Next, an example of processing by the data processing system 10 according to the first embodiment will be described.

[0028] (Example of form 1) The voice dialogue system according to an embodiment of the present invention is a system that predicts the end of a conversation and the concluding sentence, and immediately generates a response. Unlike conventional systems that require multiple steps for a conversation to be established, this voice dialogue system predicts the end of a person's conversation and the concluding sentence, and immediately generates a response, thereby shortening response delays and achieving smoother conversations. This enables natural conversations and reduces response delays. It is believed that this proposal will provide a realistic experience and improve the quality of communication, thereby promoting more commercial use. For example, dialogue systems have become more familiar, with voice dialogue agents becoming commercial services in nursing homes and voice assistants for home use becoming widespread. However, it is difficult to say that the same smooth conversations between systems and humans as between humans have been achieved, and delays occur. Therefore, the content of this proposal is to improve response delay, which is one of the major challenges of current voice dialogue systems, and to promote more commercial use. For voice dialogue systems to become widespread, it is considered a prerequisite that the dialogue is as smooth as a conversation between humans and that situations of "dialogue breakdown" do not occur frequently. Therefore, this proposal proposes a system that predicts the end of a person's conversation and the concluding sentence, and generates a response. This system shortens response delays by stopping the voice data acquisition process midway through user speech and executing the speech recognition process early. Specifically, it consists of the following steps: First, it analyzes the user's speech in real time and predicts the end of the conversation and the final sentence. Next, it immediately generates a response based on the predicted final sentence. This response generation is performed using a generative AI. The generated response is immediately provided to the user, enabling smooth conversation. This system is particularly intended for use by elderly people living alone and in nursing homes. Daily conversation is important to alleviate feelings of loneliness and cognitive decline among the elderly, and it is expected that conversing with a robot incorporating this system will solve these problems. In addition, to investigate the effectiveness of this system, an experiment was conducted in which conversational sentences with the endings removed were provided to the generative AI to detect inconsistencies in the response.As a result, it was confirmed that response time could be reduced by approximately 2.5 seconds, and smoother conversations could be achieved by tolerating 17% of dialogue breakdowns. Furthermore, in an attempt to improve response speed, the importance of conversational keywords (TF-IDF value) was measured, and important words were detected. As a result, it was confirmed that dialogue breakdowns occurred when words with high TF-IDF values ​​were omitted. This makes it possible to generate responses that take important words into consideration, potentially further reducing response delays. This system can reduce response delays in voice dialogue systems and realize smoother conversations, thereby providing a more realistic experience and improving the quality of communication. This is expected to lead to increased commercial use and, in particular, to alleviate feelings of loneliness and cognitive decline among the elderly. In short, voice dialogue systems can reduce response delays and achieve smoother conversations.

[0029] The voice dialogue system according to this embodiment comprises a prediction unit, a generation unit, and a provision unit. The prediction unit predicts the end and final sentence of a conversation. The prediction unit predicts the end and final sentence of a conversation using, for example, a generation AI. The generation AI can predict the end and final sentence of a conversation using a text generation AI (e.g., LLM). The prediction unit can also predict the end and final sentence of a conversation using a multimodal generation AI. For example, the generation AI understands the context of the conversation and predicts the next sentence. The generation unit generates a response based on the final sentence predicted by the prediction unit. The generation unit generates a response using, for example, a generation AI. The generation AI can generate a response using a text generation AI (e.g., LLM). The generation unit can also generate a response using a multimodal generation AI. For example, the generation AI generates an appropriate response based on the predicted final sentence. The provision unit provides the response generated by the generation unit. The provision unit can also provide a response using, for example, a generation AI. The provision unit can also provide a response without using a generation AI. For example, the output unit plays the generated response as audio. The output unit can also display the generated response as text. For example, the output unit displays the generated response on a smartphone screen. As a result, the voice dialogue system according to this embodiment can predict the end of a conversation and the final sentence, generate a response immediately, reduce response delay, and achieve smoother conversation.

[0030] The prediction unit predicts the end of a conversation and the final sentence. The prediction unit predicts the end of a conversation and the final sentence, for example, using a generative AI. The generative AI can predict the end of a conversation and the final sentence using a text generation AI (e.g., LLM). Specifically, the generative AI learns from a large amount of past conversation data to understand the context of the conversation and predict the next sentence to come. This allows the generative AI to analyze the intent and emotion of the user's statements and predict a final sentence to end the conversation at the appropriate time. For example, if a user says, "I had a great time today," the generative AI will predict an appropriate final sentence such as "Let's meet again" based on that context. Furthermore, by using a multimodal generative AI, predictions can be made that take into account non-verbal information such as voice, facial expressions, and gestures. This enables more natural and human-like conversation endings. In addition, the prediction unit has an algorithm to monitor the progress of the conversation in real time and predict the final sentence at the appropriate time. This allows the appropriate final sentence to be provided the moment the user feels ready to end the conversation.

[0031] The generation unit generates a response based on the predicted ending sentence by the prediction unit. The generation unit generates responses using, for example, a generation AI. The generation AI can generate responses using a text generation AI (e.g., LLM). Specifically, the generation AI utilizes natural language processing techniques to generate appropriate responses for the user based on the predicted ending sentence. For example, if the user says, "I had a great time today," the generation AI will generate a response such as, "Let's meet again." Furthermore, by using a multimodal generation AI, it becomes possible to generate responses that take into account non-verbal information such as voice, facial expressions, and gestures. This allows for the generation of more natural and human-like responses. In addition, the generation unit can generate more personalized responses by considering the user's past conversation history and preferences. For example, if the user previously said, "I like movies," the generation AI can generate a response such as, "Let's talk about movies next time." This allows the generation unit to provide users with more approachable and interesting responses.

[0032] The service provider provides responses generated by the generation unit. The service provider can, for example, use generation AI to provide responses. Specifically, the service provider is equipped with speech synthesis technology to play the generated responses as audio. This allows users to receive responses in a natural voice. The service provider can also display the generated responses as text. For example, the service provider can display the generated responses on a smartphone screen, allowing users to visually confirm the responses. Furthermore, the service provider can provide responses in the most optimal way depending on the user's device and environment. For example, it can provide audio responses when used in a car and text responses in quiet places. The service provider also has a function to collect user feedback and continuously improve the quality of responses. This allows the service provider to provide users with quick and appropriate responses, resulting in smoother conversations.

[0033] The voice dialogue system includes an acquisition unit that acquires voice data. The acquisition unit can acquire voice data using, for example, a generative AI. The acquisition unit can also acquire voice data without using a generative AI. For example, the acquisition unit can acquire voice data using a microphone. The acquisition unit can also acquire voice data using an existing voice file. For example, the acquisition unit acquires user utterances in real time. By acquiring voice data, it is possible to provide the data necessary to predict the end of a conversation and the final sentence. Voice data includes, but is not limited to, user utterances, background sounds, and environmental sounds. Some or all of the above processing in the acquisition unit may be performed using, for example, a generative AI, or without using a generative AI. For example, the acquisition unit can input voice data acquired by the microphone into a generative AI and have the generative AI perform analysis of the voice data.

[0034] The voice dialogue system includes a recognition unit that performs speech recognition processing. The recognition unit performs speech recognition processing using, for example, a generative AI. The generative AI can perform speech recognition processing using a text generation AI (e.g., LLM). The recognition unit can also perform speech recognition processing using conventional speech recognition technology. For example, the recognition unit performs speech recognition processing using deep learning. The recognition unit analyzes the acquired voice data and uses it to predict the end of a conversation and the final sentence. Thus, by performing speech recognition processing, the acquired voice data can be analyzed and used to predict the end of a conversation and the final sentence. Some or all of the above processing in the recognition unit may be performed using, for example, a generative AI, or without a generative AI. For example, the recognition unit can input the acquired voice data into the generative AI and have the generative AI perform speech recognition processing.

[0035] The voice dialogue system includes a detection unit that detects important words. The detection unit detects important words, for example, using a generative AI. The generative AI can detect important words using a text generation AI (e.g., LLM). The detection unit can also detect important words using conventional natural language processing techniques. For example, the detection unit measures the TF-IDF value to detect important words. By detecting important words, the accuracy of predicting the end of a conversation and the concluding sentence can be improved. Important words include, but are not limited to, frequently occurring words, specific keywords, and contextually important words. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or without a generative AI. For example, the detection unit can input voice data recognized by the recognition unit into the generative AI and have the generative AI perform the detection of important words.

[0036] The prediction unit can analyze past conversation history and optimize its prediction algorithm based on the user's speaking patterns. For example, the prediction unit can analyze past conversation history using generative AI. The generative AI can analyze past conversation history using text generation AI (e.g., LLM). Alternatively, the prediction unit can analyze past conversation history using conventional natural language processing techniques. For example, the prediction unit can identify frequently used phrases and expressions by the user and adjust the prediction algorithm accordingly. The prediction unit can also analyze the tempo and rhythm of the user's conversation and optimize the prediction algorithm accordingly. Furthermore, the prediction unit can detect changes in speaking style in specific situations from the user's past conversation history and adjust the prediction algorithm based on these changes. This allows for the optimization of the prediction algorithm and improvement of prediction accuracy by analyzing past conversation history. Some or all of the above-described processes in the prediction unit may be performed using, for example, generative AI, or without generative AI. For example, the prediction unit can input past conversation history into the generative AI and have the generative AI optimize the prediction algorithm.

[0037] The prediction unit can adjust the timing of its predictions by taking into account the user's speaking speed and pauses. For example, the prediction unit can use generative AI to consider the user's speaking speed and pauses. The generative AI can use text generation AI (e.g., LLM) to consider the user's speaking speed and pauses. Alternatively, the prediction unit can use conventional natural language processing techniques to consider the user's speaking speed and pauses. For example, if the user speaks quickly, the prediction unit can advance the timing of its predictions to generate a quick response. Conversely, if the user speaks slowly, the prediction unit can delay the timing of its predictions to maintain a natural flow of conversation. Furthermore, if the user tends to pause, the prediction unit can adjust the timing of its predictions to take those pauses into account. In this way, by taking into account the user's speaking speed and pauses, the timing of predictions can be optimized, resulting in a natural conversation. Some or all of the above processing in the prediction unit may be performed using, for example, generative AI, or without using generative AI. For example, the prediction unit can input data on the user's speaking speed and pauses into the generating AI, allowing the generating AI to adjust the timing of the predictions.

[0038] The prediction unit can take the user's geographical location into consideration and reflect regional expressions and dialects in its predictions. For example, the prediction unit can consider the user's geographical location using generative AI. The generative AI can reflect regional expressions and dialects in its predictions using text generation AI (e.g., LLM). Alternatively, the prediction unit can also reflect regional expressions and dialects in its predictions using conventional natural language processing techniques. For example, if the user is in the Kansai region, the prediction unit will predict a sentence ending in Kansai dialect. If the user is in the Tohoku region, the prediction unit can also predict a sentence ending in Tohoku dialect. Furthermore, if the user is overseas, the prediction unit can predict a sentence ending that takes into account the language and dialect of that region. This allows for more natural conversation by reflecting regional expressions and dialects. Some or all of the above processing in the prediction unit may be performed using, for example, generative AI, or without using generative AI. For example, the prediction unit can input the user's geographical location information into the generating AI and have the AI ​​predict regionally specific expressions and dialects.

[0039] The prediction unit can analyze a user's social media activity and incorporate recent trends and topics into its predictions. For example, the prediction unit can analyze a user's social media activity using generative AI. The generative AI can analyze social media activity using text generation AI (e.g., LLM). Alternatively, the prediction unit can analyze social media activity using conventional natural language processing techniques. For example, the prediction unit can reflect topics that the user has recently posted about in its predictions. The prediction unit can also incorporate trends from influencers the user follows into its predictions. Furthermore, the prediction unit can reflect topics from online communities the user participates in. This allows for predictions that reflect recent trends and topics by analyzing social media activity. Some or all of the above processing in the prediction unit may be performed using, for example, generative AI, or without generative AI. For example, the prediction unit can input data on the user's social media activity into the generative AI and have the generative AI perform trend and topic predictions.

[0040] The generation unit can adjust the level of detail in the response based on the importance of the predicted final sentence. The generation unit can evaluate the importance of the predicted final sentence using, for example, a generation AI. The generation AI can evaluate the importance of the final sentence using a text generation AI (e.g., LLM). Alternatively, the generation unit can evaluate the importance of the final sentence using conventional natural language processing techniques. For example, if the final sentence is important, the generation unit can generate a response that includes a detailed explanation. If the final sentence is not very important, the generation unit can also generate a concise response. Furthermore, if the final sentence is of high interest to the user, the generation unit can generate a response that includes relevant information. This allows for the generation of more appropriate responses by adjusting the level of detail in the response based on the importance of the final sentence. Some or all of the above processing in the generation unit may be performed using, for example, a generation AI, or without a generation AI. For example, the generation unit can input data of the predicted final sentence into the generation AI and have the generation AI adjust the level of detail in the response.

[0041] The generation unit can generate the optimal response by referring to the user's past response patterns when generating a response. For example, the generation unit can analyze the user's past response patterns using a generation AI. The generation AI can analyze past response patterns using a text generation AI (e.g., LLM). Alternatively, the generation unit can analyze past response patterns using conventional natural language processing techniques. For example, the generation unit can generate a response that includes phrases the user has preferred to use in the past. The generation unit can also select the most appropriate tone and style from the user's past response patterns. Furthermore, the generation unit can generate a response considering topics the user has avoided in the past. This allows for the generation of more appropriate responses by referring to past response patterns. Some or all of the above processing in the generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the generation unit can input data on the user's past response patterns into a generation AI and have the generation AI generate the optimal response.

[0042] The generation unit can customize the response content based on the user's current situation and environment when generating a response. For example, the generation unit can consider the user's current situation and environment using a generation AI. The generation AI can consider the current situation and environment using a text generation AI (e.g., LLM). Alternatively, the generation unit can consider the current situation and environment using conventional natural language processing techniques. For example, if the user is out, the generation unit can generate a concise and practical response. If the user is at home, the generation unit can also generate a response that includes a detailed explanation. Furthermore, if the user is in a meeting, the generation unit can generate a response that can be quietly reviewed. This allows for the generation of more appropriate responses by customizing the response content based on the user's current situation and environment. Some or all of the above processing in the generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the generation unit can input data on the user's current situation and environment into the generation AI and have the generation AI customize the response content.

[0043] The generation unit can add relevant information based on the user's interests when generating responses. For example, the generation unit can analyze the user's interests using a generation AI. The generation AI can analyze interests using a text generation AI (e.g., LLM). Alternatively, the generation unit can analyze interests using conventional natural language processing techniques. For example, the generation unit can include information related to topics the user is interested in in its responses. The generation unit can also customize responses based on what the user has recently searched for. Furthermore, the generation unit can reflect the opinions of influencers the user follows in its responses. This allows for the generation of more appropriate responses by adding relevant information based on the user's interests. Some or all of the above-described processes in the generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the generation unit can input data on the user's interests into a generation AI and have the generation AI add relevant information.

[0044] The response unit can select the optimal timing for providing a response by referring to the user's past responses. The response unit can analyze the user's past responses using, for example, a generative AI. The generative AI can analyze past responses using a text generation AI (e.g., LLM). Alternatively, the response unit can analyze past responses using conventional natural language processing techniques. For example, if the response unit has previously preferred a quick response, it can provide an immediate response. The response unit can also provide a detailed response if the user has previously preferred a detailed explanation. Furthermore, if the response unit has previously preferred a response during a specific time period, it can provide a response during that time period. This allows for providing responses at a more appropriate time by referring to past responses. Some or all of the above processing in the response unit may be performed using, for example, a generative AI, or without a generative AI. For example, the response unit can input data on the user's past responses into a generative AI and have the generative AI select the timing for providing the response.

[0045] The service provider can select the optimal service delivery method when providing a response, taking into account the user's device information. For example, the service provider can analyze the user's device information using a generative AI. The generative AI can analyze the device information using a text generation AI (e.g., LLM). Alternatively, the service provider can analyze the device information using conventional natural language processing techniques. For example, if the user is using a smartphone, the service provider can provide a display method that matches the screen size. If the user is using a tablet, the service provider can also provide a display method optimized for a larger screen. Furthermore, if the user is using a smartwatch, the service provider can provide a concise and highly visible display method. This allows the service provider to provide a more appropriate response by taking device information into consideration. Some or all of the above-described processes in the service provider may be performed using a generative AI, for example, or without a generative AI. For example, the service provider can input the user's device information into a generative AI and have the generative AI select the service delivery method.

[0046] The service provider can select the optimal method of providing a response, taking into account the user's geographical location. For example, the service provider can analyze the user's geographical location using a generative AI. The generative AI can analyze the geographical location using a text generation AI (e.g., LLM). Alternatively, the service provider can analyze the geographical location using conventional natural language processing techniques. For example, if the user is out, the service provider can provide a concise and practical response. If the user is at home, the service provider can provide a response that includes a detailed explanation. Furthermore, if the user is in a meeting, the service provider can provide a response that can be quietly reviewed. This allows for the provision of responses in a more appropriate manner by considering geographical location. Some or all of the above-described processes in the service provider may be performed using, for example, a generative AI, or without a generative AI. For example, the service provider can input the user's geographical location into a generative AI and have the generative AI select the method of providing the response.

[0047] The service provider can analyze the user's social media activity and provide relevant information when providing a response. For example, the service provider can analyze the user's social media activity using generative AI. The generative AI can analyze social media activity using text generation AI (e.g., LLM). Alternatively, the service provider can analyze social media activity using conventional natural language processing techniques. For example, the service provider can provide information related to topics the user has recently posted about frequently. The service provider can also provide information that reflects the opinions of influencers the user follows. Furthermore, the service provider can provide information related to topics in online communities the user participates in. In this way, relevant information can be provided by analyzing social media activity. Some or all of the above processing in the service provider may be performed using, for example, generative AI, or without generative AI. For example, the service provider can input data on the user's social media activity into a generative AI and have the generative AI provide relevant information.

[0048] The acquisition unit can analyze the user's past speech history and select the optimal acquisition method. For example, the acquisition unit can analyze the user's past speech history using a generative AI. The generative AI can analyze the past speech history using a text generation AI (e.g., LLM). The acquisition unit can also analyze the past speech history using conventional natural language processing techniques. For example, the acquisition unit can adjust the method of acquiring audio data based on phrases that the user has frequently used in the past. The acquisition unit can also analyze the user's past speech patterns and select the optimal acquisition method based on that. Furthermore, the acquisition unit can detect changes in speech in specific situations from the user's past speech history and adjust the acquisition method based on that. By analyzing the past speech history, the optimal acquisition method can be selected and the accuracy of audio data acquisition can be improved. Some or all of the above processing in the acquisition unit may be performed using a generative AI, for example, or without a generative AI. For example, the acquisition unit can input data from the user's past speech history into a generative AI and have the generative AI select the optimal acquisition method.

[0049] The acquisition unit can filter audio data based on the user's current situation and environment. For example, the acquisition unit can analyze the user's current situation and environment using a generative AI. The generative AI can analyze the current situation and environment using a text generation AI (e.g., LLM). The acquisition unit can also analyze the current situation and environment using conventional natural language processing techniques. For example, if the user is out, the acquisition unit can filter out background noise to acquire audio data. If the user is at home, the acquisition unit can also acquire audio data while considering ambient noise. Furthermore, if the user is in a meeting, the acquisition unit can filter out ambient noise to acquire audio data. By filtering based on the current situation and environment, noise can be removed and the accuracy of audio data acquisition can be improved. Some or all of the above processing in the acquisition unit may be performed using a generative AI, or without one. For example, the acquisition unit can input data on the user's current situation and environment into the generative AI and have the generative AI perform the filtering.

[0050] The acquisition unit can prioritize the acquisition of highly relevant data by considering the user's geographical location information when acquiring audio data. The acquisition unit can analyze the user's geographical location information using, for example, a generative AI. The generative AI can analyze the geographical location information using a text generation AI (e.g., LLM). The acquisition unit can also analyze the geographical location information using conventional natural language processing techniques. For example, if the user is in the Kansai region, the acquisition unit will prioritize the acquisition of audio data using the Kansai dialect. Similarly, if the user is in the Tohoku region, the acquisition unit can prioritize the acquisition of audio data using the Tohoku dialect. Furthermore, if the user is overseas, the acquisition unit can prioritize the acquisition of audio data that takes into account the language and dialect of that region. This allows for the priority acquisition of highly relevant audio data by considering geographical location information. Some or all of the above processing in the acquisition unit may be performed using, for example, a generative AI, or without using a generative AI. For example, the acquisition unit can input the user's geographical location information into the generative AI and have the generative AI acquire highly relevant data.

[0051] The acquisition unit can analyze the user's social media activity and obtain relevant data when acquiring audio data. For example, the acquisition unit can analyze the user's social media activity using a generative AI. The generative AI can analyze social media activity using a text generation AI (e.g., LLM). The acquisition unit can also analyze social media activity using conventional natural language processing techniques. For example, the acquisition unit can acquire audio data related to topics that the user has recently posted about frequently. The acquisition unit can also acquire audio data that reflects the opinions of influencers that the user follows. Furthermore, the acquisition unit can acquire audio data related to topics in online communities that the user participates in. In this way, relevant audio data can be acquired by analyzing social media activity. Some or all of the above processing in the acquisition unit may be performed using a generative AI, for example, or without a generative AI. For example, the acquisition unit can input data on the user's social media activity into a generative AI and have the generative AI acquire the relevant data.

[0052] The recognition unit can optimize its recognition algorithm during speech recognition by considering the user's speech patterns and accent. For example, the recognition unit can analyze the user's speech patterns and accent using a generative AI. The generative AI can analyze speech patterns and accent using a text generation AI (e.g., LLM). The recognition unit can also analyze speech patterns and accent using conventional natural language processing techniques. For example, if the user speaks quickly, the recognition unit can adjust its recognition algorithm by considering the speech patterns. The recognition unit can also adjust its recognition algorithm by considering the speech patterns if the user speaks slowly. Furthermore, if the user has a specific accent, the recognition unit can optimize its recognition algorithm by considering that accent. In this way, by considering speech patterns and accents, the recognition algorithm can be optimized and the accuracy of speech recognition can be improved. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input data on the user's speech patterns and accent into a generative AI and have the generative AI perform the optimization of the recognition algorithm.

[0053] The recognition unit can improve recognition accuracy by referring to the user's past speech data during speech recognition. For example, the recognition unit can analyze the user's past speech data using a generative AI. The generative AI can analyze past speech data using a text generation AI (e.g., LLM). The recognition unit can also analyze past speech data using conventional natural language processing techniques. For example, the recognition unit can improve speech recognition accuracy based on phrases the user has used in the past. The recognition unit can also analyze the user's past speech data and optimize the recognition algorithm based on it. Furthermore, the recognition unit can detect changes in speech in specific situations from the user's past speech data and improve recognition accuracy based on that. In this way, recognition accuracy can be improved by referring to past speech data. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input the user's past speech data into a generative AI and have the generative AI perform the improvement of recognition accuracy.

[0054] The recognition unit can improve recognition accuracy by considering the user's geographical location information during speech recognition. For example, the recognition unit can analyze the user's geographical location information using a generative AI. The generative AI can analyze the geographical location information using a text generation AI (e.g., LLM). The recognition unit can also analyze the geographical location information using conventional natural language processing techniques. For example, if the user is in the Kansai region, the recognition unit can improve recognition accuracy by considering the Kansai dialect. Similarly, if the user is in the Tohoku region, the recognition unit can improve recognition accuracy by considering the Tohoku dialect. Furthermore, if the user is overseas, the recognition unit can improve recognition accuracy by considering the language and dialect of that region. In this way, recognition accuracy can be improved by considering geographical location information. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input the user's geographical location information into a generative AI and have the generative AI perform the improvement of recognition accuracy.

[0055] The recognition unit can improve recognition accuracy by analyzing the user's social media activity during speech recognition. For example, the recognition unit can analyze the user's social media activity using a generative AI. The generative AI can analyze social media activity using a text generation AI (e.g., LLM). The recognition unit can also analyze social media activity using conventional natural language processing techniques. For example, the recognition unit can use phrases related to topics the user has recently posted about to improve recognition accuracy. The recognition unit can also use phrases that reflect the opinions of influencers the user follows to improve recognition accuracy. Furthermore, the recognition unit can use phrases related to topics in online communities the user participates in to improve recognition accuracy. In this way, recognition accuracy can be improved by analyzing social media activity. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input data on the user's social media activity into a generative AI and have the generative AI perform the improvement of recognition accuracy.

[0056] The detection unit can optimize its detection algorithm by considering the user's speech patterns and accent when detecting important words. For example, the detection unit can analyze the user's speech patterns and accent using a generative AI. The generative AI can analyze speech patterns and accent using a text generation AI (e.g., LLM). Alternatively, the detection unit can analyze speech patterns and accent using conventional natural language processing techniques. For example, if the user speaks quickly, the detection unit adjusts its detection algorithm considering the speech patterns. Similarly, if the user speaks slowly, the detection unit can adjust its detection algorithm considering the speech patterns. Furthermore, if the user has a specific accent, the detection unit can optimize its detection algorithm by considering that accent. This allows for the optimization of the detection algorithm and improvement of word detection accuracy by considering speech patterns and accents. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or without a generative AI. For example, the detection unit can input data on the user's speech patterns and accent into a generative AI and have the generative AI optimize the detection algorithm.

[0057] The detection unit can improve detection accuracy by referring to the user's past speech data when detecting important words. The detection unit can analyze the user's past speech data using, for example, a generative AI. The generative AI can analyze past speech data using a text generation AI (e.g., LLM). The detection unit can also analyze past speech data using conventional natural language processing techniques. For example, the detection unit can improve the detection accuracy of important words based on phrases the user has used in the past. The detection unit can also analyze the user's past speech data and optimize the detection algorithm based on it. Furthermore, the detection unit can detect changes in speech in specific situations from the user's past speech data and improve detection accuracy based on that. In this way, detection accuracy can be improved by referring to past speech data. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or without a generative AI. For example, the detection unit can input the user's past speech data into a generative AI and have the generative AI perform the improvement of detection accuracy.

[0058] The detection unit can improve detection accuracy by considering the user's geographical location information when detecting important words. The detection unit can analyze the user's geographical location information using, for example, a generative AI. The generative AI can analyze geographical location information using a text generation AI (e.g., LLM). Alternatively, the detection unit can analyze geographical location information using conventional natural language processing techniques. For example, if the user is in the Kansai region, the detection unit can improve detection accuracy by considering the Kansai dialect. Similarly, if the user is in the Tohoku region, the detection unit can improve detection accuracy by considering the Tohoku dialect. Furthermore, if the user is overseas, the detection unit can improve detection accuracy by considering the language and dialect of that region. In this way, detection accuracy can be improved by considering geographical location information. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or without using a generative AI. For example, the detection unit can input the user's geographical location information into the generative AI and have the generative AI perform the improvement of detection accuracy.

[0059] The detection unit can improve detection accuracy by analyzing the user's social media activity when detecting important words. For example, the detection unit can analyze the user's social media activity using a generative AI. The generative AI can analyze social media activity using a text generation AI (e.g., LLM). The detection unit can also analyze social media activity using conventional natural language processing techniques. For example, the detection unit can use words related to topics that the user has recently posted about to improve detection accuracy. The detection unit can also use words that reflect the opinions of influencers that the user follows to improve detection accuracy. Furthermore, the detection unit can use words related to topics in online communities that the user participates in to improve detection accuracy. In this way, detection accuracy can be improved by analyzing social media activity. Some or all of the above processing in the detection unit may be performed using a generative AI, for example, or without a generative AI. For example, the detection unit can input data on the user's social media activity into a generative AI and have the generative AI perform the detection accuracy improvement.

[0060] The system according to the embodiment is not limited to the example described above, and various modifications are possible, for example, as follows.

[0061] A voice dialogue system may include a background sound generation unit that generates appropriate background sounds based on the user's utterances. The background sound generation unit may, for example, use a generation AI to generate background sounds suitable for the user's utterances. The generation AI may use a text generation AI (e.g., LLM) to generate background sounds. Alternatively, the background sound generation unit may use conventional speech synthesis technology to generate background sounds. For example, if the user is talking about nature, the background sound generation unit may generate sounds of birds chirping or wind. If the user is talking about a city, the background sound generation unit may also generate sounds of cars or people talking. Furthermore, if the user wants to relax, the background sound generation unit may also generate sounds of waves or quiet music. This allows for a more realistic conversation experience by generating appropriate background sounds based on the user's utterances. Some or all of the above-described processes in the background sound generation unit may be performed using, for example, a generation AI, or without a generation AI. For example, the background sound generation unit may input the user's utterances into the generation AI and have the generation AI perform background sound generation.

[0062] A voice dialogue system may include a gesture generation unit that generates appropriate gestures based on the user's utterances. The gesture generation unit generates gestures suitable for the user's utterances, for example, using a generation AI. The generation AI can generate gestures using a text generation AI (e.g., LLM). The gesture generation unit can also generate gestures using conventional motion capture technology. For example, when a user expresses gratitude, the gesture generation unit can generate a bowing or waving motion. When a user expresses surprise, the gesture generation unit can also generate an outstretched hand motion. Furthermore, when a user provides an explanation, the gesture generation unit can generate pointing or hand movements. This allows for a more natural conversational experience by generating appropriate gestures based on the user's utterances. Some or all of the above-described processes in the gesture generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the gesture generation unit can input the user's utterances into the generation AI and have the generation AI perform the gesture generation.

[0063] A voice dialogue system may include a visual effect generation unit that generates appropriate visual effects based on the user's utterances. The visual effect generation unit may, for example, use a generation AI to generate visual effects suitable for the user's utterances. The generation AI can generate visual effects using a text generation AI (e.g., LLM). The visual effect generation unit can also generate visual effects using conventional CG technology. For example, if the user is talking about fireworks, the visual effect generation unit can generate images of fireworks. If the user is talking about space, the visual effect generation unit can also generate images of a starry sky or planets. Furthermore, if the user wants to relax, the visual effect generation unit can generate images of a calm landscape. This allows for a more engaging conversation experience by generating appropriate visual effects based on the user's utterances. Some or all of the above-described processes in the visual effect generation unit may be performed using, for example, a generation AI, or without using a generation AI. For example, the visual effect generation unit can input the user's utterances into the generation AI and have the generation AI perform the generation of visual effects.

[0064] A voice dialogue system may include a music generation unit that generates appropriate music based on the user's utterances. The music generation unit may, for example, use a generation AI to generate music suitable for the user's utterances. The generation AI may use a text generation AI (e.g., LLM) to generate music. Alternatively, the music generation unit may use conventional music generation technologies to generate music. For example, if the user is telling a pleasant story, the music generation unit may generate bright and rhythmic music. If the user is telling a sad story, the music generation unit may generate quiet and sentimental music. Furthermore, if the user wants to relax, the music generation unit may generate calm and soothing music. This allows for a more emotionally rich conversational experience by generating appropriate music based on the user's utterances. Some or all of the above-described processes in the music generation unit may be performed using, for example, a generation AI, or without a generation AI. For example, the music generation unit may input the user's utterances into the generation AI and have the generation AI perform music generation.

[0065] A voice dialogue system may include a scent generation unit that generates an appropriate scent based on the user's utterances. The scent generation unit may, for example, use a generation AI to generate a scent suitable for the user's utterances. The generation AI may use a text generation AI (e.g., LLM) to generate scents. Alternatively, the scent generation unit may use conventional scent generation technologies to generate scents. For example, if the user is talking about flowers, the scent generation unit will generate a floral scent. If the user is talking about food, the scent generation unit may also generate a food scent. Furthermore, if the user wants to relax, the scent generation unit may generate lavender or mint scents. This allows for a more sensory-focused conversation experience by generating an appropriate scent based on the user's utterances. Some or all of the above-described processes in the scent generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the scent generation unit may input the user's utterances into the generation AI and have the generation AI perform the scent generation.

[0066] The following briefly describes the processing flow for example form 1.

[0067] Step 1: The prediction unit predicts the end and final sentence of the conversation. The prediction unit can predict the end and final sentence of the conversation using generative AI. The generative AI uses text generation AI (e.g., LLM) or multimodal generative AI to understand the context of the conversation and predict the next sentence. Step 2: The generation unit generates a response based on the ending sentence predicted by the prediction unit. The generation unit can generate a response using a generation AI. The generation AI uses a text generation AI (e.g., LLM) or a multimodal generation AI to generate an appropriate response based on the predicted ending sentence. Step 3: The provider unit provides the response generated by the generator unit. The provider unit can provide the response using a generation AI, or it can provide the response without using a generation AI. For example, the provider unit can play the generated response as audio or display it as text.

[0068] (Example of form 2) The voice dialogue system according to an embodiment of the present invention is a system that predicts the end of a conversation and the concluding sentence, and immediately generates a response. Unlike conventional systems that require multiple steps for a conversation to be established, this voice dialogue system predicts the end of a person's conversation and the concluding sentence, and immediately generates a response, thereby shortening response delays and achieving smoother conversations. This enables natural conversations and reduces response delays. It is believed that this proposal will provide a realistic experience and improve the quality of communication, thereby promoting more commercial use. For example, dialogue systems have become more familiar, with voice dialogue agents becoming commercial services in nursing homes and voice assistants for home use becoming widespread. However, it is difficult to say that the same smooth conversations between systems and humans as between humans have been achieved, and delays occur. Therefore, the content of this proposal is to improve response delay, which is one of the major challenges of current voice dialogue systems, and to promote more commercial use. For voice dialogue systems to become widespread, it is considered a prerequisite that the dialogue is as smooth as a conversation between humans and that situations of "dialogue breakdown" do not occur frequently. Therefore, this proposal proposes a system that predicts the end of a person's conversation and the concluding sentence, and generates a response. This system shortens response delays by stopping the voice data acquisition process midway through user speech and executing the speech recognition process early. Specifically, it consists of the following steps: First, it analyzes the user's speech in real time and predicts the end of the conversation and the final sentence. Next, it immediately generates a response based on the predicted final sentence. This response generation is performed using a generative AI. The generated response is immediately provided to the user, enabling smooth conversation. This system is particularly intended for use by elderly people living alone and in nursing homes. Daily conversation is important to alleviate feelings of loneliness and cognitive decline among the elderly, and it is expected that conversing with a robot incorporating this system will solve these problems. In addition, to investigate the effectiveness of this system, an experiment was conducted in which conversational sentences with the endings removed were provided to the generative AI to detect inconsistencies in the response.As a result, it was confirmed that response time could be reduced by approximately 2.5 seconds, and smoother conversations could be achieved by tolerating 17% of dialogue breakdowns. Furthermore, in an attempt to improve response speed, the importance of conversational keywords (TF-IDF value) was measured, and important words were detected. As a result, it was confirmed that dialogue breakdowns occurred when words with high TF-IDF values ​​were omitted. This makes it possible to generate responses that take important words into consideration, potentially further reducing response delays. This system can reduce response delays in voice dialogue systems and realize smoother conversations, thereby providing a more realistic experience and improving the quality of communication. This is expected to lead to increased commercial use and, in particular, to alleviate feelings of loneliness and cognitive decline among the elderly. In short, voice dialogue systems can reduce response delays and achieve smoother conversations.

[0069] The voice dialogue system according to this embodiment comprises a prediction unit, a generation unit, and a provision unit. The prediction unit predicts the end and final sentence of a conversation. The prediction unit predicts the end and final sentence of a conversation using, for example, a generation AI. The generation AI can predict the end and final sentence of a conversation using a text generation AI (e.g., LLM). The prediction unit can also predict the end and final sentence of a conversation using a multimodal generation AI. For example, the generation AI understands the context of the conversation and predicts the next sentence. The generation unit generates a response based on the final sentence predicted by the prediction unit. The generation unit generates a response using, for example, a generation AI. The generation AI can generate a response using a text generation AI (e.g., LLM). The generation unit can also generate a response using a multimodal generation AI. For example, the generation AI generates an appropriate response based on the predicted final sentence. The provision unit provides the response generated by the generation unit. The provision unit can also provide a response using, for example, a generation AI. The provision unit can also provide a response without using a generation AI. For example, the output unit plays the generated response as audio. The output unit can also display the generated response as text. For example, the output unit displays the generated response on a smartphone screen. As a result, the voice dialogue system according to this embodiment can predict the end of a conversation and the final sentence, generate a response immediately, reduce response delay, and achieve smoother conversation.

[0070] The prediction unit predicts the end of a conversation and the final sentence. The prediction unit predicts the end of a conversation and the final sentence, for example, using a generative AI. The generative AI can predict the end of a conversation and the final sentence using a text generation AI (e.g., LLM). Specifically, the generative AI learns from a large amount of past conversation data to understand the context of the conversation and predict the next sentence to come. This allows the generative AI to analyze the intent and emotion of the user's statements and predict a final sentence to end the conversation at the appropriate time. For example, if a user says, "I had a great time today," the generative AI will predict an appropriate final sentence such as "Let's meet again" based on that context. Furthermore, by using a multimodal generative AI, predictions can be made that take into account non-verbal information such as voice, facial expressions, and gestures. This enables more natural and human-like conversation endings. In addition, the prediction unit has an algorithm to monitor the progress of the conversation in real time and predict the final sentence at the appropriate time. This allows the appropriate final sentence to be provided the moment the user feels ready to end the conversation.

[0071] The generation unit generates a response based on the predicted ending sentence by the prediction unit. The generation unit generates responses using, for example, a generation AI. The generation AI can generate responses using a text generation AI (e.g., LLM). Specifically, the generation AI utilizes natural language processing techniques to generate appropriate responses for the user based on the predicted ending sentence. For example, if the user says, "I had a great time today," the generation AI will generate a response such as, "Let's meet again." Furthermore, by using a multimodal generation AI, it becomes possible to generate responses that take into account non-verbal information such as voice, facial expressions, and gestures. This allows for the generation of more natural and human-like responses. In addition, the generation unit can generate more personalized responses by considering the user's past conversation history and preferences. For example, if the user previously said, "I like movies," the generation AI can generate a response such as, "Let's talk about movies next time." This allows the generation unit to provide users with more approachable and interesting responses.

[0072] The service provider provides responses generated by the generation unit. The service provider can, for example, use generation AI to provide responses. Specifically, the service provider is equipped with speech synthesis technology to play the generated responses as audio. This allows users to receive responses in a natural voice. The service provider can also display the generated responses as text. For example, the service provider can display the generated responses on a smartphone screen, allowing users to visually confirm the responses. Furthermore, the service provider can provide responses in the most optimal way depending on the user's device and environment. For example, it can provide audio responses when used in a car and text responses in quiet places. The service provider also has a function to collect user feedback and continuously improve the quality of responses. This allows the service provider to provide users with quick and appropriate responses, resulting in smoother conversations.

[0073] The voice dialogue system includes an acquisition unit that acquires voice data. The acquisition unit can acquire voice data using, for example, a generative AI. The acquisition unit can also acquire voice data without using a generative AI. For example, the acquisition unit can acquire voice data using a microphone. The acquisition unit can also acquire voice data using an existing voice file. For example, the acquisition unit acquires user utterances in real time. By acquiring voice data, it is possible to provide the data necessary to predict the end of a conversation and the final sentence. Voice data includes, but is not limited to, user utterances, background sounds, and environmental sounds. Some or all of the above processing in the acquisition unit may be performed using, for example, a generative AI, or without using a generative AI. For example, the acquisition unit can input voice data acquired by the microphone into a generative AI and have the generative AI perform analysis of the voice data.

[0074] The voice dialogue system includes a recognition unit that performs speech recognition processing. The recognition unit performs speech recognition processing using, for example, a generative AI. The generative AI can perform speech recognition processing using a text generation AI (e.g., LLM). The recognition unit can also perform speech recognition processing using conventional speech recognition technology. For example, the recognition unit performs speech recognition processing using deep learning. The recognition unit analyzes the acquired voice data and uses it to predict the end of a conversation and the final sentence. Thus, by performing speech recognition processing, the acquired voice data can be analyzed and used to predict the end of a conversation and the final sentence. Some or all of the above processing in the recognition unit may be performed using, for example, a generative AI, or without a generative AI. For example, the recognition unit can input the acquired voice data into the generative AI and have the generative AI perform speech recognition processing.

[0075] The voice dialogue system includes a detection unit that detects important words. The detection unit detects important words, for example, using a generative AI. The generative AI can detect important words using a text generation AI (e.g., LLM). The detection unit can also detect important words using conventional natural language processing techniques. For example, the detection unit measures the TF-IDF value to detect important words. By detecting important words, the accuracy of predicting the end of a conversation and the concluding sentence can be improved. Important words include, but are not limited to, frequently occurring words, specific keywords, and contextually important words. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or without a generative AI. For example, the detection unit can input voice data recognized by the recognition unit into the generative AI and have the generative AI perform the detection of important words.

[0076] The prediction unit can estimate the user's emotions and adjust the prediction accuracy of the conversation's ending and concluding sentence based on the estimated emotions. The prediction unit is implemented using emotion estimation functionality, for example, with an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. For example, if the user is nervous, the prediction unit may analyze past conversation history in more detail to improve prediction accuracy. The prediction unit may also slightly loosen the prediction accuracy if the user is relaxed, prioritizing a natural flow of conversation. Furthermore, if the user is in a hurry, the prediction unit may maximize prediction accuracy to generate a quick response. This allows for more natural conversation by adjusting prediction accuracy based on the user's emotions. Some or all of the above processing in the prediction unit may be performed using, for example, a generative AI, or not using a generative AI. For example, the prediction unit can input user emotion data into a generative AI and have the generative AI adjust the prediction accuracy of the conversation's ending and concluding sentence.

[0077] The prediction unit can analyze past conversation history and optimize its prediction algorithm based on the user's speaking patterns. For example, the prediction unit can analyze past conversation history using generative AI. The generative AI can analyze past conversation history using text generation AI (e.g., LLM). Alternatively, the prediction unit can analyze past conversation history using conventional natural language processing techniques. For example, the prediction unit can identify frequently used phrases and expressions by the user and adjust the prediction algorithm accordingly. The prediction unit can also analyze the tempo and rhythm of the user's conversation and optimize the prediction algorithm accordingly. Furthermore, the prediction unit can detect changes in speaking style in specific situations from the user's past conversation history and adjust the prediction algorithm based on these changes. This allows for the optimization of the prediction algorithm and improvement of prediction accuracy by analyzing past conversation history. Some or all of the above-described processes in the prediction unit may be performed using, for example, generative AI, or without generative AI. For example, the prediction unit can input past conversation history into the generative AI and have the generative AI optimize the prediction algorithm.

[0078] The prediction unit can adjust the timing of its predictions by taking into account the user's speaking speed and pauses. For example, the prediction unit can use generative AI to consider the user's speaking speed and pauses. The generative AI can use text generation AI (e.g., LLM) to consider the user's speaking speed and pauses. Alternatively, the prediction unit can use conventional natural language processing techniques to consider the user's speaking speed and pauses. For example, if the user speaks quickly, the prediction unit can advance the timing of its predictions to generate a quick response. Conversely, if the user speaks slowly, the prediction unit can delay the timing of its predictions to maintain a natural flow of conversation. Furthermore, if the user tends to pause, the prediction unit can adjust the timing of its predictions to take those pauses into account. In this way, by taking into account the user's speaking speed and pauses, the timing of predictions can be optimized, resulting in a natural conversation. Some or all of the above processing in the prediction unit may be performed using, for example, generative AI, or without using generative AI. For example, the prediction unit can input data on the user's speaking speed and pauses into the generating AI, allowing the generating AI to adjust the timing of the predictions.

[0079] The prediction unit can estimate the user's emotions and determine the priority of predicted closing sentences based on the estimated user emotions. The prediction unit is implemented using emotion estimation functionality, for example, with an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. For example, if the user is excited, the prediction unit will prioritize closing sentences that calm the emotions. The prediction unit can also prioritize comforting closing sentences if the user is sad. Furthermore, if the user is happy, the prediction unit can prioritize closing sentences that maintain that emotion. This allows for the generation of more appropriate responses by determining the priority of closing sentences based on the user's emotions. Some or all of the above processing in the prediction unit may be performed using, for example, a generative AI, or not using a generative AI. For example, the prediction unit can input user emotion data into a generative AI and have the generative AI determine the priority of closing sentences.

[0080] The prediction unit can take the user's geographical location into consideration and reflect regional expressions and dialects in its predictions. For example, the prediction unit can consider the user's geographical location using generative AI. The generative AI can reflect regional expressions and dialects in its predictions using text generation AI (e.g., LLM). Alternatively, the prediction unit can also reflect regional expressions and dialects in its predictions using conventional natural language processing techniques. For example, if the user is in the Kansai region, the prediction unit will predict a sentence ending in Kansai dialect. If the user is in the Tohoku region, the prediction unit can also predict a sentence ending in Tohoku dialect. Furthermore, if the user is overseas, the prediction unit can predict a sentence ending that takes into account the language and dialect of that region. This allows for more natural conversation by reflecting regional expressions and dialects. Some or all of the above processing in the prediction unit may be performed using, for example, generative AI, or without using generative AI. For example, the prediction unit can input the user's geographical location information into the generating AI and have the AI ​​predict regionally specific expressions and dialects.

[0081] The prediction unit can analyze a user's social media activity and incorporate recent trends and topics into its predictions. For example, the prediction unit can analyze a user's social media activity using generative AI. The generative AI can analyze social media activity using text generation AI (e.g., LLM). Alternatively, the prediction unit can analyze social media activity using conventional natural language processing techniques. For example, the prediction unit can reflect topics that the user has recently posted about in its predictions. The prediction unit can also incorporate trends from influencers the user follows into its predictions. Furthermore, the prediction unit can reflect topics from online communities the user participates in. This allows for predictions that reflect recent trends and topics by analyzing social media activity. Some or all of the above processing in the prediction unit may be performed using, for example, generative AI, or without generative AI. For example, the prediction unit can input data on the user's social media activity into the generative AI and have the generative AI perform trend and topic predictions.

[0082] The generation unit can estimate the user's emotions and adjust the expression of its response based on the estimated emotions. The generation unit is implemented using an emotion estimation function, for example, with an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is angry, the generation unit will use a calm and composed expression. If the user is sad, the generation unit may also use a comforting expression. Furthermore, if the user is happy, the generation unit may also use a bright and cheerful expression. By adjusting the expression of the response based on the user's emotions, a more appropriate response can be generated. Some or all of the above processing in the generation unit may be performed using a generative AI, for example, or without a generative AI. For example, the generation unit can input user emotion data into a generative AI and have the generative AI adjust the expression of the response.

[0083] The generation unit can adjust the level of detail in the response based on the importance of the predicted final sentence. The generation unit can evaluate the importance of the predicted final sentence using, for example, a generation AI. The generation AI can evaluate the importance of the final sentence using a text generation AI (e.g., LLM). Alternatively, the generation unit can evaluate the importance of the final sentence using conventional natural language processing techniques. For example, if the final sentence is important, the generation unit can generate a response that includes a detailed explanation. If the final sentence is not very important, the generation unit can also generate a concise response. Furthermore, if the final sentence is of high interest to the user, the generation unit can generate a response that includes relevant information. This allows for the generation of more appropriate responses by adjusting the level of detail in the response based on the importance of the final sentence. Some or all of the above processing in the generation unit may be performed using, for example, a generation AI, or without a generation AI. For example, the generation unit can input data of the predicted final sentence into the generation AI and have the generation AI adjust the level of detail in the response.

[0084] The generation unit can generate the optimal response by referring to the user's past response patterns when generating a response. For example, the generation unit can analyze the user's past response patterns using a generation AI. The generation AI can analyze past response patterns using a text generation AI (e.g., LLM). Alternatively, the generation unit can analyze past response patterns using conventional natural language processing techniques. For example, the generation unit can generate a response that includes phrases the user has preferred to use in the past. The generation unit can also select the most appropriate tone and style from the user's past response patterns. Furthermore, the generation unit can generate a response considering topics the user has avoided in the past. This allows for the generation of more appropriate responses by referring to past response patterns. Some or all of the above processing in the generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the generation unit can input data on the user's past response patterns into a generation AI and have the generation AI generate the optimal response.

[0085] The generation unit can estimate the user's emotions and adjust the length of the response based on the estimated emotions. The generation unit is implemented using emotion estimation functionality, for example, with an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is in a hurry, the generation unit can generate a short, to-the-point response. If the user is relaxed, the generation unit can also generate a longer response that includes detailed explanations. Furthermore, if the user is excited, the generation unit can generate a response with visually stimulating effects. This allows for the generation of more appropriate responses by adjusting the length of the response based on the user's emotions. Some or all of the above processing in the generation unit may be performed using a generative AI, for example, or without a generative AI. For example, the generation unit can input user emotion data into a generative AI and have the generative AI adjust the length of the response.

[0086] The generation unit can customize the response content based on the user's current situation and environment when generating a response. For example, the generation unit can consider the user's current situation and environment using a generation AI. The generation AI can consider the current situation and environment using a text generation AI (e.g., LLM). Alternatively, the generation unit can consider the current situation and environment using conventional natural language processing techniques. For example, if the user is out, the generation unit can generate a concise and practical response. If the user is at home, the generation unit can also generate a response that includes a detailed explanation. Furthermore, if the user is in a meeting, the generation unit can generate a response that can be quietly reviewed. This allows for the generation of more appropriate responses by customizing the response content based on the user's current situation and environment. Some or all of the above processing in the generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the generation unit can input data on the user's current situation and environment into the generation AI and have the generation AI customize the response content.

[0087] The generation unit can add relevant information based on the user's interests when generating responses. For example, the generation unit can analyze the user's interests using a generation AI. The generation AI can analyze interests using a text generation AI (e.g., LLM). Alternatively, the generation unit can analyze interests using conventional natural language processing techniques. For example, the generation unit can include information related to topics the user is interested in in its responses. The generation unit can also customize responses based on what the user has recently searched for. Furthermore, the generation unit can reflect the opinions of influencers the user follows in its responses. This allows for the generation of more appropriate responses by adding relevant information based on the user's interests. Some or all of the above-described processes in the generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the generation unit can input data on the user's interests into a generation AI and have the generation AI add relevant information.

[0088] The service provider can estimate the user's emotions and adjust the method of providing responses based on the estimated emotions. The service provider is implemented using an emotion estimation function, for example, with an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. For example, if the user is nervous, the service provider can provide a response in a calm voice. The service provider can also provide a response in a cheerful voice if the user is relaxed. Furthermore, if the user is in a hurry, the service provider can provide a quick and concise response. This allows for the provision of more appropriate responses by adjusting the method of providing responses based on the user's emotions. Some or all of the above processing in the service provider may be performed using, for example, a generative AI, or not using a generative AI. For example, the service provider can input user emotion data into a generative AI and have the generative AI adjust the method of providing responses.

[0089] The response unit can select the optimal timing for providing a response by referring to the user's past responses. The response unit can analyze the user's past responses using, for example, a generative AI. The generative AI can analyze past responses using a text generation AI (e.g., LLM). Alternatively, the response unit can analyze past responses using conventional natural language processing techniques. For example, if the response unit has previously preferred a quick response, it can provide an immediate response. The response unit can also provide a detailed response if the user has previously preferred a detailed explanation. Furthermore, if the response unit has previously preferred a response during a specific time period, it can provide a response during that time period. This allows for providing responses at a more appropriate time by referring to past responses. Some or all of the above processing in the response unit may be performed using, for example, a generative AI, or without a generative AI. For example, the response unit can input data on the user's past responses into a generative AI and have the generative AI select the timing for providing the response.

[0090] The service provider can select the optimal service delivery method when providing a response, taking into account the user's device information. For example, the service provider can analyze the user's device information using a generative AI. The generative AI can analyze the device information using a text generation AI (e.g., LLM). Alternatively, the service provider can analyze the device information using conventional natural language processing techniques. For example, if the user is using a smartphone, the service provider can provide a display method that matches the screen size. If the user is using a tablet, the service provider can also provide a display method optimized for a larger screen. Furthermore, if the user is using a smartwatch, the service provider can provide a concise and highly visible display method. This allows the service provider to provide a more appropriate response by taking device information into consideration. Some or all of the above-described processes in the service provider may be performed using a generative AI, for example, or without a generative AI. For example, the service provider can input the user's device information into a generative AI and have the generative AI select the service delivery method.

[0091] The service provider can estimate the user's emotions and adjust the display method of the response based on the estimated user emotions. The service provider is implemented using emotion estimation functionality, for example, with an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. For example, if the user is nervous, the service provider provides a simple and highly visible display method. If the user is relaxed, the service provider can also provide a display method that includes detailed information. Furthermore, if the user is in a hurry, the service provider can provide a concise display method. This allows for the provision of more appropriate responses by adjusting the display method of the response based on the user's emotions. Some or all of the above processing in the service provider may be performed using, for example, a generative AI, or not using a generative AI. For example, the service provider can input user emotion data into a generative AI and have the generative AI perform the adjustment of the display method.

[0092] The service provider can select the optimal method of providing a response, taking into account the user's geographical location. For example, the service provider can analyze the user's geographical location using a generative AI. The generative AI can analyze the geographical location using a text generation AI (e.g., LLM). Alternatively, the service provider can analyze the geographical location using conventional natural language processing techniques. For example, if the user is out, the service provider can provide a concise and practical response. If the user is at home, the service provider can provide a response that includes a detailed explanation. Furthermore, if the user is in a meeting, the service provider can provide a response that can be quietly reviewed. This allows for the provision of responses in a more appropriate manner by considering geographical location. Some or all of the above-described processes in the service provider may be performed using, for example, a generative AI, or without a generative AI. For example, the service provider can input the user's geographical location into a generative AI and have the generative AI select the method of providing the response.

[0093] The service provider can analyze the user's social media activity and provide relevant information when providing a response. For example, the service provider can analyze the user's social media activity using generative AI. The generative AI can analyze social media activity using text generation AI (e.g., LLM). Alternatively, the service provider can analyze social media activity using conventional natural language processing techniques. For example, the service provider can provide information related to topics the user has recently posted about frequently. The service provider can also provide information that reflects the opinions of influencers the user follows. Furthermore, the service provider can provide information related to topics in online communities the user participates in. In this way, relevant information can be provided by analyzing social media activity. Some or all of the above processing in the service provider may be performed using, for example, generative AI, or without generative AI. For example, the service provider can input data on the user's social media activity into a generative AI and have the generative AI provide relevant information.

[0094] The acquisition unit can estimate the user's emotions and adjust the timing of voice data acquisition based on the estimated emotions. The acquisition unit is implemented using an emotion estimation function, for example, with an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. For example, if the user is nervous, the acquisition unit can advance the timing of voice data acquisition to enable a quick response. The acquisition unit can also delay the timing of voice data acquisition to maintain a natural flow of conversation if the user is relaxed. Furthermore, if the user is in a hurry, the acquisition unit can advance the timing of voice data acquisition to the maximum extent possible to enable a quick response. In this way, by adjusting the timing of voice data acquisition based on the user's emotions, voice data can be acquired at a more appropriate time. Some or all of the above processing in the acquisition unit may be performed using, for example, a generative AI, or not using a generative AI. For example, the acquisition unit can input user emotion data into a generative AI and have the generative AI perform the adjustment of the acquisition timing.

[0095] The acquisition unit can analyze the user's past speech history and select the optimal acquisition method. For example, the acquisition unit can analyze the user's past speech history using a generative AI. The generative AI can analyze the past speech history using a text generation AI (e.g., LLM). The acquisition unit can also analyze the past speech history using conventional natural language processing techniques. For example, the acquisition unit can adjust the method of acquiring audio data based on phrases that the user has frequently used in the past. The acquisition unit can also analyze the user's past speech patterns and select the optimal acquisition method based on that. Furthermore, the acquisition unit can detect changes in speech in specific situations from the user's past speech history and adjust the acquisition method based on that. By analyzing the past speech history, the optimal acquisition method can be selected and the accuracy of audio data acquisition can be improved. Some or all of the above processing in the acquisition unit may be performed using a generative AI, for example, or without a generative AI. For example, the acquisition unit can input data from the user's past speech history into a generative AI and have the generative AI select the optimal acquisition method.

[0096] The acquisition unit can filter audio data based on the user's current situation and environment. For example, the acquisition unit can analyze the user's current situation and environment using a generative AI. The generative AI can analyze the current situation and environment using a text generation AI (e.g., LLM). The acquisition unit can also analyze the current situation and environment using conventional natural language processing techniques. For example, if the user is out, the acquisition unit can filter out background noise to acquire audio data. If the user is at home, the acquisition unit can also acquire audio data while considering ambient noise. Furthermore, if the user is in a meeting, the acquisition unit can filter out ambient noise to acquire audio data. By filtering based on the current situation and environment, noise can be removed and the accuracy of audio data acquisition can be improved. Some or all of the above processing in the acquisition unit may be performed using a generative AI, or without one. For example, the acquisition unit can input data on the user's current situation and environment into the generative AI and have the generative AI perform the filtering.

[0097] The acquisition unit can estimate the user's emotions and determine the priority of audio data to acquire based on the estimated user emotions. The acquisition unit is implemented using an emotion estimation function, for example, with an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is excited, the acquisition unit will prioritize acquiring audio data to calm their emotions. The acquisition unit can also prioritize acquiring comforting audio data if the user is sad. Furthermore, if the user is happy, the acquisition unit can prioritize acquiring audio data to maintain that emotion. In this way, more appropriate audio data can be acquired by determining the priority of audio data based on the user's emotions. Some or all of the above processing in the acquisition unit may be performed using a generative AI, for example, or without a generative AI. For example, the acquisition unit can input user emotion data into a generative AI and have the generative AI perform the determination of the priority of audio data.

[0098] The acquisition unit can prioritize the acquisition of highly relevant data by considering the user's geographical location information when acquiring audio data. The acquisition unit can analyze the user's geographical location information using, for example, a generative AI. The generative AI can analyze the geographical location information using a text generation AI (e.g., LLM). The acquisition unit can also analyze the geographical location information using conventional natural language processing techniques. For example, if the user is in the Kansai region, the acquisition unit will prioritize the acquisition of audio data using the Kansai dialect. Similarly, if the user is in the Tohoku region, the acquisition unit can prioritize the acquisition of audio data using the Tohoku dialect. Furthermore, if the user is overseas, the acquisition unit can prioritize the acquisition of audio data that takes into account the language and dialect of that region. This allows for the priority acquisition of highly relevant audio data by considering geographical location information. Some or all of the above processing in the acquisition unit may be performed using, for example, a generative AI, or without using a generative AI. For example, the acquisition unit can input the user's geographical location information into the generative AI and have the generative AI acquire highly relevant data.

[0099] The acquisition unit can analyze the user's social media activity and obtain relevant data when acquiring audio data. For example, the acquisition unit can analyze the user's social media activity using a generative AI. The generative AI can analyze social media activity using a text generation AI (e.g., LLM). The acquisition unit can also analyze social media activity using conventional natural language processing techniques. For example, the acquisition unit can acquire audio data related to topics that the user has recently posted about frequently. The acquisition unit can also acquire audio data that reflects the opinions of influencers that the user follows. Furthermore, the acquisition unit can acquire audio data related to topics in online communities that the user participates in. In this way, relevant audio data can be acquired by analyzing social media activity. Some or all of the above processing in the acquisition unit may be performed using a generative AI, for example, or without a generative AI. For example, the acquisition unit can input data on the user's social media activity into a generative AI and have the generative AI acquire the relevant data.

[0100] The recognition unit can estimate the user's emotions and adjust the accuracy of speech recognition based on the estimated emotions. The recognition unit is implemented using emotion estimation functionality, for example, with an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. For example, if the user is nervous, the recognition unit can increase the accuracy of speech recognition to prevent misrecognition. The recognition unit can also slightly loosen the accuracy of speech recognition to maintain a natural flow of conversation if the user is relaxed. Furthermore, if the user is in a hurry, the recognition unit can maximize the accuracy of speech recognition to enable a quick response. In this way, more accurate speech recognition can be achieved by adjusting the accuracy of speech recognition based on the user's emotions. Some or all of the above processing in the recognition unit may be performed using, for example, a generative AI, or not using a generative AI. For example, the recognition unit can input user emotion data into a generative AI and have the generative AI perform the adjustment of speech recognition accuracy.

[0101] The recognition unit can optimize its recognition algorithm during speech recognition by considering the user's speech patterns and accent. For example, the recognition unit can analyze the user's speech patterns and accent using a generative AI. The generative AI can analyze speech patterns and accent using a text generation AI (e.g., LLM). The recognition unit can also analyze speech patterns and accent using conventional natural language processing techniques. For example, if the user speaks quickly, the recognition unit can adjust its recognition algorithm by considering the speech patterns. The recognition unit can also adjust its recognition algorithm by considering the speech patterns if the user speaks slowly. Furthermore, if the user has a specific accent, the recognition unit can optimize its recognition algorithm by considering that accent. In this way, by considering speech patterns and accents, the recognition algorithm can be optimized and the accuracy of speech recognition can be improved. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input data on the user's speech patterns and accent into a generative AI and have the generative AI perform the optimization of the recognition algorithm.

[0102] The recognition unit can improve recognition accuracy by referring to the user's past speech data during speech recognition. For example, the recognition unit can analyze the user's past speech data using a generative AI. The generative AI can analyze past speech data using a text generation AI (e.g., LLM). The recognition unit can also analyze past speech data using conventional natural language processing techniques. For example, the recognition unit can improve speech recognition accuracy based on phrases the user has used in the past. The recognition unit can also analyze the user's past speech data and optimize the recognition algorithm based on it. Furthermore, the recognition unit can detect changes in speech in specific situations from the user's past speech data and improve recognition accuracy based on that. In this way, recognition accuracy can be improved by referring to past speech data. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input the user's past speech data into a generative AI and have the generative AI perform the improvement of recognition accuracy.

[0103] The recognition unit can estimate the user's emotions and adjust the display method of the recognition results based on the estimated user emotions. The recognition unit is implemented using an emotion estimation function, for example, with an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is nervous, the recognition unit provides a simple and highly visible display method. The recognition unit can also provide a display method that includes detailed information if the user is relaxed. Furthermore, if the user is in a hurry, the recognition unit can provide a concise display method. In this way, a more appropriate display can be provided by adjusting the display method of the recognition results based on the user's emotions. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input user emotion data into a generative AI and have the generative AI perform the adjustment of the display method.

[0104] The recognition unit can improve recognition accuracy by considering the user's geographical location information during speech recognition. For example, the recognition unit can analyze the user's geographical location information using a generative AI. The generative AI can analyze the geographical location information using a text generation AI (e.g., LLM). Alternatively, the recognition unit can analyze the geographical location information using conventional natural language processing techniques. For example, if the user is in the Kansai region, the recognition unit can improve recognition accuracy by considering the Kansai dialect. Similarly, if the user is in the Tohoku region, the recognition unit can improve recognition accuracy by considering the Tohoku dialect. Furthermore, if the user is overseas, the recognition unit can improve recognition accuracy by considering the local language and dialect. In this way, recognition accuracy can be improved by considering geographical location information. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input the user's geographical location information into a generative AI and have the generative AI perform the improvement of recognition accuracy.

[0105] The recognition unit can improve recognition accuracy by analyzing the user's social media activity during speech recognition. For example, the recognition unit can analyze the user's social media activity using a generative AI. The generative AI can analyze social media activity using a text generation AI (e.g., LLM). The recognition unit can also analyze social media activity using conventional natural language processing techniques. For example, the recognition unit can use phrases related to topics the user has recently posted about to improve recognition accuracy. The recognition unit can also use phrases that reflect the opinions of influencers the user follows to improve recognition accuracy. Furthermore, the recognition unit can use phrases related to topics in online communities the user participates in to improve recognition accuracy. In this way, recognition accuracy can be improved by analyzing social media activity. Some or all of the above processing in the recognition unit may be performed using a generative AI, for example, or without a generative AI. For example, the recognition unit can input data on the user's social media activity into a generative AI and have the generative AI perform the improvement of recognition accuracy.

[0106] The detection unit can estimate the user's emotions and adjust the detection accuracy of important words based on the estimated user emotions. The detection unit is implemented using emotion estimation functionality, for example, with an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. For example, if the user is nervous, the detection unit can increase the detection accuracy of important words to prevent false positives. The detection unit can also slightly loosen the detection accuracy to maintain a natural flow of conversation if the user is relaxed. Furthermore, if the user is in a hurry, the detection unit can maximize the detection accuracy to enable a quick response. In this way, more accurate word detection can be achieved by adjusting the detection accuracy based on the user's emotions. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or not using a generative AI. For example, the detection unit can input user emotion data into a generative AI and have the generative AI perform the adjustment of detection accuracy.

[0107] The detection unit can optimize its detection algorithm by considering the user's speech patterns and accent when detecting important words. For example, the detection unit can analyze the user's speech patterns and accent using a generative AI. The generative AI can analyze speech patterns and accent using a text generation AI (e.g., LLM). Alternatively, the detection unit can analyze speech patterns and accent using conventional natural language processing techniques. For example, if the user speaks quickly, the detection unit adjusts its detection algorithm considering the speech patterns. Similarly, if the user speaks slowly, the detection unit can adjust its detection algorithm considering the speech patterns. Furthermore, if the user has a specific accent, the detection unit can optimize its detection algorithm by considering that accent. This allows for the optimization of the detection algorithm and improvement of word detection accuracy by considering speech patterns and accents. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or without a generative AI. For example, the detection unit can input data on the user's speech patterns and accent into a generative AI and have the generative AI optimize the detection algorithm.

[0108] The detection unit can improve detection accuracy by referring to the user's past speech data when detecting important words. The detection unit can analyze the user's past speech data using, for example, a generative AI. The generative AI can analyze past speech data using a text generation AI (e.g., LLM). The detection unit can also analyze past speech data using conventional natural language processing techniques. For example, the detection unit can improve the detection accuracy of important words based on phrases the user has used in the past. The detection unit can also analyze the user's past speech data and optimize the detection algorithm based on it. Furthermore, the detection unit can detect changes in speech in specific situations from the user's past speech data and improve detection accuracy based on that. In this way, detection accuracy can be improved by referring to past speech data. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or without a generative AI. For example, the detection unit can input the user's past speech data into a generative AI and have the generative AI perform the improvement of detection accuracy.

[0109] The detection unit can estimate the user's emotions and adjust the display method of the detection results based on the estimated user emotions. The detection unit is implemented using an emotion estimation function, for example, with an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is tense, the detection unit provides a simple and highly visible display method. The detection unit can also provide a display method that includes detailed information if the user is relaxed. Furthermore, if the user is in a hurry, the detection unit can provide a concise display method. In this way, by adjusting the display method of the detection results based on the user's emotions, a more appropriate display can be provided. Some or all of the above processing in the detection unit may be performed using a generative AI, for example, or without a generative AI. For example, the detection unit can input user emotion data into a generative AI and have the generative AI perform the adjustment of the display method.

[0110] The detection unit can improve detection accuracy by considering the user's geographical location information when detecting important words. The detection unit can analyze the user's geographical location information using, for example, a generative AI. The generative AI can analyze geographical location information using a text generation AI (e.g., LLM). Alternatively, the detection unit can analyze geographical location information using conventional natural language processing techniques. For example, if the user is in the Kansai region, the detection unit can improve detection accuracy by considering the Kansai dialect. Similarly, if the user is in the Tohoku region, the detection unit can improve detection accuracy by considering the Tohoku dialect. Furthermore, if the user is overseas, the detection unit can improve detection accuracy by considering the language and dialect of that region. In this way, detection accuracy can be improved by considering geographical location information. Some or all of the above processing in the detection unit may be performed using, for example, a generative AI, or without using a generative AI. For example, the detection unit can input the user's geographical location information into the generative AI and have the generative AI perform the improvement of detection accuracy.

[0111] The detection unit can improve detection accuracy by analyzing the user's social media activity when detecting important words. For example, the detection unit can analyze the user's social media activity using a generative AI. The generative AI can analyze social media activity using a text generation AI (e.g., LLM). The detection unit can also analyze social media activity using conventional natural language processing techniques. For example, the detection unit can use words related to topics that the user has recently posted about to improve detection accuracy. The detection unit can also use words that reflect the opinions of influencers that the user follows to improve detection accuracy. Furthermore, the detection unit can use words related to topics in online communities that the user participates in to improve detection accuracy. In this way, detection accuracy can be improved by analyzing social media activity. Some or all of the above processing in the detection unit may be performed using a generative AI, for example, or without a generative AI. For example, the detection unit can input data on the user's social media activity into a generative AI and have the generative AI perform the detection accuracy improvement.

[0112] The system according to the embodiment is not limited to the example described above, and various modifications are possible, for example, as follows.

[0113] A voice dialogue system may include a background sound generation unit that generates appropriate background sounds based on the user's utterances. The background sound generation unit may, for example, use a generation AI to generate background sounds suitable for the user's utterances. The generation AI may use a text generation AI (e.g., LLM) to generate background sounds. Alternatively, the background sound generation unit may use conventional speech synthesis technology to generate background sounds. For example, if the user is talking about nature, the background sound generation unit may generate sounds of birds chirping or wind. If the user is talking about a city, the background sound generation unit may also generate sounds of cars or people talking. Furthermore, if the user wants to relax, the background sound generation unit may also generate sounds of waves or quiet music. This allows for a more realistic conversation experience by generating appropriate background sounds based on the user's utterances. Some or all of the above-described processes in the background sound generation unit may be performed using, for example, a generation AI, or without a generation AI. For example, the background sound generation unit may input the user's utterances into the generation AI and have the generation AI perform background sound generation.

[0114] A voice dialogue system may include a gesture generation unit that generates appropriate gestures based on the user's utterances. The gesture generation unit generates gestures suitable for the user's utterances, for example, using a generation AI. The generation AI can generate gestures using a text generation AI (e.g., LLM). The gesture generation unit can also generate gestures using conventional motion capture technology. For example, when a user expresses gratitude, the gesture generation unit can generate a bowing or waving motion. When a user expresses surprise, the gesture generation unit can also generate an outstretched hand motion. Furthermore, when a user provides an explanation, the gesture generation unit can generate pointing or hand movements. This allows for a more natural conversational experience by generating appropriate gestures based on the user's utterances. Some or all of the above-described processes in the gesture generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the gesture generation unit can input the user's utterances into the generation AI and have the generation AI perform the gesture generation.

[0115] A voice dialogue system may include a visual effect generation unit that generates appropriate visual effects based on the user's utterances. The visual effect generation unit may, for example, use a generation AI to generate visual effects suitable for the user's utterances. The generation AI can generate visual effects using a text generation AI (e.g., LLM). The visual effect generation unit can also generate visual effects using conventional CG technology. For example, if the user is talking about fireworks, the visual effect generation unit can generate images of fireworks. If the user is talking about space, the visual effect generation unit can also generate images of a starry sky or planets. Furthermore, if the user wants to relax, the visual effect generation unit can generate images of a calm landscape. This allows for a more engaging conversation experience by generating appropriate visual effects based on the user's utterances. Some or all of the above-described processes in the visual effect generation unit may be performed using, for example, a generation AI, or without using a generation AI. For example, the visual effect generation unit can input the user's utterances into the generation AI and have the generation AI perform the generation of visual effects.

[0116] A voice dialogue system may include a music generation unit that generates appropriate music based on the user's utterances. The music generation unit may, for example, use a generation AI to generate music suitable for the user's utterances. The generation AI may use a text generation AI (e.g., LLM) to generate music. Alternatively, the music generation unit may use conventional music generation technologies to generate music. For example, if the user is telling a pleasant story, the music generation unit may generate bright and rhythmic music. If the user is telling a sad story, the music generation unit may generate quiet and sentimental music. Furthermore, if the user wants to relax, the music generation unit may generate calm and soothing music. This allows for a more emotionally rich conversational experience by generating appropriate music based on the user's utterances. Some or all of the above-described processes in the music generation unit may be performed using, for example, a generation AI, or without a generation AI. For example, the music generation unit may input the user's utterances into the generation AI and have the generation AI perform music generation.

[0117] A voice dialogue system may include a scent generation unit that generates an appropriate scent based on the user's utterances. The scent generation unit may, for example, use a generation AI to generate a scent suitable for the user's utterances. The generation AI may use a text generation AI (e.g., LLM) to generate scents. Alternatively, the scent generation unit may use conventional scent generation technologies to generate scents. For example, if the user is talking about flowers, the scent generation unit will generate a floral scent. If the user is talking about food, the scent generation unit may also generate a food scent. Furthermore, if the user wants to relax, the scent generation unit may generate lavender or mint scents. This allows for a more sensory-focused conversation experience by generating an appropriate scent based on the user's utterances. Some or all of the above-described processes in the scent generation unit may be performed using a generation AI, for example, or without a generation AI. For example, the scent generation unit may input the user's utterances into the generation AI and have the generation AI perform the scent generation.

[0118] A voice dialogue system may include a tempo adjustment unit that estimates the user's emotions and adjusts the pace of the conversation based on the estimated emotions. The tempo adjustment unit is implemented using emotion estimation functionality, for example, with an emotion engine or generative AI. The generative AI is, but is not limited to, text generation AI (e.g., LLM) or multimodal generation AI. For example, if the user is excited, the tempo adjustment unit can speed up the pace of the conversation to promote a lively dialogue. The tempo adjustment unit can also slow down the pace of the conversation to maintain a calm dialogue if the user is relaxed. Furthermore, if the user is in a hurry, the tempo adjustment unit can speed up the pace of the conversation to enable a quick response. In this way, a more natural and appropriate dialogue can be achieved by adjusting the pace of the conversation based on the user's emotions. Some or all of the above processing in the tempo adjustment unit may be performed using, for example, a generative AI, or not using a generative AI. For example, the tempo adjustment unit can input user emotion data into a generative AI and have the generative AI perform the adjustment of the conversation pace.

[0119] A voice dialogue system may include a topic selection unit that estimates the user's emotions and selects conversation topics based on the estimated emotions. The topic selection unit is implemented using an emotion estimation function, for example, by using an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is sad, the topic selection unit will select a comforting topic. If the user is happy, the topic selection unit may also select a topic that maintains that emotion. Furthermore, if the user is excited, the topic selection unit may also select a topic that soothes that excitement. By selecting conversation topics based on the user's emotions, a more appropriate dialogue can be achieved. Some or all of the above processing in the topic selection unit may be performed using a generative AI, for example, or without a generative AI. For example, the topic selection unit can input the user's emotion data into a generative AI and have the generative AI perform the selection of conversation topics.

[0120] A voice dialogue system may include a length adjustment unit that estimates the user's emotions and adjusts the length of the conversation based on the estimated emotions. The length adjustment unit is implemented using an emotion estimation function, for example, with an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is in a hurry, the length adjustment unit can shorten the conversation to provide a concise dialogue. The length adjustment unit can also lengthen the conversation to provide a more detailed dialogue if the user is relaxed. Furthermore, if the user is excited, the length adjustment unit can adjust the conversation length to alleviate that excitement. In this way, a more appropriate dialogue can be achieved by adjusting the conversation length based on the user's emotions. Some or all of the above processing in the length adjustment unit may be performed using a generative AI, for example, or without a generative AI. For example, the length adjustment unit can input the user's emotion data into a generative AI and have the generative AI perform the adjustment of the conversation length.

[0121] A voice dialogue system may include a tone adjustment unit that estimates the user's emotions and adjusts the tone of conversation based on the estimated emotions. The tone adjustment unit is implemented using an emotion estimation function, for example, with an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is angry, the tone adjustment unit will use a calm and composed tone. The tone adjustment unit may also use a comforting tone if the user is sad. Furthermore, the tone adjustment unit may use a bright and cheerful tone if the user is happy. By adjusting the tone of conversation based on the user's emotions, a more appropriate dialogue can be achieved. Some or all of the above processing in the tone adjustment unit may be performed using, for example, a generative AI, or without a generative AI. For example, the tone adjustment unit can input the user's emotion data into a generative AI and have the generative AI perform the tone adjustment of the conversation.

[0122] A voice dialogue system may include a content customization unit that estimates the user's emotions and customizes the conversation content based on the estimated emotions. The content customization unit is implemented using an emotion estimation function, for example, with an emotion engine or a generative AI. The generative AI is, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. For example, if the user is nervous, the content customization unit may provide content that helps them relax. Also, if the user is relaxed, the content customization unit may provide interesting topics. Furthermore, if the user is in a hurry, the content customization unit may provide concise content. By customizing the conversation content based on the user's emotions, a more appropriate dialogue can be achieved. Some or all of the above processing in the content customization unit may be performed using, for example, a generative AI, or without a generative AI. For example, the content customization unit may input the user's emotion data into a generative AI and have the generative AI perform the customization of the conversation content.

[0123] The following briefly describes the processing flow for example form 2.

[0124] Step 1: The prediction unit predicts the end and final sentence of the conversation. The prediction unit can predict the end and final sentence of the conversation using generative AI. The generative AI uses text generation AI (e.g., LLM) or multimodal generative AI to understand the context of the conversation and predict the next sentence. Step 2: The generation unit generates a response based on the ending sentence predicted by the prediction unit. The generation unit can generate a response using a generation AI. The generation AI uses a text generation AI (e.g., LLM) or a multimodal generation AI to generate an appropriate response based on the predicted ending sentence. Step 3: The provider unit provides the response generated by the generator unit. The provider unit can provide the response using a generation AI, or it can provide the response without using a generation AI. For example, the provider unit can play the generated response as audio or display it as text.

[0125] 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.

[0126] Data generation model 58 is a form of 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> Examples of generative AI include text generation AI, image generation AI, and multimodal generation AI. 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 (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats from audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVMs), k-means clustering, convolutional neural networks (CNNs), recurrent neural networks (RNNs), generative adversarial networks (GANs), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each of the above parts is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example.Furthermore, processing performed by AI, including generative AI, may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by AI, including generative AI.

[0127] Furthermore, the processing performed by the data processing system 10 described above is carried out by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but it may also be carried out by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart device 14 or an external device, and the smart device 14 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0128] Each of the multiple elements described above, including the prediction unit, generation unit, provision unit, acquisition unit, recognition unit, and detection unit, is implemented, for example, in at least one of the smart device 14 and the data processing device 12. For example, the prediction unit is implemented by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. The generation unit is implemented, for example, by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. The provision unit is implemented, for example, by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. The acquisition unit is implemented, for example, by the microphone 38B of the smart device 14 or the specific processing unit 290 of the data processing device 12. The recognition unit is implemented, for example, by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. The detection unit is implemented, for example, by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. The correspondence between each part and the device or control unit is not limited to the examples described above, and various modifications are possible.

[0129] [Second Embodiment] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0130] 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.

[0131] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0132] 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.

[0133] The microphone 238 receives voice commands and other instructions from the user by receiving voice signals. The microphone 238 captures the voice signals from the user, 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.

[0134] 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, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0135] 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.

[0136] 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 by the processor 28. The storage 32 stores the specific processing program 56.

[0137] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0138] 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. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0139] In the smart glasses 214, specific processing is performed by the processor 46. The storage 50 stores a specific processing program 60. The processor 46 reads the specific processing program 60 from the storage 50 and executes the read specific processing program 60 on the RAM 48. The specific processing is realized by the processor 46 acting as a control unit 46A according to the specific processing program 60 executed on the RAM 48. The smart glasses 214 also have a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0140] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0141] 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.

[0142] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. 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 inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0143] The data processing system 210 according to the second embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 210 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but it may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart glasses 214 or an external device, and the smart glasses 214 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0144] Each of the multiple elements described above, including the prediction unit, generation unit, provision unit, acquisition unit, recognition unit, and detection unit, is implemented, for example, in at least one of the smart glasses 214 and the data processing device 12. For example, the prediction unit is implemented by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. The generation unit is implemented, for example, by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. The provision unit is implemented, for example, by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. The acquisition unit is implemented, for example, by the microphone 238 of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. The recognition unit is implemented, for example, by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. The detection unit is implemented, for example, by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. The correspondence between each part and the device or control unit is not limited to the examples described above, and various modifications are possible.

[0145] [Third Embodiment] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0146] 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.

[0147] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0148] 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.

[0149] The microphone 238 receives voice commands and other instructions from the user by receiving voice signals. The microphone 238 captures the voice signals from the user, 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.

[0150] 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, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0151] 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.

[0152] 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.

[0153] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0154] 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. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0155] In the headset terminal 314, specific processing is performed by the processor 46. The storage 50 stores a specific program 60. The processor 46 reads the specific program 60 from the storage 50 and executes the read specific program 60 on the RAM 48. The specific processing is realized by the processor 46 acting as a control unit 46A according to the specific program 60 executed on the RAM 48. The headset terminal 314 also has a data generation model 58 and an emotion identification model 59, similar to the data generation model and emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0156] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0157] 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.

[0158] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. 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 inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0159] The data processing system 310 according to the third embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 310 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the headset terminal 314, but may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the headset terminal 314. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the headset terminal 314 or an external device, and the headset terminal 314 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0160] Each of the multiple elements described above, including the prediction unit, generation unit, provision unit, acquisition unit, recognition unit, and detection unit, is implemented, for example, in at least one of the headset terminal 314 and the data processing device 12. For example, the prediction unit is implemented by the control unit 46A of the headset terminal 314 or the specific processing unit 290 of the data processing device 12. The generation unit is implemented, for example, by the control unit 46A of the headset terminal 314 or the specific processing unit 290 of the data processing device 12. The provision unit is implemented, for example, by the control unit 46A of the headset terminal 314 or the specific processing unit 290 of the data processing device 12. The acquisition unit is implemented, for example, by the microphone 238 of the headset terminal 314 or the specific processing unit 290 of the data processing device 12. The recognition unit is implemented, for example, by the control unit 46A of the headset terminal 314 or the specific processing unit 290 of the data processing device 12. The detection unit is implemented, for example, by the control unit 46A of the headset terminal 314 or the specific processing unit 290 of the data processing device 12. The correspondence between each part and the device or control unit is not limited to the examples described above, and various modifications are possible.

[0161] [Fourth Embodiment] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0162] 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.

[0163] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. 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 and / or LAN.

[0164] 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.

[0165] The microphone 238 receives voice commands and other instructions from the user by receiving voice signals. The microphone 238 captures the voice signals from the user, 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.

[0166] 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 image sensor or CCD image sensor, which captures images of the area around the user (for example, an imaging range defined by a field of view equivalent to the field of vision of a typical healthy person).

[0167] 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.

[0168] 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. The robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0169] 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.

[0170] The processor 28 reads a 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 acting as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0171] 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. The identification processing unit 290 can estimate the user's emotions using the emotion identification model 59 and perform identification processing using the user's emotions. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotions, including but not limited to these examples. Furthermore, emotion estimation and prediction also include, for example, emotion analysis.

[0172] In robot 414, specific processing is performed by processor 46. A specific program 60 is stored in storage 50. Processor 46 reads the specific program 60 from storage 50 and executes it on RAM 48. The specific processing is achieved by processor 46 acting as a control unit 46A according to the specific program 60 executed on RAM 48. Robot 414 also has data generation model 58 and emotion identification model 59, similar to those of the robot, and can perform processing similar to that of the specific processing unit 290 using these models.

[0173] Furthermore, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 obtains processing results (such as prediction results) using the data generation model 58 by communicating with the server device that has the data generation model 58. Also, the data processing device 12 may be a server device or a terminal device owned by the user (for example, a mobile phone, robot, home appliance, etc.).

[0174] 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.

[0175] The data generation model 58 is a so-called generative AI. An example of a data generation model 58 is a generative AI such as ChatGPT. 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 inference data such as audio data representing speech, text data representing text, and image data representing images (e.g., still image data or video data). The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference result in one or more data formats such as audio data, text data, and image data. The data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specific processing unit 290 performs the specific processing described above using the data generation model 58. The data generation model 58 may be a fine-tuned model that outputs inference results from prompts that do not contain instructions, in which case the data generation model 58 can output inference results from prompts that do not contain instructions. In the data processing device 12, etc., there are multiple types of data generation models 58, and the data generation model 58 includes AI other than generative AI. AI other than generative AI includes, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. Also, the AI ​​may be an AI agent. Furthermore, when the processing of each part described above is performed by the AI, the processing may be performed by the AI ​​in part or in whole, but is not limited to this example. Also, processing performed by an AI including a generative AI may be replaced by rule-based processing, and rule-based processing may be replaced by processing performed by an AI including a generative AI.

[0176] The data processing system 410 according to the fourth embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 410 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but it may also be performed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. In addition, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the robot 414 or an external device, and the robot 414 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0177] Each of the multiple elements described above, including the prediction unit, generation unit, provision unit, acquisition unit, recognition unit, and detection unit, is implemented, for example, in at least one of the robot 414 and the data processing device 12. For example, the prediction unit is implemented by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. The generation unit is implemented, for example, by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. The provision unit is implemented, for example, by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. The acquisition unit is implemented, for example, by the microphone 238 of the robot 414 or the specific processing unit 290 of the data processing device 12. The recognition unit is implemented, for example, by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. The detection unit is implemented, for example, by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. The correspondence between each part and the device or control unit is not limited to the examples described above, and various modifications are possible.

[0178] 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.

[0179] Figure 9 shows the 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.

[0180] 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.

[0181] 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.

[0182] 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, and motorcycles, 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 based, for example, 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.

[0183] 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."

[0184] 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.

[0185] 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 method for the specific process may be used, which includes computer 22 and multiple other computers.

[0186] 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.

[0187] 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.

[0188] 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.

[0189] 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.

[0190] 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.

[0191] 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.

[0192] 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.

[0193] Furthermore, although the above-described examples were divided into four embodiments, some or all of these embodiments may be combined. Also, the smart device 14, smart glasses 214, headset terminal 314, and robot 414 are just examples, and they may be combined, or other devices may be used. Also, although the above-described examples were divided into two embodiments, Embodiment 1 and Embodiment 2, these may be combined.

[0194] 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 other things 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.

[0195] 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.

[0196] (Note 1) A prediction unit that predicts the end of a conversation and the final sentence, A generation unit that generates a response based on the ending sentence predicted by the prediction unit, A providing unit that provides the response generated by the generation unit, Equipped with A system characterized by the following features. (Note 2) It includes an acquisition unit for acquiring audio data. The system described in Appendix 1, characterized by the features described herein. (Note 3) It includes a recognition unit that performs speech recognition processing. The system described in Appendix 1, characterized by the features described herein. (Note 4) It is equipped with a detection unit that detects important words. The system described in Appendix 1, characterized by the features described herein. (Note 5) The prediction unit, It estimates the user's emotions and adjusts the accuracy of predicting the end of the conversation and the concluding sentence based on the estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 6) The prediction unit, By analyzing past conversation history, the predictive algorithm is optimized based on the user's speaking patterns. The system described in Appendix 1, characterized by the features described herein. (Note 7) The prediction unit, The prediction timing is adjusted considering the user's speaking speed and pauses. The system described in Appendix 1, characterized by the features described herein. (Note 8) The prediction unit, It estimates the user's sentiment and determines the priority of predicted closing sentences based on the estimated user sentiment. The system described in Appendix 1, characterized by the features described herein. (Note 9) The prediction unit, The prediction system takes into account the user's geographical location and incorporates regional expressions and dialects. The system described in Appendix 1, characterized by the features described herein. (Note 10) The prediction unit, Analyze users' social media activity and incorporate recent trends and topics into predictions. The system described in Appendix 1, characterized by the features described herein. (Note 11) The generating unit is It estimates the user's emotions and adjusts the way responses are expressed based on those estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 12) The generating unit is Adjust the level of detail in the response based on the predicted importance of the closing sentence. The system described in Appendix 1, characterized by the features described herein. (Note 13) The generating unit is When generating a response, the system refers to the user's past response patterns to generate the most suitable response. The system described in Appendix 1, characterized by the features described herein. (Note 14) The generating unit is It estimates the user's emotions and adjusts the length of the response based on the estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 15) The generating unit is When generating a response, customize the response content based on the user's current situation and environment. The system described in Appendix 1, characterized by the features described herein. (Note 16) The generating unit is When generating responses, relevant information is added based on the user's interests and preferences. The system described in Appendix 1, characterized by the features described herein. (Note 17) The aforementioned supply unit is, It estimates the user's emotions and adjusts how responses are provided based on those estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 18) The aforementioned supply unit is, When providing a response, the system will refer to the user's past responses to select the optimal timing for providing the response. The system described in Appendix 1, characterized by the features described herein. (Note 19) The aforementioned supply unit is, When providing a response, the optimal method of delivery will be selected, taking into account the user's device information. The system described in Appendix 1, characterized by the features described herein. (Note 20) The aforementioned supply unit is, It estimates the user's emotions and adjusts how responses are displayed based on those estimated emotions. The system described in Appendix 1, characterized by the features described herein. (Note 21) The aforementioned supply unit is, When providing a response, the optimal method of delivery will be selected, taking into account the user's geographical location. The system described in Appendix 1, characterized by the features described herein. (Note 22) The aforementioned supply unit is, When providing a response, we analyze the user's social media activity and provide relevant information. The system described in Appendix 1, characterized by the features described herein. (Note 23) The acquisition unit is, The system estimates the user's emotions and adjusts the timing of voice data acquisition based on the estimated emotions. The system described in Appendix 2, characterized by the features described herein. (Note 24) The acquisition unit is, Analyze the user's past speech history and select the optimal acquisition method. The system described in Appendix 2, characterized by the features described herein. (Note 25) The acquisition unit is, When acquiring audio data, filtering is performed based on the user's current situation and environment. The system described in Appendix 2, characterized by the features described herein. (Note 26) The acquisition unit is, It estimates the user's emotions and determines the priority of audio data to acquire based on the estimated user emotions. The system described in Appendix 2, characterized by the features described herein. (Note 27) The acquisition unit is, When acquiring audio data, the system prioritizes acquiring data that is highly relevant, taking into account the user's geographical location. The system described in Appendix 2, characterized by the features described herein. (Note 28) The acquisition unit is, When acquiring voice data, the system analyzes the user's social media activity and retrieves relevant data. The system described in Appendix 2, characterized by the features described herein. (Note 29) The aforementioned recognition unit, It estimates the user's emotions and adjusts the accuracy of speech recognition based on the estimated emotions. The system described in Appendix 3, characterized by the features described herein. (Note 30) The aforementioned recognition unit, During speech recognition, the recognition algorithm is optimized by taking into account the user's speech patterns and accent. The system described in Appendix 3, characterized by the features described herein. (Note 31) The aforementioned recognition unit, During speech recognition, the system improves recognition accuracy by referencing the user's past speech data. The system described in Appendix 3, characterized by the features described herein. (Note 32) The aforementioned recognition unit, It estimates the user's emotions and adjusts how the recognition results are displayed based on the estimated emotions. The system described in Appendix 3, characterized by the features described herein. (Note 33) The aforementioned recognition unit, When performing speech recognition, the accuracy of recognition is improved by taking into account the user's geographical location. The system described in Appendix 3, characterized by the features described herein. (Note 34) The aforementioned recognition unit, During speech recognition, the system analyzes the user's social media activity to improve recognition accuracy. The system described in Appendix 3, characterized by the features described herein. (Note 35) The detection unit, It estimates the user's emotions and adjusts the accuracy of detecting important words based on the estimated user emotions. The system described in Appendix 4, characterized by the features described herein. (Note 36) The detection unit, When detecting important words, the detection algorithm is optimized by taking into account the user's speech patterns and accent. The system described in Appendix 4, characterized by the features described herein. (Note 37) The detection unit, When detecting important words, the system improves detection accuracy by referencing the user's past speech data. The system described in Appendix 4, characterized by the features described herein. (Note 38) The detection unit, It estimates the user's emotions and adjusts how the detection results are displayed based on the estimated emotions. The system described in Appendix 4, characterized by the features described herein. (Note 39) The detection unit, When detecting important words, the system improves detection accuracy by taking into account the user's geographical location. The system described in Appendix 4, characterized by the features described herein. (Note 40) The detection unit, When detecting important words, we analyze the user's social media activity to improve detection accuracy. The system described in Appendix 4, characterized by the features described herein. [Explanation of symbols]

[0197] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots

Claims

1. A prediction unit that predicts the end of a conversation and the final sentence, A generation unit that generates a response based on the ending sentence predicted by the prediction unit, A providing unit that provides the response generated by the generation unit, Equipped with A system characterized by the following features.

2. It includes an acquisition unit for acquiring audio data. The system according to feature 1.

3. It includes a recognition unit that performs speech recognition processing. The system according to feature 1.

4. It is equipped with a detection unit that detects important words. The system according to feature 1.

5. The prediction unit, It estimates the user's emotions and adjusts the accuracy of predicting the end of the conversation and the concluding sentence based on the estimated emotions. The system according to feature 1.

6. The prediction unit, By analyzing past conversation history, the predictive algorithm is optimized based on the user's speaking patterns. The system according to feature 1.

7. The prediction unit, The prediction timing is adjusted considering the user's speaking speed and pauses. The system according to feature 1.

8. The prediction unit, It estimates the user's sentiment and determines the priority of predicted closing sentences based on the estimated user sentiment. The system according to feature 1.

9. The prediction unit, The prediction system takes into account the user's geographical location and incorporates regional expressions and dialects. The system according to feature 1.

10. The prediction unit, Analyze users' social media activity and incorporate recent trends and topics into predictions. The system according to feature 1.

Images