system

Abstract

The system according to the embodiment aims to prevent users from using offensive words and to replace them with appropriate words. [Solution] A system according to an embodiment includes a reception unit, an analysis unit, a generation unit, a presentation unit, and a learning unit. The reception unit receives text input from a user. The analysis unit analyzes the text received by the reception unit and detects offensive words. The generation unit generates candidates for replacing the offensive words detected by the analysis unit with appropriate words. The presentation unit presents the candidates generated by the generation unit to the user. The learning unit learns words selected by the user from the candidates presented by the presentation unit.

Description

[Technical Field]

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

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventional techniques do not adequately provide means to prevent users from using offensive language, and there is room for improvement.

[0005] The system according to the embodiment aims to prevent users from using offensive words and to replace them with appropriate words. [Means for solving the problem]

[0006] The system according to the embodiment includes a reception unit, an analysis unit, a generation unit, a presentation unit, and a learning unit. The reception unit receives text input from a user. The analysis unit analyzes the text received by the reception unit and detects offensive words. The generation unit generates candidates for replacing the offensive words detected by the analysis unit with appropriate words. The presentation unit presents the candidates generated by the generation unit to the user. The learning unit learns words selected by the user from the candidates presented by the presentation unit. [Effects of the Invention]

[0007] The system according to the embodiment can prevent users from using offensive words and replace them with appropriate words. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10]1 shows an emotion map onto which multiple emotions are mapped. DETAILED DESCRIPTION OF THE INVENTION

[0009] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0010] First, the terms used in the following description will be explained.

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

[0012] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0013] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0014] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), and Bluetooth (registered trademark).

[0015] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

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

[0017] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[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, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[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. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0020] The reception device 38 includes a touch panel 38A and a microphone 38B, and receives user input. The touch panel 38A detects contact with a pointer (for example, a pen or a finger) to receive user input by the touch of the pointer. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 (see FIG. 2) acquires the data indicating the user input.

[0021] Output device 40 includes a display 40A and a speaker 40B, and presents data to a user by outputting the data in a form of expression that the user can perceive (e.g., audio and / or text). Display 40A displays visible information such as text and images in accordance with instructions from processor 46. Speaker 40B outputs audio in accordance with instructions from processor 46. Camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0022] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

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

[0024] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0025] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate a user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotion, including estimation and prediction of the user's emotion, but is not limited to these examples. Furthermore, the estimation and prediction of emotion also includes, for example, emotion analysis.

[0026] In the smart device 14, the specific processing is performed by the processor 46. The storage 50 stores a specific processing program 60. The specific processing program 60 is used together 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 the control unit 46A in accordance with the specific processing program 60 executed on the RAM 48. Note that the smart device 14 has a data generation model and an emotion identification model similar to the data generation model 58 and the emotion identification model 59, and can also perform processing similar to that of the specific processing unit 290 using these models.

[0027] Note that a device other than the data processing device 12 may 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 a processing result (prediction result, etc.) using the data generation model 58 by communicating with the server device having the data generation model 58. Furthermore, the data processing device 12 may be a server device, or may be a terminal device owned by a user (e.g., a mobile phone, a robot, a home appliance, etc.). Next, an example of processing by the data processing system 10 according to the first embodiment will be described.

[0028] (Example 1) A system for paraphrasing offensive words according to an embodiment of the present invention analyzes text entered by a user and detects offensive words and expressions. The system generates candidates for replacing the detected offensive words with appropriate words and presents them to the user. The user can select an appropriate word from the presented candidates. For example, the word "stupid" can be replaced with an expression such as "foolish." The system analyzes text using natural language processing technology to generate paraphrase candidates. Furthermore, the system has a function for learning user selections and reflecting them in subsequent candidate generation. The system also has a function for generating paraphrase candidates taking context into account, enabling more natural paraphrasing. For example, if the text entered by a user contains the word "stupid," the system detects the word and generates appropriate candidate words such as "stupid" or "ignorant." The user can select from these candidates to correct the text. The system saves the user's selection history and reflects it in subsequent candidate generation, thereby providing more accurate paraphrase candidates. Furthermore, because paraphrase candidates are generated taking context into account, natural paraphrasing is possible without compromising the overall meaning of the text. For example, the system analyzes the context of the text entered by the user and generates appropriate paraphrase candidates, thereby allowing the user to enjoy a comfortable text entry environment. Thus, the system for paraphrasing offensive words into other words analyzes the text entered by the user and replaces the offensive words with appropriate words, thereby providing the user with a comfortable text entry environment.

[0029] A system for paraphrasing offensive words according to an embodiment includes a reception unit, an analysis unit, a generation unit, a presentation unit, and a learning unit. The reception unit receives text input from a user. The text input from the user includes, but is not limited to, handwritten text and digital text. For example, the reception unit digitizes and reads the handwritten text using scanning technology. The reception unit can also directly read text submitted in digital format. The reception unit can also read printed text using OCR technology. For example, the reception unit scans handwritten text with a high-resolution scanner and converts it into character information using OCR technology. Digital text submitted in a specific file format can also be directly read. OCR technology recognizes printed characters with high accuracy and converts them into digital text. The analysis unit analyzes the text received by the reception unit to detect offensive words. The analysis is performed using, for example, natural language processing technology, but is not limited to, for example. For example, the analysis unit analyzes words and phrases in the text to detect offensive words. The analysis unit can also detect offensive words by taking context into consideration. The generation unit generates candidates for replacing the offensive words detected by the analysis unit with appropriate words. The generation is performed, for example, using natural language processing technology, but is not limited to such an example. For example, the generation unit generates appropriate word candidates using a text generation AI (e.g., LLM). The generation unit can also generate appropriate word candidates using a multimodal generation AI. The generation unit can also generate appropriate word candidates by taking context into consideration using the generation AI. For example, the text generation AI has learned a large amount of text data and has advanced natural language processing capabilities. The multimodal generation AI can handle multiple modalities, including not only text but also images and audio. The generation AI uses keyword extraction technology to pick out particularly important information in the text and generate appropriate word candidates based on that information. The presentation unit presents the candidates generated by the generation unit to the user. The presentation is performed, for example, using a user interface, but is not limited to such an example.For example, the presentation unit presents the generated candidates to the user in a list format. The presentation unit can also present the generated candidates to the user in a card format. The presentation unit can also present the generated candidates to the user in a pop-up format. The learning unit learns words selected by the user from the candidates presented by the presentation unit. Learning is performed, for example, using machine learning technology, but is not limited to this example. For example, the learning unit saves the user's selection history and reflects it in generating candidates from the next time onwards. The learning unit can also receive user feedback and use it to improve the system. As a result, the system for paraphrasing offensive words according to the embodiment can provide a comfortable text input environment for the user by analyzing the user's text input and replacing the offensive words with appropriate words.

[0030] The context analysis unit can generate paraphrase candidates based on the context. For example, when the generative AI is summarizing, the context analysis unit can automatically collect relevant background information and refer to it to understand the context. For example, the context analysis unit collects related news articles and academic papers. The context analysis unit can also use a topic model to understand the context when the generative AI is summarizing. For example, the context analysis unit extracts related keywords and phrases based on the topic model. The topic model can be realized using techniques such as LDA (Latent Dirichlet Allocation), non-negative matrix factorization, or topic clustering. The context analysis unit can also build a system for understanding the context by referring to relevant background information and topic models when the generative AI is summarizing. For example, the context analysis unit automatically collects relevant information and reflects it in the summary. This allows for more accurate summarization by referring to background information and topic models to understand the context. Furthermore, the context analysis unit can collect relevant information in real time to understand the context and reflect it in the summary. For example, the context analysis unit collects the latest news articles and social media posts from the Internet and uses them as information sources for understanding the context. The context analysis unit can also use topic models to create a feedback loop to improve the accuracy of summaries. For example, the context analysis unit evaluates the quality of the generated summaries and updates the topic model based on the evaluation results. This allows the context analysis unit to always understand the context based on the latest information and provide more accurate summaries. This allows for more natural paraphrasing by taking context into consideration.

[0031] The history storage unit can store the user's selection history. The history storage unit stores, for example, words selected by the user and the date and time of selection. For example, the history storage unit stores the words selected by the user in a database. The history storage unit can also store the date and time of the user's selection as a timestamp. The history storage unit can also analyze the user's selection history and reflect it in subsequent candidate generation. For example, the history storage unit analyzes patterns of words selected by the user in the past and prioritizes generating similar words as candidates. The history storage unit can also generate word candidates that are optimal for a specific time period based on the user's selection history. In this way, by storing the user's selection history, it can be reflected in subsequent candidate generation.

[0032] The feedback receiving unit can receive user feedback. The feedback receiving unit can receive, for example, user opinions and evaluations. For example, the feedback receiving unit can receive text feedback entered by the user. The feedback receiving unit can also receive feedback provided by the user via voice. The feedback receiving unit can also analyze the user feedback and use it to improve the system. For example, the feedback receiving unit can store the user feedback content in a database and identify areas for improvement in the system. The feedback receiving unit can also develop an algorithm for improving the system's functions based on the user feedback. In this way, receiving user feedback can be used to improve the system.

[0033] The generator can generate paraphrase candidates based on context. For example, when the generation AI generates a summary, the generator can automatically collect relevant background information and refer to it to understand the context. For example, the generator can collect related news articles and academic papers. The generator can also use a topic model to understand the context when the generation AI generates a summary. For example, the generator can extract related keywords and phrases based on the topic model. The topic model can be realized using techniques such as LDA (Latent Dirichlet Allocation), non-negative matrix factorization, or topic clustering. The generator can also build a system for understanding the context by referring to relevant background information and topic models when the generation AI generates a summary. For example, the generator can automatically collect relevant information and incorporate it into the summary. This enables more accurate summaries by referring to background information and topic models to understand the context. Furthermore, the generator can collect relevant information in real time to understand the context and incorporate it into the summary. For example, the generator can collect the latest news articles and social media posts from the Internet and use them as sources of information for understanding the context. The generator can also use the topic model to create a feedback loop to improve the accuracy of summaries. For example, the generator evaluates the quality of the generated summaries and updates the topic model based on the evaluation results. This allows the generator to always understand the context based on the latest information and provide more accurate summaries. This allows for more natural paraphrasing by taking context into account.

[0034] The learning unit can save the user's selection history and reflect it in subsequent candidate generation. The learning unit, for example, saves words selected by the user and the date and time of selection. For example, the learning unit saves the words selected by the user in a database. The learning unit can also save the date and time of the user's selection as a timestamp. The learning unit can also analyze the user's selection history and reflect it in subsequent candidate generation. For example, the learning unit analyzes patterns of words selected by the user in the past and prioritizes generating similar words as candidates. The learning unit can also generate candidate words that are optimal for a specific time period based on the user's selection history. In this way, by reflecting the user's selection history, the accuracy of candidate generation can be improved in subsequent candidate generation.

[0035] The reception unit can analyze the user's past text input history and select the optimal reception method. For example, the reception unit preferentially suggests input methods (voice input, text input, etc.) that the user has frequently used in the past. The reception unit can also predict and suggest the optimal input method for a specific time period based on the user's past input history. The reception unit can also analyze the content of text that the user has previously input and provide an appropriate input interface. In this way, the optimal reception method can be provided by analyzing the user's past input history.

[0036] The reception unit can perform filtering based on the user's current situation and areas of interest when inputting text. For example, the reception unit preferentially receives topics and keywords related to the user's current situation. The reception unit can also preferentially receive related text based on the user's areas of interest. The reception unit can also filter unnecessary text based on the user's current situation and areas of interest. In this way, filtering based on the user's current situation and areas of interest can preferentially receive highly relevant text.

[0037] When inputting text, the reception unit can preferentially receive highly relevant text by taking into account the user's geographical location information. For example, when the user is in a specific area, the reception unit can preferentially receive text related to that area. Furthermore, when the user is traveling, the reception unit can preferentially receive text related to the travel destination. Furthermore, when the user is at home, the reception unit can preferentially receive text related to information about the area around the user's home. In this way, highly relevant text can be preferentially received by taking into account the user's geographical location information.

[0038] The reception unit can analyze the user's social media activity and receive related text when the user inputs text. The reception unit can, for example, preferentially receive related text based on keywords frequently used by the user on social media. The reception unit can also analyze the user's social media activity history and preferentially receive related topics. The reception unit can also preferentially receive related text based on the content of posts from accounts the user follows on social media. In this way, by analyzing the user's social media activity, it is possible to preferentially receive related text.

[0039] When analyzing text, the analysis unit can adjust the analysis algorithm by referring to past analysis data. For example, the analysis unit can apply a more accurate offensive language detection algorithm based on the past analysis data. The analysis unit can also analyze the past analysis data and adjust the analysis algorithm based on specific patterns. The analysis unit can also improve the performance of the analysis algorithm by referring to the past analysis data. In this way, the accuracy of the analysis algorithm is improved by referring to the past analysis data.

[0040] The analysis unit can perform text analysis while taking into account user attribute information. The analysis unit selects an appropriate analysis algorithm based on the user's age and gender, for example. The analysis unit can also preferentially detect relevant offensive words based on the user's occupation and interests. The analysis unit can also apply an appropriate analysis algorithm while taking into account the user's cultural background. This allows for more appropriate analysis by taking into account the user's attribute information.

[0041] The analysis unit can perform text analysis while taking into account the geographic distribution of the text. For example, the analysis unit can preferentially detect offensive words that are frequently used in a specific region. The analysis unit can also analyze the geographic distribution of the text to understand the trends in offensive words by region. The analysis unit can also apply an appropriate analysis algorithm based on the geographic distribution. In this way, by taking the geographic distribution of the text into account, it is possible to understand the trends in offensive words by region.

[0042] The analysis unit can improve the accuracy of the analysis by referring to related literature during text analysis. For example, the analysis unit can refer to related literature and optimize the analysis algorithm based on the latest research results. The analysis unit can also analyze related literature and adjust the analysis algorithm based on specific patterns. The analysis unit can also introduce new techniques for improving the accuracy of the analysis by referring to related literature. In this way, the accuracy of the analysis is improved by referring to related literature.

[0043] When generating paraphrase candidates, the generation unit can adjust the level of detail of the candidates based on the importance of the text. For example, in the case of important text, the generation unit generates detailed paraphrase candidates. In addition, in the case of text with low importance, the generation unit can also generate simple paraphrase candidates. In addition, the generation unit can also generate paraphrase candidates with an appropriate level of detail according to the importance of the text. In this way, paraphrase candidates with an appropriate level of detail can be generated according to the importance of the text.

[0044] When generating paraphrase candidates, the generation unit can apply different generation algorithms depending on the text category. For example, in the case of a business document, the generation unit generates formal paraphrase candidates. In addition, in the case of casual text, the generation unit can also generate familiar paraphrase candidates. In addition, the generation unit can apply an appropriate generation algorithm depending on the text category. In this way, by applying an appropriate generation algorithm depending on the text category, more appropriate paraphrase candidates can be provided.

[0045] When generating paraphrase candidates, the generation unit can determine the priority of the candidates based on the time of submission of the text. For example, in the case of recently submitted text, the generation unit can generate paraphrase candidates with priority. In addition, in the case of text that was submitted earlier, the generation unit can also generate paraphrase candidates later. In addition, the generation unit can generate paraphrase candidates with an appropriate priority depending on the time of submission of the text. In this way, paraphrase candidates can be generated with an appropriate priority depending on the time of submission of the text.

[0046] When generating paraphrase candidates, the generation unit can adjust the order of the candidates based on the relevance of the text. For example, the generation unit preferentially displays paraphrase candidates that are highly relevant to the content of the text. The generation unit can also display paraphrase candidates that are less relevant later. The generation unit can also display paraphrase candidates in an appropriate order based on the relevance of the text. This makes it possible to display paraphrase candidates in an appropriate order based on the relevance of the text.

[0047] When presenting candidates, the presentation unit can select the optimal presentation method by referring to the user's past selection history. For example, the presentation unit analyzes patterns of paraphrase candidates previously selected by the user and preferentially presents similar candidates. The presentation unit can also preferentially provide a specific presentation method (list format, card format, etc.) based on the user's past selection history. The presentation unit can also determine the optimal presentation order based on the user's past selection history. In this way, the optimal presentation method can be provided by referring to the user's past selection history.

[0048] The presentation unit can present candidates while taking into consideration the user's attribute information. The presentation unit selects an appropriate presentation method based on, for example, the user's age and gender. The presentation unit can also preferentially present related paraphrase candidates based on the user's occupation and interests. The presentation unit can also provide an appropriate presentation method by taking into consideration the user's cultural background. This allows for more appropriate presentation by taking into consideration the user's attribute information.

[0049] When presenting candidates, the presentation unit can select the optimal presentation method by taking into consideration the user's geographical location information. For example, when the user is in a specific area, the presentation unit can preferentially present paraphrase candidates related to that area. Furthermore, when the user is traveling, the presentation unit can also preferentially present paraphrase candidates related to the user's travel destination. Furthermore, when the user is at home, the presentation unit can also preferentially present paraphrase candidates related to information about the area around the user's home. In this way, by taking into consideration the user's geographical location information, highly relevant candidates can be preferentially presented.

[0050] When presenting candidates, the presentation unit can analyze the user's social media activity and adjust the presentation method. For example, the presentation unit can prioritize presenting related paraphrase candidates based on keywords frequently used by the user on social media. The presentation unit can also analyze the user's social media activity history and prioritize presenting related topics. The presentation unit can also prioritize presenting related paraphrase candidates based on the content of posts from accounts the user follows on social media. In this way, by analyzing the user's social media activity, it is possible to prioritize presenting highly relevant candidates.

[0051] During learning, the learning unit can optimize the learning algorithm by referring to past learning data. For example, the learning unit applies a more accurate learning algorithm based on the past learning data. The learning unit can also analyze the past learning data and adjust the learning algorithm based on a specific pattern. The learning unit can also improve the performance of the learning algorithm by referring to the past learning data. In this way, the accuracy of the learning algorithm is improved by referring to the past learning data.

[0052] The learning unit can perform learning while taking into account the user's attribute information. The learning unit selects appropriate learning data based on the user's age and gender, for example. The learning unit can also preferentially use related learning data based on the user's occupation and interests. The learning unit can also use appropriate learning data taking into account the user's cultural background. This allows for more appropriate learning by taking into account the user's attribute information.

[0053] During learning, the learning unit can weight the learning data based on the time of submission of the text. For example, the learning unit can weight recently submitted texts higher and perform learning. The learning unit can also weight older submitted texts lower and perform learning. The learning unit can also perform learning with appropriate weighting depending on the time of submission of the text. This allows learning to be performed with appropriate weighting depending on the time of submission of the text.

[0054] The learning unit can improve the accuracy of learning by referring to related literature during learning. For example, the learning unit can refer to related literature and optimize the learning algorithm based on the latest research results. The learning unit can also analyze related literature and adjust the learning algorithm based on specific patterns. The learning unit can also introduce new techniques for improving the accuracy of learning by referring to related literature. As a result, the accuracy of learning is improved by referring to related literature.

[0055] The context analysis unit can optimize the analysis algorithm by referring to past context data during context analysis. For example, the context analysis unit applies a more accurate context analysis algorithm based on the past context data. The context analysis unit can also analyze the past context data and adjust the analysis algorithm based on a specific pattern. The context analysis unit can also improve the performance of the analysis algorithm by referring to the past context data. As a result, the accuracy of the analysis algorithm is improved by referring to the past context data.

[0056] The context analysis unit can apply different analysis methods to different text categories during context analysis. For example, the context analysis unit applies a formal context analysis method to a business document. The context analysis unit can also apply a more familiar context analysis method to casual text. The context analysis unit can also apply an appropriate context analysis method depending on the text category. This allows for more appropriate context analysis by applying an appropriate analysis method depending on the text category.

[0057] The context analysis unit can take the geographical distribution of text into consideration when analyzing the context. For example, the context analysis unit can preferentially detect offensive words that are frequently used in a specific region. The context analysis unit can also analyze the geographical distribution of text to understand the trends in offensive words by region. The context analysis unit can also apply an appropriate analysis algorithm based on the geographical distribution. In this way, by taking the geographical distribution of text into consideration, it is possible to understand the trends in offensive words by region.

[0058] The context analysis unit can improve the accuracy of the analysis by referring to related literature during context analysis. For example, the context analysis unit can refer to related literature and optimize the analysis algorithm based on the latest research results. The context analysis unit can also analyze related literature and adjust the analysis algorithm based on specific patterns. The context analysis unit can also introduce new techniques for improving the accuracy of the analysis by referring to related literature. In this way, the accuracy of the analysis is improved by referring to related literature.

[0059] The history storage unit can optimize the storage algorithm by referring to past history data when storing history. For example, the history storage unit applies a more efficient storage algorithm based on the past history data. The history storage unit can also analyze the past history data and adjust the storage algorithm based on a specific pattern. The history storage unit can also improve the performance of the storage algorithm by referring to the past history data. As a result, the accuracy of the storage algorithm is improved by referring to the past history data.

[0060] When saving the history, the history saving unit can weight the saved data based on the time of submission of the text. For example, the history saving unit saves text that has been submitted recently with a high weight. The history saving unit can also save text that has been submitted earlier with a low weight. The history saving unit can also save the text with an appropriate weight depending on the time of submission of the text. This allows the text to be saved with an appropriate weight depending on the time of submission.

[0061] When receiving feedback, the feedback receiving unit can select the optimal reception method by referring to the user's past feedback history. For example, the feedback receiving unit preferentially suggests feedback methods (voice, text, etc.) that the user has used in the past. The feedback receiving unit can also predict and suggest the optimal reception method for a specific time period based on the user's past feedback history. The feedback receiving unit can also analyze the user's past feedback history and provide an appropriate reception interface. In this way, the optimal reception method can be provided by referring to the user's past feedback history.

[0062] When receiving feedback, the feedback receiving unit can select the optimal receiving method by taking into consideration the user's geographical location information. For example, when the user is in a specific area, the feedback receiving unit can preferentially receive feedback related to that area. Furthermore, when the user is traveling, the feedback receiving unit can also preferentially receive feedback related to the travel destination. Furthermore, when the user is at home, the feedback receiving unit can also preferentially receive feedback related to information around the user's home. In this way, by taking into consideration the user's geographical location information, it is possible to preferentially receive highly relevant feedback.

[0063] The system according to the embodiment is not limited to the above-described example, and various modifications are possible, for example, as follows.

[0064] When analyzing text, the analysis unit can adjust the analysis algorithm by referring to past analysis data. For example, it can apply a more accurate algorithm for detecting offensive words based on past analysis data. It can also analyze past analysis data and adjust the analysis algorithm based on specific patterns. It can also improve the performance of the analysis algorithm by referring to past analysis data. In this way, the accuracy of the analysis algorithm is improved by referring to past analysis data.

[0065] When generating paraphrase candidates, the generation unit can adjust the level of detail of the candidates based on the importance of the text. For example, for important text, detailed paraphrase candidates can be generated. For less important text, simple paraphrase candidates can also be generated. Furthermore, paraphrase candidates with an appropriate level of detail can be generated depending on the importance of the text. This makes it possible to generate paraphrase candidates with an appropriate level of detail depending on the importance of the text.

[0066] When presenting candidates, the presentation unit can select the optimal presentation method by referring to the user's past selection history. For example, the presentation unit can analyze patterns of paraphrase candidates previously selected by the user and preferentially present similar candidates. It can also preferentially provide a specific presentation method (list format, card format, etc.) based on the user's past selection history. Furthermore, it can also determine the optimal presentation order based on the user's past selection history. In this way, the optimal presentation method can be provided by referring to the user's past selection history.

[0067] When receiving feedback, the feedback receiving unit can select the optimal reception method by referring to the user's past feedback history. For example, it can preferentially suggest feedback methods (voice, text, etc.) that the user has used in the past. It can also predict and suggest the optimal reception method for a specific time period based on the user's past feedback history. It can also analyze the user's past feedback history and provide an appropriate reception interface. In this way, it can provide the optimal reception method by referring to the user's past feedback history.

[0068] The context analysis unit can apply different analysis methods to different text categories during context analysis. For example, a formal context analysis method can be applied to business documents. A more familiar context analysis method can also be applied to casual text. Furthermore, an appropriate context analysis method can be applied depending on the text category. This allows for more appropriate context analysis by applying the appropriate analysis method depending on the text category.

[0069] The processing flow of the first embodiment will be briefly explained below.

[0070] Step 1: The reception unit receives text input from the user. The text input from the user includes handwritten text, digital text, etc. The reception unit digitizes and reads handwritten text using scanning technology. It can also directly read text submitted in digital format. It can also read printed text using OCR technology. Step 2: The analysis unit analyzes the text received by the reception unit and detects offensive words. The analysis is performed using natural language processing technology. The analysis unit analyzes words and phrases in the text and detects offensive words. It can also detect offensive words by taking context into account. Step 3: The generation unit generates candidates for replacing the offensive words detected by the analysis unit with appropriate words. Generation is performed using natural language processing technology. The generation unit generates appropriate word candidates using text generation AI (e.g., LLM) or multimodal generation AI. The generation AI uses keyword extraction technology to pick out particularly important information in the text and generate appropriate word candidates based on that information. Step 4: The presenting unit presents the candidates generated by the generating unit to the user. The presentation is performed using a user interface. The presenting unit can present the generated candidates to the user in a list format, a card format, or a pop-up format. Step 5: The learning unit learns the words selected by the user from the candidates presented by the presentation unit. Learning is performed using machine learning technology. The learning unit saves the user's selection history and reflects it in generating candidates from the next time onwards. It can also accept user feedback and use it to improve the system.

[0071] (Example 2) A system for paraphrasing offensive words according to an embodiment of the present invention analyzes text entered by a user and detects offensive words and expressions. The system generates candidates for replacing the detected offensive words with appropriate words and presents them to the user. The user can select an appropriate word from the presented candidates. For example, the word "stupid" can be replaced with an expression such as "foolish." The system analyzes text using natural language processing technology to generate paraphrase candidates. Furthermore, the system has a function for learning user selections and reflecting them in subsequent candidate generation. The system also has a function for generating paraphrase candidates taking context into account, enabling more natural paraphrasing. For example, if the text entered by a user contains the word "stupid," the system detects the word and generates appropriate candidate words such as "stupid" or "ignorant." The user can select from these candidates to correct the text. The system saves the user's selection history and reflects it in subsequent candidate generation, thereby providing more accurate paraphrase candidates. Furthermore, because paraphrase candidates are generated taking context into account, natural paraphrasing is possible without compromising the overall meaning of the text. For example, the system analyzes the context of the text entered by the user and generates appropriate paraphrase candidates, thereby allowing the user to enjoy a comfortable text entry environment. Thus, the system for paraphrasing offensive words into other words analyzes the text entered by the user and replaces the offensive words with appropriate words, thereby providing the user with a comfortable text entry environment.

[0072] A system for paraphrasing offensive words according to an embodiment includes a reception unit, an analysis unit, a generation unit, a presentation unit, and a learning unit. The reception unit receives text input from a user. The text input from the user includes, but is not limited to, handwritten text and digital text. For example, the reception unit digitizes and reads the handwritten text using scanning technology. The reception unit can also directly read text submitted in digital format. The reception unit can also read printed text using OCR technology. For example, the reception unit scans handwritten text with a high-resolution scanner and converts it into character information using OCR technology. Digital text submitted in a specific file format can also be directly read. OCR technology recognizes printed characters with high accuracy and converts them into digital text. The analysis unit analyzes the text received by the reception unit to detect offensive words. The analysis is performed using, for example, natural language processing technology, but is not limited to, for example. For example, the analysis unit analyzes words and phrases in the text to detect offensive words. The analysis unit can also detect offensive words by taking context into consideration. The generation unit generates candidates for replacing the offensive words detected by the analysis unit with appropriate words. The generation is performed, for example, using natural language processing technology, but is not limited to such an example. For example, the generation unit generates appropriate word candidates using a text generation AI (e.g., LLM). The generation unit can also generate appropriate word candidates using a multimodal generation AI. The generation unit can also generate appropriate word candidates by taking context into consideration using the generation AI. For example, the text generation AI has learned a large amount of text data and has advanced natural language processing capabilities. The multimodal generation AI can handle multiple modalities, including not only text but also images and audio. The generation AI uses keyword extraction technology to pick out particularly important information in the text and generate appropriate word candidates based on that information. The presentation unit presents the candidates generated by the generation unit to the user. The presentation is performed, for example, using a user interface, but is not limited to such an example.For example, the presentation unit presents the generated candidates to the user in a list format. The presentation unit can also present the generated candidates to the user in a card format. The presentation unit can also present the generated candidates to the user in a pop-up format. The learning unit learns words selected by the user from the candidates presented by the presentation unit. Learning is performed, for example, using machine learning technology, but is not limited to this example. For example, the learning unit saves the user's selection history and reflects it in generating candidates from the next time onwards. The learning unit can also receive user feedback and use it to improve the system. As a result, the system for paraphrasing offensive words according to the embodiment can provide a comfortable text input environment for the user by analyzing the user's text input and replacing the offensive words with appropriate words.

[0073] The context analysis unit can generate paraphrase candidates based on the context. For example, when the generative AI is summarizing, the context analysis unit can automatically collect relevant background information and refer to it to understand the context. For example, the context analysis unit collects related news articles and academic papers. The context analysis unit can also use a topic model to understand the context when the generative AI is summarizing. For example, the context analysis unit extracts related keywords and phrases based on the topic model. The topic model can be realized using techniques such as LDA (Latent Dirichlet Allocation), non-negative matrix factorization, or topic clustering. The context analysis unit can also build a system for understanding the context by referring to relevant background information and topic models when the generative AI is summarizing. For example, the context analysis unit automatically collects relevant information and reflects it in the summary. This allows for more accurate summarization by referring to background information and topic models to understand the context. Furthermore, the context analysis unit can collect relevant information in real time to understand the context and reflect it in the summary. For example, the context analysis unit collects the latest news articles and social media posts from the Internet and uses them as information sources for understanding the context. The context analysis unit can also use topic models to create a feedback loop to improve the accuracy of summaries. For example, the context analysis unit evaluates the quality of the generated summaries and updates the topic model based on the evaluation results. This allows the context analysis unit to always understand the context based on the latest information and provide more accurate summaries. This allows for more natural paraphrasing by taking context into consideration.

[0074] The history storage unit can store the user's selection history. The history storage unit stores, for example, words selected by the user and the date and time of selection. For example, the history storage unit stores the words selected by the user in a database. The history storage unit can also store the date and time of the user's selection as a timestamp. The history storage unit can also analyze the user's selection history and reflect it in subsequent candidate generation. For example, the history storage unit analyzes patterns of words selected by the user in the past and prioritizes generating similar words as candidates. The history storage unit can also generate word candidates that are optimal for a specific time period based on the user's selection history. In this way, by storing the user's selection history, it can be reflected in subsequent candidate generation.

[0075] The feedback receiving unit can receive user feedback. The feedback receiving unit can receive, for example, user opinions and evaluations. For example, the feedback receiving unit can receive text feedback entered by the user. The feedback receiving unit can also receive feedback provided by the user via voice. The feedback receiving unit can also analyze the user feedback and use it to improve the system. For example, the feedback receiving unit can store the user feedback content in a database and identify areas for improvement in the system. The feedback receiving unit can also develop an algorithm for improving the system's functions based on the user feedback. In this way, receiving user feedback can be used to improve the system.

[0076] The generator can generate paraphrase candidates based on context. For example, when the generation AI generates a summary, the generator can automatically collect relevant background information and refer to it to understand the context. For example, the generator can collect related news articles and academic papers. The generator can also use a topic model to understand the context when the generation AI generates a summary. For example, the generator can extract related keywords and phrases based on the topic model. The topic model can be realized using techniques such as LDA (Latent Dirichlet Allocation), non-negative matrix factorization, or topic clustering. The generator can also build a system for understanding the context by referring to relevant background information and topic models when the generation AI generates a summary. For example, the generator can automatically collect relevant information and incorporate it into the summary. This enables more accurate summaries by referring to background information and topic models to understand the context. Furthermore, the generator can collect relevant information in real time to understand the context and incorporate it into the summary. For example, the generator can collect the latest news articles and social media posts from the Internet and use them as sources of information for understanding the context. The generator can also use the topic model to create a feedback loop to improve the accuracy of summaries. For example, the generator evaluates the quality of the generated summaries and updates the topic model based on the evaluation results. This allows the generator to always understand the context based on the latest information and provide more accurate summaries. This allows for more natural paraphrasing by taking context into account.

[0077] The learning unit can save the user's selection history and reflect it in subsequent candidate generation. The learning unit, for example, saves words selected by the user and the date and time of selection. For example, the learning unit saves the words selected by the user in a database. The learning unit can also save the date and time of the user's selection as a timestamp. The learning unit can also analyze the user's selection history and reflect it in subsequent candidate generation. For example, the learning unit analyzes patterns of words selected by the user in the past and prioritizes generating similar words as candidates. The learning unit can also generate candidate words that are optimal for a specific time period based on the user's selection history. In this way, by reflecting the user's selection history, the accuracy of candidate generation can be improved in subsequent candidate generation.

[0078] The reception unit can estimate the user's emotions and adjust the timing of accepting text input based on the estimated user emotions. For example, if the user is feeling stressed, the reception unit can temporarily delay accepting the input to provide time for the user to relax. Furthermore, if the user is relaxed, the reception unit can immediately accept the text input to promote smooth operation. Furthermore, if the user is in a hurry, the reception unit can quickly accept the input and immediately start analyzing it. By adjusting the timing of accepting the text input according to the user's emotions, it is possible to accept the text input at a more appropriate timing. The emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or a generation AI. The generation AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0079] The reception unit can analyze the user's past text input history and select the optimal reception method. For example, the reception unit preferentially suggests input methods (voice input, text input, etc.) that the user has frequently used in the past. The reception unit can also predict and suggest the optimal input method for a specific time period based on the user's past input history. The reception unit can also analyze the content of text that the user has previously input and provide an appropriate input interface. In this way, the optimal reception method can be provided by analyzing the user's past input history.

[0080] The reception unit can perform filtering based on the user's current situation and areas of interest when inputting text. For example, the reception unit preferentially receives topics and keywords related to the user's current situation. The reception unit can also preferentially receive related text based on the user's areas of interest. The reception unit can also filter unnecessary text based on the user's current situation and areas of interest. In this way, filtering based on the user's current situation and areas of interest can preferentially receive highly relevant text.

[0081] The reception unit can estimate the user's emotions and determine the priority of text to be received based on the estimated user emotions. For example, when the user is feeling stressed, the reception unit postpones less important text and prioritizes receiving important text. Furthermore, when the user is relaxed, the reception unit can equally accept all text. Furthermore, when the user is in a hurry, the reception unit can prioritize urgent text. In this way, by determining the priority of text according to the user's emotions, important text can be prioritized. Emotion estimation is realized using an emotion estimation function using, for example, an emotion engine or a generation AI. The generation AI can be, for example, a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to these examples.

[0082] When inputting text, the reception unit can preferentially receive highly relevant text by taking into account the user's geographical location information. For example, when the user is in a specific area, the reception unit can preferentially receive text related to that area. Furthermore, when the user is traveling, the reception unit can preferentially receive text related to the travel destination. Furthermore, when the user is at home, the reception unit can preferentially receive text related to information about the area around the user's home. In this way, highly relevant text can be preferentially received by taking into account the user's geographical location information.

[0083] The reception unit can analyze the user's social media activity and receive related text when the user inputs text. The reception unit can, for example, preferentially receive related text based on keywords frequently used by the user on social media. The reception unit can also analyze the user's social media activity history and preferentially receive related topics. The reception unit can also preferentially receive related text based on the content of posts from accounts the user follows on social media. In this way, by analyzing the user's social media activity, it is possible to preferentially receive related text.

[0084] The analysis unit can estimate the user's emotions and adjust the detection criteria for offensive words based on the estimated user emotions. For example, if the user is stressed, the analysis unit can detect offensive words using stricter criteria. Furthermore, if the user is relaxed, the analysis unit can also detect offensive words using normal criteria. Furthermore, if the user is in a hurry, the analysis unit can adjust the criteria to quickly detect offensive words. This allows for more appropriate detection by adjusting the detection criteria for offensive words according to the user's emotions. The emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or a generation AI. The generation AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0085] When analyzing text, the analysis unit can adjust the analysis algorithm by referring to past analysis data. For example, the analysis unit can apply a more accurate offensive language detection algorithm based on the past analysis data. The analysis unit can also analyze the past analysis data and adjust the analysis algorithm based on specific patterns. The analysis unit can also improve the performance of the analysis algorithm by referring to the past analysis data. In this way, the accuracy of the analysis algorithm is improved by referring to the past analysis data.

[0086] The analysis unit can perform text analysis while taking into account user attribute information. The analysis unit selects an appropriate analysis algorithm based on the user's age and gender, for example. The analysis unit can also preferentially detect relevant offensive words based on the user's occupation and interests. The analysis unit can also apply an appropriate analysis algorithm while taking into account the user's cultural background. This allows for more appropriate analysis by taking into account the user's attribute information.

[0087] The analysis unit can estimate the user's emotions and adjust the display order of the analysis results based on the estimated user emotions. For example, when the user is feeling stressed, the analysis unit can prioritize displaying important analysis results. Furthermore, when the user is relaxed, the analysis unit can also display all analysis results evenly. Furthermore, when the user is in a hurry, the analysis unit can summarize and display the analysis results so that they can be checked quickly. In this way, by adjusting the display order of the analysis results according to the user's emotions, important analysis results can be prioritized and displayed. Emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to these examples.

[0088] The analysis unit can perform text analysis while taking into account the geographic distribution of the text. For example, the analysis unit can preferentially detect offensive words that are frequently used in a specific region. The analysis unit can also analyze the geographic distribution of the text to understand the trends in offensive words by region. The analysis unit can also apply an appropriate analysis algorithm based on the geographic distribution. In this way, by taking the geographic distribution of the text into account, it is possible to understand the trends in offensive words by region.

[0089] The analysis unit can improve the accuracy of the analysis by referring to related literature during text analysis. For example, the analysis unit can refer to related literature and optimize the analysis algorithm based on the latest research results. The analysis unit can also analyze related literature and adjust the analysis algorithm based on specific patterns. The analysis unit can also introduce new techniques for improving the accuracy of the analysis by referring to related literature. In this way, the accuracy of the analysis is improved by referring to related literature.

[0090] The generation unit can estimate the user's emotions and adjust the method for generating paraphrase candidates based on the estimated user's emotions. For example, if the user is feeling stressed, the generation unit generates simple and easy-to-understand paraphrase candidates. Furthermore, if the user is relaxed, the generation unit can also generate detailed paraphrase candidates. Furthermore, if the user is in a hurry, the generation unit can also generate short paraphrase candidates that can be quickly checked. This allows for more appropriate paraphrase candidates to be provided by adjusting the method for generating paraphrase candidates according to the user's emotions. The emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or generation AI. The generation AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0091] When generating paraphrase candidates, the generation unit can adjust the level of detail of the candidates based on the importance of the text. For example, in the case of important text, the generation unit generates detailed paraphrase candidates. In addition, in the case of text with low importance, the generation unit can also generate simple paraphrase candidates. In addition, the generation unit can also generate paraphrase candidates with an appropriate level of detail according to the importance of the text. In this way, paraphrase candidates with an appropriate level of detail can be generated according to the importance of the text.

[0092] When generating paraphrase candidates, the generation unit can apply different generation algorithms depending on the text category. For example, in the case of a business document, the generation unit generates formal paraphrase candidates. In addition, in the case of casual text, the generation unit can also generate familiar paraphrase candidates. In addition, the generation unit can apply an appropriate generation algorithm depending on the text category. In this way, by applying an appropriate generation algorithm depending on the text category, more appropriate paraphrase candidates can be provided.

[0093] The generation unit can estimate the user's emotions and adjust the length of the paraphrase candidates based on the estimated user's emotions. For example, if the user is feeling stressed, the generation unit can generate short and concise paraphrase candidates. Furthermore, if the user is relaxed, the generation unit can also generate detailed paraphrase candidates. Furthermore, if the user is in a hurry, the generation unit can also generate short paraphrase candidates that can be quickly checked. This allows for adjusting the length of the paraphrase candidates according to the user's emotions, thereby providing more appropriate paraphrase candidates. The emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or generation AI. The generation AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0094] When generating paraphrase candidates, the generation unit can determine the priority of the candidates based on the time of submission of the text. For example, in the case of recently submitted text, the generation unit can generate paraphrase candidates with priority. In addition, in the case of text that was submitted earlier, the generation unit can also generate paraphrase candidates later. In addition, the generation unit can generate paraphrase candidates with an appropriate priority depending on the time of submission of the text. In this way, paraphrase candidates can be generated with an appropriate priority depending on the time of submission of the text.

[0095] When generating paraphrase candidates, the generation unit can adjust the order of the candidates based on the relevance of the text. For example, the generation unit preferentially displays paraphrase candidates that are highly relevant to the content of the text. The generation unit can also display paraphrase candidates that are less relevant later. The generation unit can also display paraphrase candidates in an appropriate order based on the relevance of the text. This makes it possible to display paraphrase candidates in an appropriate order based on the relevance of the text.

[0096] The presentation unit can estimate the user's emotions and adjust the way in which candidates are presented based on the estimated user emotions. For example, if the user is feeling stressed, the presentation unit can provide a simple, highly visible presentation method. If the user is relaxed, the presentation unit can also provide a presentation method that includes detailed information. If the user is in a hurry, the presentation unit can also provide a presentation method that focuses on the main points. This allows for more appropriate presentation by adjusting the way in which candidates are presented according to the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generation AI. The generation AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0097] When presenting candidates, the presentation unit can select the optimal presentation method by referring to the user's past selection history. For example, the presentation unit analyzes patterns of paraphrase candidates previously selected by the user and preferentially presents similar candidates. The presentation unit can also preferentially provide a specific presentation method (list format, card format, etc.) based on the user's past selection history. The presentation unit can also determine the optimal presentation order based on the user's past selection history. In this way, the optimal presentation method can be provided by referring to the user's past selection history.

[0098] The presentation unit can present candidates while taking into consideration the user's attribute information. The presentation unit selects an appropriate presentation method based on, for example, the user's age and gender. The presentation unit can also preferentially present related paraphrase candidates based on the user's occupation and interests. The presentation unit can also provide an appropriate presentation method by taking into consideration the user's cultural background. This allows for more appropriate presentation by taking into consideration the user's attribute information.

[0099] The presentation unit can estimate the user's emotions and adjust the display order of candidates based on the estimated user emotions. For example, when the user is feeling stressed, the presentation unit can prioritize displaying important candidates. Furthermore, when the user is relaxed, the presentation unit can also display all candidates evenly. Furthermore, when the user is in a hurry, the presentation unit can summarize and display candidates so that they can be checked quickly. In this way, by adjusting the display order of candidates according to the user's emotions, important candidates can be prioritized and displayed. Emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to these examples.

[0100] When presenting candidates, the presentation unit can select the optimal presentation method by taking into consideration the user's geographical location information. For example, when the user is in a specific area, the presentation unit can preferentially present paraphrase candidates related to that area. Furthermore, when the user is traveling, the presentation unit can also preferentially present paraphrase candidates related to the user's travel destination. Furthermore, when the user is at home, the presentation unit can also preferentially present paraphrase candidates related to information about the area around the user's home. In this way, by taking into consideration the user's geographical location information, highly relevant candidates can be preferentially presented.

[0101] When presenting candidates, the presentation unit can analyze the user's social media activity and adjust the presentation method. For example, the presentation unit can prioritize presenting related paraphrase candidates based on keywords frequently used by the user on social media. The presentation unit can also analyze the user's social media activity history and prioritize presenting related topics. The presentation unit can also prioritize presenting related paraphrase candidates based on the content of posts from accounts the user follows on social media. In this way, by analyzing the user's social media activity, it is possible to prioritize presenting highly relevant candidates.

[0102] The learning unit can estimate the user's emotions and select training data based on the estimated user emotions. For example, if the user is feeling stressed, the learning unit prioritizes learning data that is useful for stress reduction. The learning unit can also use normal training data when the user is relaxed. The learning unit can also select data that can be quickly learned when the user is in a hurry. This allows for more appropriate learning by selecting training data according to the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. Generative AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0103] During learning, the learning unit can optimize the learning algorithm by referring to past learning data. For example, the learning unit applies a more accurate learning algorithm based on the past learning data. The learning unit can also analyze the past learning data and adjust the learning algorithm based on a specific pattern. The learning unit can also improve the performance of the learning algorithm by referring to the past learning data. In this way, the accuracy of the learning algorithm is improved by referring to the past learning data.

[0104] The learning unit can perform learning while taking into account the user's attribute information. The learning unit selects appropriate learning data based on the user's age and gender, for example. The learning unit can also preferentially use related learning data based on the user's occupation and interests. The learning unit can also use appropriate learning data taking into account the user's cultural background. This allows for more appropriate learning by taking into account the user's attribute information.

[0105] The learning unit can estimate the user's emotions and adjust the frequency of learning based on the estimated user emotions. For example, if the user is feeling stressed, the learning unit can reduce the frequency of learning and provide time for relaxation. The learning unit can also perform learning at a normal frequency when the user is relaxed. The learning unit can also increase the frequency of learning and perform learning quickly when the user is in a hurry. This allows for more appropriate learning by adjusting the frequency of learning according to the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generative AI. The generative AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to these examples.

[0106] During learning, the learning unit can weight the learning data based on the time of submission of the text. For example, the learning unit can weight recently submitted texts higher and perform learning. The learning unit can also weight older submitted texts lower and perform learning. The learning unit can also perform learning with appropriate weighting depending on the time of submission of the text. This allows learning to be performed with appropriate weighting depending on the time of submission of the text.

[0107] The learning unit can improve the accuracy of learning by referring to related literature during learning. For example, the learning unit can refer to related literature and optimize the learning algorithm based on the latest research results. The learning unit can also analyze related literature and adjust the learning algorithm based on specific patterns. The learning unit can also introduce new techniques for improving the accuracy of learning by referring to related literature. As a result, the accuracy of learning is improved by referring to related literature.

[0108] The context analysis unit can estimate the user's emotions and adjust the criteria for context analysis based on the estimated user emotions. For example, if the user is stressed, the context analysis unit can perform context analysis using stricter criteria. Furthermore, if the user is relaxed, the context analysis unit can also perform context analysis using normal criteria. Furthermore, if the user is in a hurry, the context analysis unit can also adjust the criteria to perform context analysis quickly. This allows for more appropriate context analysis by adjusting the criteria for context analysis according to the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or a generative AI. The generative AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0109] The context analysis unit can optimize the analysis algorithm by referring to past context data during context analysis. For example, the context analysis unit applies a more accurate context analysis algorithm based on the past context data. The context analysis unit can also analyze the past context data and adjust the analysis algorithm based on a specific pattern. The context analysis unit can also improve the performance of the analysis algorithm by referring to the past context data. As a result, the accuracy of the analysis algorithm is improved by referring to the past context data.

[0110] The context analysis unit can apply different analysis methods to different text categories during context analysis. For example, the context analysis unit applies a formal context analysis method to a business document. The context analysis unit can also apply a more familiar context analysis method to casual text. The context analysis unit can also apply an appropriate context analysis method depending on the text category. This allows for more appropriate context analysis by applying an appropriate analysis method depending on the text category.

[0111] The context analysis unit can estimate the user's emotions and adjust the order in which the context analysis results are displayed based on the estimated user emotions. For example, when the user is stressed, the context analysis unit can prioritize displaying important analysis results. Furthermore, when the user is relaxed, the context analysis unit can also display all analysis results evenly. Furthermore, when the user is in a hurry, the context analysis unit can summarize and display the analysis results so that they can be checked quickly. Thus, by adjusting the order in which the context analysis results are displayed according to the user's emotions, important analysis results can be prioritized. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or a generation AI. The generation AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0112] The context analysis unit can take the geographical distribution of text into consideration when analyzing the context. For example, the context analysis unit can preferentially detect offensive words that are frequently used in a specific region. The context analysis unit can also analyze the geographical distribution of text to understand the trends in offensive words by region. The context analysis unit can also apply an appropriate analysis algorithm based on the geographical distribution. In this way, by taking the geographical distribution of text into consideration, it is possible to understand the trends in offensive words by region.

[0113] The context analysis unit can improve the accuracy of the analysis by referring to related literature during context analysis. For example, the context analysis unit can refer to related literature and optimize the analysis algorithm based on the latest research results. The context analysis unit can also analyze related literature and adjust the analysis algorithm based on specific patterns. The context analysis unit can also introduce new techniques for improving the accuracy of the analysis by referring to related literature. In this way, the accuracy of the analysis is improved by referring to related literature.

[0114] The history storage unit can estimate the user's emotions and select the history to save based on the estimated user emotions. For example, if the user is feeling stressed, the history storage unit saves only important history. Furthermore, if the user is relaxed, the history storage unit can save all history. Furthermore, if the user is in a hurry, the history storage unit can save summarized history for quick review. This allows for more appropriate history saving by selecting the history to save based on the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generation AI. The generation AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0115] The history storage unit can optimize the storage algorithm by referring to past history data when storing history. For example, the history storage unit applies a more efficient storage algorithm based on the past history data. The history storage unit can also analyze the past history data and adjust the storage algorithm based on a specific pattern. The history storage unit can also improve the performance of the storage algorithm by referring to the past history data. As a result, the accuracy of the storage algorithm is improved by referring to the past history data.

[0116] The history storage unit can estimate the user's emotions and adjust the frequency of history storage based on the estimated user emotions. For example, if the user is feeling stressed, the history storage unit can reduce the storage frequency to provide time for the user to relax. The history storage unit can also store history at a normal frequency when the user is relaxed. The history storage unit can also increase the storage frequency to quickly store history when the user is in a hurry. This allows for more appropriate history storage by adjusting the history storage frequency according to the user's emotions. Emotion estimation is achieved using an emotion estimation function, for example, using an emotion engine or generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to these examples.

[0117] When saving the history, the history saving unit can weight the saved data based on the time of submission of the text. For example, the history saving unit saves text that has been submitted recently with a high weight. The history saving unit can also save text that has been submitted earlier with a low weight. The history saving unit can also save the text with an appropriate weight depending on the time of submission of the text. This allows the text to be saved with an appropriate weight depending on the time of submission.

[0118] The feedback receiving unit can estimate the user's emotions and adjust the feedback receiving method based on the estimated user's emotions. For example, when the user is feeling stressed, the feedback receiving unit can provide a simple feedback form and quickly accept feedback. Furthermore, when the user is relaxed, the feedback receiving unit can provide a detailed feedback form and carefully accept feedback. Furthermore, when the user is in a hurry, the feedback receiving unit can prioritize voice input and quickly accept feedback. This allows for more appropriate feedback reception by adjusting the feedback receiving method according to the user's emotions. Emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or generation AI. The generation AI can be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI.

[0119] When receiving feedback, the feedback receiving unit can select the optimal reception method by referring to the user's past feedback history. For example, the feedback receiving unit preferentially suggests feedback methods (voice, text, etc.) that the user has used in the past. The feedback receiving unit can also predict and suggest the optimal reception method for a specific time period based on the user's past feedback history. The feedback receiving unit can also analyze the user's past feedback history and provide an appropriate reception interface. In this way, the optimal reception method can be provided by referring to the user's past feedback history.

[0120] The feedback receiving unit can estimate the user's emotions and determine the priority of feedback based on the estimated user's emotions. For example, when the user is feeling stressed, the feedback receiving unit postpones less important feedback and prioritizes important feedback. Furthermore, when the user is relaxed, the feedback receiving unit can equally accept all feedback. Furthermore, when the user is in a hurry, the feedback receiving unit can prioritize urgent feedback. In this way, by determining the priority of feedback according to the user's emotions, important feedback can be prioritized. The emotion estimation is realized using an emotion estimation function using, for example, an emotion engine or a generation AI. The generation AI can be, for example, a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to these examples.

[0121] When receiving feedback, the feedback receiving unit can select the optimal receiving method by taking into consideration the user's geographical location information. For example, when the user is in a specific area, the feedback receiving unit can preferentially receive feedback related to that area. Furthermore, when the user is traveling, the feedback receiving unit can also preferentially receive feedback related to the travel destination. Furthermore, when the user is at home, the feedback receiving unit can also preferentially receive feedback related to information around the user's home. In this way, by taking into consideration the user's geographical location information, it is possible to preferentially receive highly relevant feedback. === Hard Collateral 1-1 === Each of the above-described elements, including the reception unit, analysis unit, generation unit, presentation unit, and learning unit, is implemented, for example, by at least one of the smart device 14 and the data processing device 12. For example, the reception unit is implemented by the reception device 38 of the smart device 14 and receives text input from a user. The analysis unit is implemented, for example, by the specific processing unit 290 of the data processing device 12 and analyzes the received text to detect offensive words. The generation unit is implemented, for example, by the specific processing unit 290 of the data processing device 12 and generates candidates for replacing the detected offensive words with appropriate words. The presentation unit is implemented, for example, by the output device 40 of the smart device 14 and presents the generated candidates to the user. The learning unit is implemented, for example, by the specific processing unit 290 of the data processing device 12 and learns the user's selection history and reflects it in subsequent candidate generation. === Hard Collateral 1-2 === Each of the multiple elements including the above-mentioned reception unit, analysis unit, generation unit, presentation unit, and learning unit is realized, for example, by at least one of the smart glasses 214 and the data processing device 12. For example, the reception unit is realized by the microphone 238 of the smart glasses 214 and receives voice input from the user. The analysis unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and analyzes the received text to detect offensive words. The generation unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and generates candidates for replacing the detected offensive words with appropriate words. The presentation unit is realized, for example, by the speaker 240 of the smart glasses 214 and presents the generated candidates to the user. The learning unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and learns the user's selection history and reflects it in subsequent candidate generation. === Hard Collateral 1-3 === Each of the multiple elements including the above-mentioned reception unit, analysis unit, generation unit, presentation unit, and learning unit is realized, for example, by at least one of the headset-type terminal 314 and the data processing device 12. For example, the reception unit is realized by the microphone 238 of the headset-type terminal 314 and receives voice input from the user. The analysis unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and analyzes the received text to detect offensive words. The generation unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and generates candidates for replacing the detected offensive words with appropriate words. The presentation unit is realized, for example, by the speaker 240 of the headset-type terminal 314 and presents the generated candidates to the user. The learning unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and learns the user's selection history and reflects it in subsequent candidate generation. === Hard Collateral 1-4 === Each of the multiple elements including the above-mentioned reception unit, analysis unit, generation unit, presentation unit, and learning unit is realized, for example, by at least one of the robot 414 and the data processing device 12. For example, the reception unit is realized by the microphone 238 of the robot 414 and receives voice input from the user. The analysis unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and analyzes the received text to detect offensive words. The generation unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and generates candidates for replacing the detected offensive words with appropriate words. The presentation unit is realized, for example, by the speaker 240 of the robot 414 and presents the generated candidates to the user. The learning unit is realized, for example, by the specific processing unit 290 of the data processing device 12 and learns the user's selection history and reflects it in subsequent candidate generation.

[0122] The system according to the embodiment is not limited to the above-described example, and various modifications are possible, for example, as follows.

[0123] The analysis unit can estimate the user's emotions and adjust the detection criteria for offensive words based on the estimated user emotions. For example, if the user is feeling stressed, offensive words can be detected using stricter criteria. Alternatively, if the user is relaxed, offensive words can be detected using normal criteria. Furthermore, if the user is in a hurry, the criteria can be adjusted to quickly detect offensive words. This allows for more appropriate detection by adjusting the detection criteria for offensive words according to the user's emotions.

[0124] The generation unit can estimate the user's emotions and adjust the method for generating paraphrase candidates based on the estimated user's emotions. For example, if the user is feeling stressed, simple and easy-to-understand paraphrase candidates can be generated. If the user is relaxed, detailed paraphrase candidates can be generated. Furthermore, if the user is in a hurry, short paraphrase candidates can be generated so that the user can quickly check them. In this way, by adjusting the method for generating paraphrase candidates according to the user's emotions, more appropriate paraphrase candidates can be provided.

[0125] The presentation unit can estimate the user's emotions and adjust the way in which candidates are presented based on the estimated user emotions. For example, if the user is feeling stressed, a simple, highly visible presentation method can be provided. If the user is relaxed, a presentation method including detailed information can be provided. Furthermore, if the user is in a hurry, a presentation method that focuses on the main points can be provided. This allows for more appropriate presentation by adjusting the way in which candidates are presented according to the user's emotions.

[0126] The learning unit can estimate the user's emotions and select learning data based on the estimated user emotions. For example, if the user is feeling stressed, data that is useful for stress reduction is prioritized for learning. Also, if the user is relaxed, regular learning data can be used. Furthermore, if the user is in a hurry, data that can be learned quickly can be selected. In this way, more appropriate learning is possible by selecting learning data according to the user's emotions.

[0127] The history storage unit can estimate the user's emotions and select the history to be saved based on the estimated user emotions. For example, if the user is feeling stressed, only important history can be saved. If the user is relaxed, all history can be saved. Furthermore, if the user is in a hurry, a summary of history can be saved for quick confirmation. This allows for more appropriate history saving by selecting the history to be saved according to the user's emotions.

[0128] When analyzing text, the analysis unit can adjust the analysis algorithm by referring to past analysis data. For example, it can apply a more accurate algorithm for detecting offensive words based on past analysis data. It can also analyze past analysis data and adjust the analysis algorithm based on specific patterns. It can also improve the performance of the analysis algorithm by referring to past analysis data. In this way, the accuracy of the analysis algorithm is improved by referring to past analysis data.

[0129] When generating paraphrase candidates, the generation unit can adjust the level of detail of the candidates based on the importance of the text. For example, for important text, detailed paraphrase candidates can be generated. For less important text, simple paraphrase candidates can also be generated. Furthermore, paraphrase candidates with an appropriate level of detail can be generated depending on the importance of the text. This makes it possible to generate paraphrase candidates with an appropriate level of detail depending on the importance of the text.

[0130] When presenting candidates, the presentation unit can select the optimal presentation method by referring to the user's past selection history. For example, the presentation unit can analyze patterns of paraphrase candidates previously selected by the user and preferentially present similar candidates. It can also preferentially provide a specific presentation method (list format, card format, etc.) based on the user's past selection history. Furthermore, it can also determine the optimal presentation order based on the user's past selection history. In this way, the optimal presentation method can be provided by referring to the user's past selection history.

[0131] When receiving feedback, the feedback receiving unit can select the optimal reception method by referring to the user's past feedback history. For example, it can preferentially suggest feedback methods (voice, text, etc.) that the user has used in the past. It can also predict and suggest the optimal reception method for a specific time period based on the user's past feedback history. It can also analyze the user's past feedback history and provide an appropriate reception interface. In this way, it can provide the optimal reception method by referring to the user's past feedback history.

[0132] The context analysis unit can apply different analysis methods to different text categories during context analysis. For example, a formal context analysis method can be applied to business documents. A more familiar context analysis method can also be applied to casual text. Furthermore, an appropriate context analysis method can be applied depending on the text category. This allows for more appropriate context analysis by applying the appropriate analysis method depending on the text category.

[0133] The processing flow of the second embodiment will be briefly explained below.

[0134] Step 1: The reception unit receives text input from the user. The text input from the user includes handwritten text, digital text, etc. The reception unit digitizes and reads handwritten text using scanning technology. It can also directly read text submitted in digital format. It can also read printed text using OCR technology. Step 2: The analysis unit analyzes the text received by the reception unit and detects offensive words. The analysis is performed using natural language processing technology. The analysis unit analyzes words and phrases in the text and detects offensive words. It can also detect offensive words by taking context into account. Step 3: The generation unit generates candidates for replacing the offensive words detected by the analysis unit with appropriate words. Generation is performed using natural language processing technology. The generation unit generates appropriate word candidates using text generation AI (e.g., LLM) or multimodal generation AI. The generation AI uses keyword extraction technology to pick out particularly important information in the text and generate appropriate word candidates based on that information. Step 4: The presenting unit presents the candidates generated by the generating unit to the user. The presentation is performed using a user interface. The presenting unit can present the generated candidates to the user in a list format, a card format, or a pop-up format. Step 5: The learning unit learns the words selected by the user from the candidates presented by the presentation unit. Learning is performed using machine learning technology. The learning unit saves the user's selection history and reflects it in generating candidates from the next time onwards. It can also accept user feedback and use it to improve the system.

[0135] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0136] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> Examples of the generative AI include a neural network (NN) and a neural network (NN). The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 receives a prompt containing an instruction, as well as inference data such as voice data representing speech, text data representing text, and image data representing an image (e.g., still image data or video data). The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in one or more data formats of voice data, text data, image data, etc. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specification processing unit 290 performs the above-mentioned specification processing using the data generation model 58. The data generation model 58 may be a fine-tuned model so as to output an inference result from a prompt that does not include an instruction. In this case, the data generation model 58 can output an inference result from a prompt that does not include an instruction. The data processing device 12 and the like include multiple types of data generation models 58, and the data generation model 58 includes AIs other than the generative AI. The AI ​​other than the generative AI may be, for example, linear regression, logistic regression, decision tree, random forest, support vector machine (SVM), k-means clustering, convolutional neural network (CNN), recurrent neural network (RNN), generative adversarial network (GAN), or naive Bayes, and may perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processing of each of the above-mentioned parts is performed by an AI, the processing may be performed in part or entirely by the AI, but is not limited to these examples. The processing performed by an AI including the generative AI may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by an AI including the generative AI.

[0137] Furthermore, the processing by the data processing system 10 described above is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but may also be executed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. Furthermore, 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.

[0138] The correspondence between each part and the device or control part is not limited to the example described above, and various modifications are possible.

[0139] [Second embodiment] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0140] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0141] 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, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN and / or a LAN.

[0142] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0143] The microphone 238 receives instructions and the like from the user by receiving voice uttered by the user. The microphone 238 captures the voice uttered by the user, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to instructions from the processor 46.

[0144] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the user's surroundings (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0145] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0146] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0147] The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0148] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate a user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotion, including estimation and prediction of the user's emotion, but is not limited to these examples. Furthermore, the estimation and prediction of emotion also includes, for example, emotion analysis.

[0149] In the smart glasses 214, the specific processing is performed by the processor 46. A specific processing program 60 is stored in the storage 50. 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 the control unit 46A in accordance with the specific processing program 60 executed on the RAM 48. The smart glasses 214 also have a data generation model and an emotion identification model similar to the data generation model 58 and the emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0150] Note that a device other than the data processing device 12 may 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 communicates with the server device having the data generation model 58 to obtain a processing result (such as a prediction result) using the data generation model 58. Furthermore, the data processing device 12 may be a server device, or may be a terminal device (for example, a mobile phone, a robot, a home appliance, etc.) owned by a user.

[0151] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0152] The data generation model 58 is a so-called generative AI. An example of the 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 receives a prompt including an instruction, as well as inference data such as audio data indicating speech, text data indicating text, and image data indicating an image (e.g., still image data or video data). The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in one or more data formats, such as audio data, text data, and image data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The identification processing unit 290 performs the above-mentioned identification processing using the data generation model 58. The data generation model 58 may be a fine-tuned model so as to output an inference result from a prompt that does not include an instruction. In this case, the data generation model 58 can output an inference result from a prompt that does not include an instruction. The data processing device 12 and the like include multiple types of data generation models 58, and the data generation models 58 include AIs other than the generation AI. Examples of AIs other than the generation AI include, but are not limited to, 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), and naive Bayes. These AIs can perform various types of processing, but are not limited to these examples. The AI ​​may also be an AI agent. When the processing of each of the above-described parts is performed by an AI, the processing may be performed in part or entirely by the AI, but is not limited to these examples. Processing performed by an AI, including the generation AI, may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by an AI, including the generation AI.

[0153] 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 executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but may also be executed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the smart glasses 214 or an external device, etc., and the smart glasses 214 acquires or collects information required for processing from the data processing device 12 or an external device, etc.

[0154] The correspondence between each part and the device or control part is not limited to the example described above, and various modifications are possible.

[0155] [Third embodiment] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0156] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0157] 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, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN and / or a LAN.

[0158] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0159] The microphone 238 receives instructions and the like from the user by receiving voice uttered by the user. The microphone 238 captures the voice uttered by the user, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to instructions from the processor 46.

[0160] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the user's surroundings (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0161] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0162] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0163] The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0164] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate a user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotion, including estimation and prediction of the user's emotion, but is not limited to these examples. Furthermore, the estimation and prediction of emotion also includes, for example, emotion analysis.

[0165] In the headset type terminal 314, the identification process is performed by the processor 46. A identification program 60 is stored in the storage 50. The processor 46 reads the identification program 60 from the storage 50 and executes the read identification program 60 on the RAM 48. The identification process is realized by the processor 46 operating as a control unit 46A in accordance with the identification program 60 executed on the RAM 48. Note that the headset type terminal 314 has a data generation model and an emotion identification model similar to the data generation model 58 and the emotion identification model 59, and can also perform processing similar to that of the identification processing unit 290 using these models.

[0166] Note that a device other than the data processing device 12 may 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 communicates with the server device having the data generation model 58 to obtain a processing result (such as a prediction result) using the data generation model 58. Furthermore, the data processing device 12 may be a server device, or may be a terminal device (for example, a mobile phone, a robot, a home appliance, etc.) owned by a user.

[0167] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0168] The data generation model 58 is a so-called generative AI. An example of the 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 receives a prompt including an instruction, as well as inference data such as audio data indicating speech, text data indicating text, and image data indicating an image (e.g., still image data or video data). The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in one or more data formats, such as audio data, text data, and image data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The identification processing unit 290 performs the above-mentioned identification processing using the data generation model 58. The data generation model 58 may be a fine-tuned model so as to output an inference result from a prompt that does not include an instruction. In this case, the data generation model 58 can output an inference result from a prompt that does not include an instruction. The data processing device 12 and the like include multiple types of data generation models 58, and the data generation models 58 include AIs other than the generation AI. Examples of AIs other than the generation AI include, but are not limited to, 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), and naive Bayes. These AIs can perform various types of processing, but are not limited to these examples. The AI ​​may also be an AI agent. When the processing of each of the above-described parts is performed by an AI, the processing may be performed in part or entirely by the AI, but is not limited to these examples. Processing performed by an AI, including the generation AI, may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by an AI, including the generation AI.

[0169] 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 executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the headset type terminal 314, but may also be executed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the headset type terminal 314. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the headset type terminal 314 or an external device, etc., and the headset type terminal 314 acquires or collects information required for processing from the data processing device 12 or an external device, etc.

[0170] The correspondence between each part and the device or control part is not limited to the example described above, and various modifications are possible.

[0171] [Fourth embodiment] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[0172] 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.

[0173] 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, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN and / or a LAN.

[0174] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[0175] The microphone 238 receives instructions and the like from the user by receiving voice uttered by the user. The microphone 238 captures the voice uttered by the user, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to instructions from the processor 46.

[0176] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS image sensor or a CCD image sensor, and captures images of the user's surroundings (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0177] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0178] The control 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 emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[0179] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0180] The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0181] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate a user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotion, including estimation and prediction of the user's emotion, but is not limited to these examples. Furthermore, the estimation and prediction of emotion also includes, for example, emotion analysis.

[0182] In the robot 414, the processor 46 performs the identification process. The storage 50 stores the identification program 60. The processor 46 reads the identification program 60 from the storage 50 and executes the read identification program 60 on the RAM 48. The identification process is realized by the processor 46 operating as the control unit 46A in accordance with the identification program 60 executed on the RAM 48. The robot 414 also has a data generation model and an emotion identification model similar to the data generation model 58 and the emotion identification model 59, and can perform the same process as the identification processing unit 290 using these models.

[0183] Note that a device other than the data processing device 12 may 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 communicates with the server device having the data generation model 58 to obtain a processing result (such as a prediction result) using the data generation model 58. Furthermore, the data processing device 12 may be a server device, or may be a terminal device (for example, a mobile phone, a robot, a home appliance, etc.) owned by a user.

[0184] 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 control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[0185] The data generation model 58 is a so-called generative AI. An example of the 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 receives a prompt including an instruction, as well as inference data such as audio data indicating speech, text data indicating text, and image data indicating an image (e.g., still image data or video data). The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in one or more data formats, such as audio data, text data, and image data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The identification processing unit 290 performs the above-mentioned identification processing using the data generation model 58. The data generation model 58 may be a fine-tuned model so as to output an inference result from a prompt that does not include an instruction. In this case, the data generation model 58 can output an inference result from a prompt that does not include an instruction. The data processing device 12 and the like include multiple types of data generation models 58, and the data generation models 58 include AIs other than the generation AI. Examples of AIs other than the generation AI include, but are not limited to, 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), and naive Bayes. These AIs can perform various types of processing, but are not limited to these examples. The AI ​​may also be an AI agent. When the processing of each of the above-described parts is performed by an AI, the processing may be performed in part or entirely by the AI, but is not limited to these examples. Processing performed by an AI, including the generation AI, may be replaced with rule-based processing, and rule-based processing may be replaced with processing performed by an AI, including the generation AI.

[0186] 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 executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but may also be executed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the robot 414 or an external device, etc., and the robot 414 acquires or collects information required for processing from the data processing device 12 or an external device, etc.

[0187] The correspondence between each part and the device or control part is not limited to the example described above, and various modifications are possible.

[0188] The emotion identification model 59 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 an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0189] FIG. 9 illustrates an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and behaviors arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion encompasses both emotions and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[0190] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[0191] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[0192] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is expressed, and when they approach the ideal, a state of pleasure is expressed. Emotions can also be created for robots, cars, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is expressed, and when they approach the ideal, a state of pleasure is expressed. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on speech emotion recognition and brain physiological signal analysis systems for emotions, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the area called "reaction," where sensation is dominant. The right half of the emotion map lists emotions belonging to the area called "situation," where situational awareness is dominant.

[0193] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs 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 a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[0194] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[0195] In the above embodiment, an example was given in which a specific process is performed by one computer 22, but the technology disclosed herein is not limited to this, and distributed processing of the specific process may be performed by multiple computers including computer 22.

[0196] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[0197] 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.

[0198] It is not necessary to store all 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 all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[0199] The hardware resource for executing a specific process can be any of the following types of processors: A processor, for example, is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. A processor also includes a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[0200] The hardware resource that executes the specific process may be configured with one of these various processors, or may be configured with 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). Also, the hardware resource that executes the specific process may be a single processor.

[0201] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[0202] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[0203] In the above example, the first to fourth embodiments have been described separately, but some or all of these embodiments may be combined. The smart device 14, smart glasses 214, headset terminal 314, and robot 414 are merely examples, and they may be combined, or other devices may be used. In the above example, the first and second embodiments have been described separately, but they may be combined.

[0204] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[0205] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[0206] [Explanation of symbols]

[0207] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot

Claims

1. a reception unit that receives text input from a user; an analysis unit that analyzes the text received by the reception unit and detects offensive words; a generator that generates candidates for replacing the unpleasant words detected by the analyzer with appropriate words; a presentation unit that presents the candidates generated by the generation unit to a user; a learning unit that learns a word selected by a user from the candidates presented by the presentation unit. A system characterized by:

2. Equipped with a context analysis unit that generates paraphrase candidates based on context 2. The system of claim 1.

3. A history storage unit is provided to store the user's selection history.

2. The system of claim 1.

4. A feedback receiving unit is provided to receive user feedback.

2. The system of claim 1.

5. The generation unit Generate paraphrase candidates based on context 2. The system of claim 1.

6. The learning unit Save the user's selection history and reflect it in the next candidate generation.

2. The system of claim 1.

7. The reception unit Estimate the user's emotions and adjust the timing of accepting text input based on the estimated user emotions.

2. The system of claim 1.

8. The reception unit Analyze the user's past text input history and select the reception method 2. The system of claim 1.

Images