system

Abstract

An object of a system according to an embodiment is to select an LLM model that matches the difficulty level of an input task of a user.SOLUTION: A system includes a reception unit, an estimation unit, and a selection unit. The reception unit receives an input task of a user. The estimation unit analyzes the content of the task received by the reception unit and estimates the degree of difficulty. The selection unit selects an LLM model based on the difficulty level estimated by the estimation unit and based on a criterion.SELECTED DRAWING: Figure 1

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] In conventional technologies, users tend to use the highest performance LLM regardless of the difficulty of the task, which can result in excessive power consumption.

[0005] The system according to the embodiment aims to select an LLM model that matches the difficulty of the user's input task. [Means for solving the problem]

[0006] The system according to the embodiment includes a receiving unit, an estimating unit, and a selecting unit. The receiving unit receives a task input from a user. The estimating unit analyzes the content of the task received by the receiving unit and estimates the difficulty level. The selecting unit selects an LLM model based on a criterion based on the difficulty level estimated by the estimating unit. [Effects of the Invention]

[0007] The system according to the embodiment can select an LLM model that matches the difficulty of the user's input task. [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 according to an embodiment of the present invention estimates the difficulty of a user's input task and selects an LLM model appropriate for that difficulty. In this system, a user inputs a task, the system estimates the difficulty of the input task, and selects an optimal LLM model based on the estimated difficulty. This reduces excessive power consumption and contributes to the realization of a society in which LLMs and humans coexist. For example, if a user inputs "Please answer a simple question," the system determines that the task is low in difficulty and selects an LLM model with low power consumption. This allows the user to use an LLM model appropriate for their purpose. For example, an LLM model with low power consumption is used for simple questions, and an LLM model with high performance is used for complex tasks. This optimizes the power consumption of LLMs and contributes to the realization of a society in which LLMs and humans coexist.

[0029] A task difficulty estimation system according to an embodiment includes a receiving unit, an estimation unit, and a selection unit. The receiving unit receives a task input from a user. For example, if the user inputs the task in text format, the receiving unit receives the text. The receiving unit can also receive voice input and image input. The estimation unit analyzes the content of the task received by the receiving unit and estimates the difficulty level. For example, the estimation unit analyzes the content of the task using natural language processing technology and estimates the difficulty level using a machine learning model based on past data. The estimation unit performs, for example, morphological analysis, grammatical analysis, and semantic analysis to evaluate the complexity, processing time, and required resources of the task. The selection unit selects an optimal LLM model based on the difficulty level estimated by the estimation unit. For example, the selection unit selects an LLM model taking into account a balance between power consumption and performance. The selection unit uses, for example, a method of trying models in order of lowest power consumption or a method of selecting based on predefined criteria. As a result, the task difficulty estimation system according to an embodiment can reduce excessive power consumption by selecting an optimal LLM model according to the difficulty level of the user's input task. For example, the selection unit can select an LLM model using an AI model that takes the difficulty estimated by the estimation unit as input and outputs the optimal LLM model.

[0030] The estimation unit can analyze the content of the task using natural language processing technology and estimate the difficulty level using a machine learning model based on past data. Natural language processing technology includes, for example, morphological analysis, grammatical analysis, and semantic analysis. For example, the estimation unit can use morphological analysis to break down the content of the task and analyze the meaning of each word. The estimation unit can also use grammatical analysis to analyze the sentence structure of the task and understand the meaning of the sentence. The estimation unit can also use semantic analysis to grasp the overall meaning of the task and estimate the difficulty level. Machine learning models include, for example, neural networks and support vector machines. For example, the estimation unit can use a neural network to learn past data and estimate the difficulty level of the task. The estimation unit can also use a support vector machine to extract task features and estimate the difficulty level. In this way, the use of natural language processing technology and machine learning models allows for accurate estimation of task difficulty. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without AI. For example, the estimation unit can estimate the difficulty level using an AI model that inputs the content of the task and outputs the difficulty level.

[0031] The selection unit can select an LLM model based on criteria that considers the balance between power consumption and performance. Power consumption evaluation criteria include, for example, wattage and energy efficiency. The selection unit can evaluate the power consumption of the LLM model based on, for example, wattage. The selection unit can also evaluate the power consumption of the LLM model based on energy efficiency. Performance evaluation criteria include, for example, processing speed, response time, and throughput. The selection unit can evaluate the performance of the LLM model based on, for example, processing speed. The selection unit can also evaluate the performance of the LLM model based on response time. The selection unit can also evaluate the performance of the LLM model based on throughput. This enables efficient selection of an LLM model by considering the balance between power consumption and performance. Some or all of the above-described processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can select an LLM model using an AI model that inputs the evaluation criteria for power consumption and performance and outputs an optimal LLM model.

[0032] The selection unit may use a method of trying models in order of lowest power consumption or a method of selecting models based on predefined criteria. Criteria for low-power models include, for example, a power consumption threshold or a model with high energy efficiency. The selection unit may select models based on, for example, a power consumption threshold and try models in order of lowest power consumption. The selection unit may also preferentially select models with high energy efficiency. Predefined criteria may include, for example, power consumption, performance, and response time. The selection unit may select models based on, for example, a power consumption criterion. The selection unit may also select models based on a performance criterion. The selection unit may also select models based on a response time criterion. This enables efficient model selection by trying models in order of lowest power consumption. Some or all of the above-described processing in the selection unit may be performed using, for example, AI, or may be performed without AI. For example, the selection unit may select an LLM model using an AI model that inputs a power consumption threshold or a performance criterion and outputs an optimal LLM model.

[0033] The reception unit can provide criteria for determining which LLM model is appropriate depending on the type and content of the task input by the user. Examples of the type and content of the task include text processing, image processing, and speech processing. For example, the reception unit can propose an LLM model suitable for text processing for a text processing task. Furthermore, the reception unit can also propose an LLM model suitable for image processing for an image processing task. Furthermore, the reception unit can also propose an LLM model suitable for speech processing for a speech processing task. By providing criteria according to the type and content of the task, it becomes possible to select the optimal LLM model. Some or all of the above-described processing in the reception unit may be performed using, for example, AI, or may be performed without using AI. For example, the reception unit can provide the criteria using an AI model that inputs the type and content of the task and outputs the optimal LLM model.

[0034] The reception unit can analyze the user's past task input history and select the optimal reception method. For example, the reception unit can automatically display tasks that the user has frequently input in the past as candidates. The reception unit can also prioritize and suggest input methods (voice, text, etc.) that the user has used in the past. The reception unit can also predict and suggest tasks to be used in a specific time period based on the user's past input history. In this way, by analyzing the past task input history, the optimal reception method can be provided to the user. Some or all of the above-mentioned processing in the reception unit may be performed using, or without, AI, for example. For example, the reception unit can input the user's past input history data into a generation AI and have the generation AI select the optimal reception method.

[0035] When receiving a task, the reception unit can filter the tasks based on the user's current project or area of ​​interest. For example, the reception unit can preferentially display tasks related to the user's current project. The reception unit can also filter and suggest related tasks based on the user's area of ​​interest. The reception unit can also suggest related tasks by referring to the user's past project history. In this way, by filtering based on the current project or area of ​​interest, highly relevant tasks can be preferentially accepted. Some or all of the above-described processing in the reception unit may be performed using AI, for example, or may be performed without using AI. For example, the reception unit can input the user's project data into a generation AI and cause the generation AI to filter related tasks.

[0036] When accepting a task, the acceptance unit can select the optimal acceptance means depending on the user's input method. For example, if the user inputs the task by voice, the acceptance unit accepts the task using voice recognition technology. Furthermore, if the user inputs the task using text, the acceptance unit can also accept the task using text analysis technology. Furthermore, if the user inputs the task using an image, the acceptance unit can also accept the task using image recognition technology. By selecting an acceptance means depending on the user's input method, the acceptance of tasks becomes more efficient. Some or all of the above-described processing in the acceptance unit may be performed using, for example, AI, or may be performed without using AI. For example, the acceptance unit can input voice data, text data, and image data into a generation AI and have the generation AI select the optimal acceptance means.

[0037] When accepting a task, the reception unit can prioritize accepting highly relevant tasks by taking into account the user's geographical location information. For example, the reception unit prioritizes accepting tasks to be performed in a location close to the user's current location. The reception unit can also suggest related tasks based on the user's geographical location information. Furthermore, if the user is in a specific area, the reception unit can also prioritize accepting tasks related to that area. In this way, highly relevant tasks can be prioritized by taking the geographical location information into consideration. Some or all of the above-described processing in the reception unit may be performed using AI, for example, or may be performed without using AI. For example, the reception unit can input the user's location information data into a generation AI and cause the generation AI to suggest related tasks.

[0038] When accepting a task, the reception unit can analyze the user's social media activity and accept related tasks. For example, the reception unit can preferentially accept tasks mentioned by the user on social media. The reception unit can also analyze the user's social media activity and suggest related tasks. The reception unit can also accept related tasks with reference to the activity of the user's friends on social media. In this way, by analyzing social media activity, highly relevant tasks can be preferentially accepted. Some or all of the above-described processing in the reception unit may be performed using, for example, AI, or may be performed without using AI. For example, the reception unit can input the user's social media data into a generation AI and have the generation AI suggest related tasks.

[0039] When accepting a task, the reception unit can customize the reception method by reflecting the user's past feedback. The reception unit, for example, suggests an optimal reception method based on feedback provided by the user in the past. The reception unit can also customize the reception interface by reflecting the user's past feedback. The reception unit can also optimize the reception procedure by referring to the user's past feedback. In this way, by reflecting the past feedback, the optimal reception method can be provided to the user. Some or all of the above-described processing in the reception unit may be performed using, for example, AI, or may be performed without using AI. For example, the reception unit can input the user's feedback data into a generation AI and have the generation AI customize the reception method.

[0040] When estimating the difficulty of a task, the estimation unit can adjust the level of detail of the estimation based on the importance of the task. For example, the estimation unit performs a detailed analysis on a task of high importance to accurately estimate the difficulty. The estimation unit can also perform a simple analysis on a task of low importance to quickly estimate the difficulty. The estimation unit can also perform an analysis with an appropriate level of detail on a task of medium importance to estimate the difficulty. This enables efficient difficulty estimation by adjusting the level of detail of the estimation based on the importance of the task. Some or all of the above-described processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input task importance data to a generation AI and cause the generation AI to adjust the level of detail of the estimation.

[0041] When estimating the difficulty of a task, the estimation unit can apply different estimation algorithms depending on the task category. For example, the estimation unit estimates the difficulty by applying a specialized algorithm to a technical task. The estimation unit can also estimate the difficulty by applying an algorithm that evaluates creativity to a creative task. The estimation unit can also estimate the difficulty by applying an algorithm that evaluates efficiency to an administrative task. This enables more accurate difficulty estimation by applying an estimation algorithm depending on the task category. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input task category data to a generation AI and cause the generation AI to apply an estimation algorithm.

[0042] The estimation unit can improve the accuracy of the estimation when estimating the difficulty of a task by referring to the user's past estimation results. For example, the estimation unit estimates the difficulty of the current task based on the difficulty estimation results of tasks previously performed by the user. The estimation unit can also analyze the user's past estimation results and optimize the estimation algorithm. The estimation unit can also improve the accuracy of the estimation by referring to the user's past estimation results. In this way, the accuracy of the estimation is improved by referring to the past estimation results. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input past estimation result data into the generation AI and cause the generation AI to improve the accuracy of the estimation.

[0043] When estimating the difficulty of a task, the estimation unit can determine the estimation priority based on the submission time of the task. For example, the estimation unit prioritizes tasks with an upcoming deadline. The estimation unit can also postpone tasks with a distant submission time. The estimation unit can also moderately prioritize tasks with a medium submission time. This enables efficient difficulty estimation by determining the estimation priority based on the submission time. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input task submission time data into the generation AI and have the generation AI determine the estimation priority.

[0044] The estimation unit can adjust the estimation order based on the relevance of the tasks when estimating the difficulty of the tasks. For example, the estimation unit prioritizes highly relevant tasks. The estimation unit can also postpone tasks with low relevance. The estimation unit can also moderately prioritize tasks with medium relevance. This enables efficient difficulty estimation by adjusting the estimation order based on relevance. Some or all of the above-described processing in the estimation unit may be performed using AI, for example, or may be performed without using AI. For example, the estimation unit can input task relevance data to a generation AI and cause the generation AI to adjust the estimation order.

[0045] When estimating the difficulty of a task, the estimation unit can adjust the use of technical terminology in the estimation according to the user's level of expertise. For example, the estimation unit can provide an estimation result that uses a lot of technical terminology for a user with high level of expertise. The estimation unit can also provide an estimation result in simple language for a user with low level of expertise. The estimation unit can also provide an estimation result using appropriate technical terminology for a user with intermediate level of expertise. In this way, by adjusting the use of technical terminology according to the level of expertise, it is possible to provide an estimation result that is suitable for the user. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input the user's level of expertise data into the generation AI and cause the generation AI to adjust the use of technical terminology.

[0046] When selecting an LLM model, the selection unit can adjust the level of detail of the selection based on the model's importance. For example, the selection unit may perform a detailed analysis for a task with high importance and select an optimal LLM model. The selection unit may also perform a simple analysis for a task with low importance and quickly select an LLM model. The selection unit may also perform an analysis with an appropriate level of detail for a task with medium importance and select an LLM model. This enables efficient model selection by adjusting the level of detail of the selection based on the model's importance. Some or all of the above-described processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit may input model importance data to a generation AI and cause the generation AI to adjust the level of detail of the selection.

[0047] When selecting an LLM model, the selection unit can apply different selection algorithms depending on the model category. For example, the selection unit can apply a specialized algorithm to select an LLM model for a technical task. The selection unit can also apply an algorithm that evaluates creativity to select an LLM model for a creative task. The selection unit can also apply an algorithm that evaluates efficiency to select an LLM model for an administrative task. This allows for more accurate model selection by applying a selection algorithm depending on the model category. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input model category data to the generation AI and cause the generation AI to apply the selection algorithm.

[0048] When selecting an LLM model, the selection unit can improve the accuracy of the selection by referring to the user's past selection results. For example, the selection unit selects the LLM model that is optimal for the current task based on the results of LLM models previously selected by the user. The selection unit can also analyze the user's past selection results and optimize the selection algorithm. The selection unit can also improve the accuracy of the selection by referring to the user's past selection results. In this way, the accuracy of the selection is improved by referring to the past selection results. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input past selection result data into the generation AI and cause the generation AI to improve the accuracy of the selection.

[0049] When selecting an LLM model, the selection unit can determine the selection priority based on the submission date of the model. For example, the selection unit prioritizes the selection of tasks with upcoming deadlines. The selection unit can also postpone tasks with distant submission dates. The selection unit can also moderately prioritize tasks with medium submission dates. This enables efficient model selection by determining the selection priority based on the submission date. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input task submission date data into the generation AI and have the generation AI determine the selection priority.

[0050] When selecting an LLM model, the selection unit can adjust the selection order based on the relevance of the model. For example, the selection unit preferentially selects highly relevant tasks. The selection unit can also postpone tasks with low relevance. The selection unit can also moderately prioritize tasks with medium relevance. This enables efficient model selection by adjusting the selection order based on relevance. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input task relevance data to the generation AI and cause the generation AI to adjust the selection order.

[0051] When selecting an LLM model, the selection unit can adjust the use of technical terminology in the selection according to the user's level of expertise. For example, the selection unit can provide selection results using a lot of technical terminology to a user with high level of expertise. The selection unit can also provide selection results in simple language to a user with low level of expertise. The selection unit can also provide selection results using appropriate technical terminology to a user with medium level of expertise. In this way, by adjusting the use of technical terminology according to the level of expertise, selection results suitable for the user can be provided. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input the user's level of expertise data into the generation AI and cause the generation AI to adjust the use of technical terminology.

[0052] The system according to the embodiment is not limited to the above-described example, and various modifications are possible, for example, as follows.

[0053] The reception unit can dynamically change the priority of tasks based on the content of the tasks input by the user. For example, if the user inputs a highly urgent task, the reception unit processes that task with priority over other tasks. Also, if the user inputs a task related to a long-term project, the reception unit can postpone that task. Furthermore, if the user inputs multiple tasks simultaneously, the reception unit can set priorities based on the importance and urgency of each task. This enables efficient task management according to the content of the tasks input by the user.

[0054] The estimation unit can estimate not only the difficulty of a task but also the required time for the task based on the content of the task input by the user. For example, the estimation unit analyzes the content of the task input by the user and estimates the time required to complete the task. The estimation unit can also make an estimate based on past data and the required time for similar tasks. Furthermore, the estimation unit can provide appropriate advice to the user by taking into account both the difficulty and required time of the task. This allows the user to understand not only the difficulty of the task but also the required time, enabling efficient task management.

[0055] The selector can not only select an LLM model but also divide the task based on the content of the user's input task. For example, if the task entered by the user is very complex, the selector can divide the task into multiple smaller tasks and select an appropriate LLM model for each. The selector can also shorten the processing time for each task by dividing the task. Furthermore, the selector can monitor the progress of each task after dividing the task and perform re-division or re-selection as necessary. This enables efficient processing of complex tasks.

[0056] The reception unit can dynamically change the priority of tasks based on the content of the tasks input by the user. For example, if the user inputs a highly urgent task, the reception unit processes that task with priority over other tasks. Also, if the user inputs a task related to a long-term project, the reception unit can postpone that task. Furthermore, if the user inputs multiple tasks simultaneously, the reception unit can set priorities based on the importance and urgency of each task. This enables efficient task management according to the content of the tasks input by the user.

[0057] The reception unit can dynamically change the priority of tasks based on the content of the tasks input by the user. For example, if the user inputs a highly urgent task, the reception unit processes that task with priority over other tasks. Also, if the user inputs a task related to a long-term project, the reception unit can postpone that task. Furthermore, if the user inputs multiple tasks simultaneously, the reception unit can set priorities based on the importance and urgency of each task. This enables efficient task management according to the content of the tasks input by the user.

[0058] The processing flow of the first embodiment will be briefly explained below.

[0059] Step 1: The reception unit receives a task input from the user. For example, if the user inputs a task in text format, the reception unit receives the text. The reception unit can also receive voice input and image input. Step 2: The estimation unit analyzes the content of the task received by the reception unit and estimates the difficulty. For example, the estimation unit analyzes the content of the task using natural language processing technology and estimates the difficulty using a machine learning model based on past data. The estimation unit performs morphological analysis, grammatical analysis, and semantic analysis to evaluate the complexity, processing time, and required resources of the task. Step 3: The selection unit selects the optimal LLM model based on the difficulty estimated by the estimation unit. For example, the selection unit selects an LLM model taking into account the balance between power consumption and performance. The selection unit uses a method of trying models in order of power consumption, or a method of selecting based on predefined criteria. This makes it possible to reduce excessive power consumption by selecting the optimal LLM model according to the difficulty of the user's input task. The selection unit can select an LLM model using an AI model that inputs the difficulty estimated by the estimation unit and outputs the optimal LLM model.

[0060] (Example 2) A system according to an embodiment of the present invention estimates the difficulty of a user's input task and selects an LLM model appropriate for that difficulty. In this system, a user inputs a task, the system estimates the difficulty of the input task, and selects an optimal LLM model based on the estimated difficulty. This reduces excessive power consumption and contributes to the realization of a society in which LLMs and humans coexist. For example, if a user inputs "Please answer a simple question," the system determines that the task is low in difficulty and selects an LLM model with low power consumption. This allows the user to use an LLM model appropriate for their purpose. For example, an LLM model with low power consumption is used for simple questions, and an LLM model with high performance is used for complex tasks. This optimizes the power consumption of LLMs and contributes to the realization of a society in which LLMs and humans coexist.

[0061] A task difficulty estimation system according to an embodiment includes a receiving unit, an estimation unit, and a selection unit. The receiving unit receives a task input from a user. For example, if the user inputs the task in text format, the receiving unit receives the text. The receiving unit can also receive voice input and image input. The estimation unit analyzes the content of the task received by the receiving unit and estimates the difficulty level. For example, the estimation unit analyzes the content of the task using natural language processing technology and estimates the difficulty level using a machine learning model based on past data. The estimation unit performs, for example, morphological analysis, grammatical analysis, and semantic analysis to evaluate the complexity, processing time, and required resources of the task. The selection unit selects an optimal LLM model based on the difficulty level estimated by the estimation unit. For example, the selection unit selects an LLM model taking into account a balance between power consumption and performance. The selection unit uses, for example, a method of trying models in order of lowest power consumption or a method of selecting based on predefined criteria. As a result, the task difficulty estimation system according to an embodiment can reduce excessive power consumption by selecting an optimal LLM model according to the difficulty level of the user's input task. For example, the selection unit can select an LLM model using an AI model that takes the difficulty estimated by the estimation unit as input and outputs the optimal LLM model.

[0062] The estimation unit can analyze the content of the task using natural language processing technology and estimate the difficulty level using a machine learning model based on past data. Natural language processing technology includes, for example, morphological analysis, grammatical analysis, and semantic analysis. For example, the estimation unit can use morphological analysis to break down the content of the task and analyze the meaning of each word. The estimation unit can also use grammatical analysis to analyze the sentence structure of the task and understand the meaning of the sentence. The estimation unit can also use semantic analysis to grasp the overall meaning of the task and estimate the difficulty level. Machine learning models include, for example, neural networks and support vector machines. For example, the estimation unit can use a neural network to learn past data and estimate the difficulty level of the task. The estimation unit can also use a support vector machine to extract task features and estimate the difficulty level. In this way, the use of natural language processing technology and machine learning models allows for accurate estimation of task difficulty. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without AI. For example, the estimation unit can estimate the difficulty level using an AI model that inputs the content of the task and outputs the difficulty level.

[0063] The selection unit can select an LLM model based on criteria that considers the balance between power consumption and performance. Power consumption evaluation criteria include, for example, wattage and energy efficiency. The selection unit can evaluate the power consumption of the LLM model based on, for example, wattage. The selection unit can also evaluate the power consumption of the LLM model based on energy efficiency. Performance evaluation criteria include, for example, processing speed, response time, and throughput. The selection unit can evaluate the performance of the LLM model based on, for example, processing speed. The selection unit can also evaluate the performance of the LLM model based on response time. The selection unit can also evaluate the performance of the LLM model based on throughput. This enables efficient selection of an LLM model by considering the balance between power consumption and performance. Some or all of the above-described processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can select an LLM model using an AI model that inputs the evaluation criteria for power consumption and performance and outputs an optimal LLM model.

[0064] The selection unit may use a method of trying models in order of lowest power consumption or a method of selecting models based on predefined criteria. Criteria for low-power models include, for example, a power consumption threshold or a model with high energy efficiency. The selection unit may select models based on, for example, a power consumption threshold and try models in order of lowest power consumption. The selection unit may also preferentially select models with high energy efficiency. Predefined criteria may include, for example, power consumption, performance, and response time. The selection unit may select models based on, for example, a power consumption criterion. The selection unit may also select models based on a performance criterion. The selection unit may also select models based on a response time criterion. This enables efficient model selection by trying models in order of lowest power consumption. Some or all of the above-described processing in the selection unit may be performed using, for example, AI, or may be performed without AI. For example, the selection unit may select an LLM model using an AI model that inputs a power consumption threshold or a performance criterion and outputs an optimal LLM model.

[0065] The reception unit can provide criteria for determining which LLM model is appropriate depending on the type and content of the task input by the user. Examples of the type and content of the task include text processing, image processing, and speech processing. For example, the reception unit can propose an LLM model suitable for text processing for a text processing task. Furthermore, the reception unit can also propose an LLM model suitable for image processing for an image processing task. Furthermore, the reception unit can also propose an LLM model suitable for speech processing for a speech processing task. By providing criteria according to the type and content of the task, it becomes possible to select the optimal LLM model. Some or all of the above-described processing in the reception unit may be performed using, for example, AI, or may be performed without using AI. For example, the reception unit can provide the criteria using an AI model that inputs the type and content of the task and outputs the optimal LLM model.

[0066] The reception unit can estimate the user's emotions and change the task reception method based on the estimated user emotions. For example, if the user is feeling stressed, the reception unit can provide a simple interface and minimize input steps. Furthermore, if the user is relaxed, the reception unit can provide detailed input options and suggest customizable input methods. Furthermore, if the user is in a hurry, the reception unit can prioritize voice input and enable quick task input. This improves the user experience by adjusting the task reception method according to the user's emotions. Emotion estimation is realized using an emotion estimation function, for example, 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. Some or all of the above-described processing in the reception unit can be performed using AI, for example, or without AI. For example, the reception unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0067] The reception unit can analyze the user's past task input history and select the optimal reception method. For example, the reception unit can automatically display tasks that the user has frequently input in the past as candidates. The reception unit can also prioritize and suggest input methods (voice, text, etc.) that the user has used in the past. The reception unit can also predict and suggest tasks to be used in a specific time period based on the user's past input history. In this way, by analyzing the past task input history, the optimal reception method can be provided to the user. Some or all of the above-mentioned processing in the reception unit may be performed using, or without, AI, for example. For example, the reception unit can input the user's past input history data into a generation AI and have the generation AI select the optimal reception method.

[0068] When receiving a task, the reception unit can filter the tasks based on the user's current project or area of ​​interest. For example, the reception unit can preferentially display tasks related to the user's current project. The reception unit can also filter and suggest related tasks based on the user's area of ​​interest. The reception unit can also suggest related tasks by referring to the user's past project history. In this way, by filtering based on the current project or area of ​​interest, highly relevant tasks can be preferentially accepted. Some or all of the above-described processing in the reception unit may be performed using AI, for example, or may be performed without using AI. For example, the reception unit can input the user's project data into a generation AI and cause the generation AI to filter related tasks.

[0069] When accepting a task, the acceptance unit can select the optimal acceptance means depending on the user's input method. For example, if the user inputs the task by voice, the acceptance unit accepts the task using voice recognition technology. Furthermore, if the user inputs the task using text, the acceptance unit can also accept the task using text analysis technology. Furthermore, if the user inputs the task using an image, the acceptance unit can also accept the task using image recognition technology. By selecting an acceptance means depending on the user's input method, the acceptance of tasks becomes more efficient. Some or all of the above-described processing in the acceptance unit may be performed using, for example, AI, or may be performed without using AI. For example, the acceptance unit can input voice data, text data, and image data into a generation AI and have the generation AI select the optimal acceptance means.

[0070] The reception unit can estimate the user's emotions and set the priority of tasks to be received based on the estimated user emotions. For example, if the user is feeling stressed, the reception unit can postpone tasks of lower importance. Furthermore, if the user is relaxed, the reception unit can also prioritize tasks of higher importance. Furthermore, if the user is in a hurry, the reception unit can also prioritize tasks of higher urgency. Thus, by determining the priority of tasks according to the user's emotions, important tasks can be processed preferentially. The emotion estimation is realized using an emotion estimation function, 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 such examples. Some or all of the above-described processing in the reception unit may be performed using AI, or may be performed without AI. For example, the reception unit can input the user's facial expression data into the generation AI and have the generation AI execute emotion estimation.

[0071] When accepting a task, the reception unit can prioritize accepting highly relevant tasks by taking into account the user's geographical location information. For example, the reception unit prioritizes accepting tasks to be performed in a location close to the user's current location. The reception unit can also suggest related tasks based on the user's geographical location information. Furthermore, if the user is in a specific area, the reception unit can also prioritize accepting tasks related to that area. In this way, highly relevant tasks can be prioritized by taking the geographical location information into consideration. Some or all of the above-described processing in the reception unit may be performed using AI, for example, or may be performed without using AI. For example, the reception unit can input the user's location information data into a generation AI and cause the generation AI to suggest related tasks.

[0072] When accepting a task, the reception unit can analyze the user's social media activity and accept related tasks. For example, the reception unit can preferentially accept tasks mentioned by the user on social media. The reception unit can also analyze the user's social media activity and suggest related tasks. The reception unit can also accept related tasks with reference to the activity of the user's friends on social media. In this way, by analyzing social media activity, highly relevant tasks can be preferentially accepted. Some or all of the above-described processing in the reception unit may be performed using, for example, AI, or may be performed without using AI. For example, the reception unit can input the user's social media data into a generation AI and have the generation AI suggest related tasks.

[0073] When accepting a task, the reception unit can customize the reception method by reflecting the user's past feedback. The reception unit, for example, suggests an optimal reception method based on feedback provided by the user in the past. The reception unit can also customize the reception interface by reflecting the user's past feedback. The reception unit can also optimize the reception procedure by referring to the user's past feedback. In this way, by reflecting the past feedback, the optimal reception method can be provided to the user. Some or all of the above-described processing in the reception unit may be performed using, for example, AI, or may be performed without using AI. For example, the reception unit can input the user's feedback data into a generation AI and have the generation AI customize the reception method.

[0074] The estimation unit can estimate the user's emotions and change the task difficulty estimation method based on the estimated user emotions. For example, when the user is relaxed, the estimation unit performs a detailed analysis to accurately estimate the difficulty. Furthermore, when the user is in a hurry, the estimation unit can perform a simple analysis to quickly estimate the difficulty. Furthermore, when the user is stressed, the estimation unit can perform a simple analysis to estimate the difficulty. This enables more accurate difficulty estimation by adjusting the difficulty estimation method according to the user's emotions. The emotion estimation is realized using an emotion estimation function, for example, an emotion engine or a generation AI. The generation AI may be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. Some or all of the above-described processing in the estimation unit may be performed using, for example, an AI, or may be performed without using an AI. For example, the estimation unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0075] When estimating the difficulty of a task, the estimation unit can adjust the level of detail of the estimation based on the importance of the task. For example, the estimation unit performs a detailed analysis on a task of high importance to accurately estimate the difficulty. The estimation unit can also perform a simple analysis on a task of low importance to quickly estimate the difficulty. The estimation unit can also perform an analysis with an appropriate level of detail on a task of medium importance to estimate the difficulty. This enables efficient difficulty estimation by adjusting the level of detail of the estimation based on the importance of the task. Some or all of the above-described processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input task importance data to a generation AI and cause the generation AI to adjust the level of detail of the estimation.

[0076] When estimating the difficulty of a task, the estimation unit can apply different estimation algorithms depending on the task category. For example, the estimation unit estimates the difficulty by applying a specialized algorithm to a technical task. The estimation unit can also estimate the difficulty by applying an algorithm that evaluates creativity to a creative task. The estimation unit can also estimate the difficulty by applying an algorithm that evaluates efficiency to an administrative task. This enables more accurate difficulty estimation by applying an estimation algorithm depending on the task category. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input task category data to a generation AI and cause the generation AI to apply an estimation algorithm.

[0077] The estimation unit can improve the accuracy of the estimation when estimating the difficulty of a task by referring to the user's past estimation results. For example, the estimation unit estimates the difficulty of the current task based on the difficulty estimation results of tasks previously performed by the user. The estimation unit can also analyze the user's past estimation results and optimize the estimation algorithm. The estimation unit can also improve the accuracy of the estimation by referring to the user's past estimation results. In this way, the accuracy of the estimation is improved by referring to the past estimation results. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input past estimation result data into the generation AI and cause the generation AI to improve the accuracy of the estimation.

[0078] The estimation unit estimates the user's emotion and can change the length of the estimation based on the estimated user emotion. For example, if the user is in a hurry, the estimation unit completes the estimation in a short time. Furthermore, if the user is relaxed, the estimation unit can perform a detailed analysis and take more time to estimate. Furthermore, if the user is stressed, the estimation unit can perform a simple analysis and quickly complete the estimation. This enables efficient estimation by adjusting the length of the estimation according to the user's emotion. Emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or a generation AI. The generation AI may be a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to such examples. Some or all of the above-described processing in the estimation unit may be performed using, for example, an AI, or may be performed without using an AI. For example, the estimation unit may input the user's facial expression data into the generation AI and cause the generation AI to estimate the emotion.

[0079] When estimating the difficulty of a task, the estimation unit can determine the estimation priority based on the submission time of the task. For example, the estimation unit prioritizes tasks with an upcoming deadline. The estimation unit can also postpone tasks with a distant submission time. The estimation unit can also moderately prioritize tasks with a medium submission time. This enables efficient difficulty estimation by determining the estimation priority based on the submission time. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input task submission time data into the generation AI and have the generation AI determine the estimation priority.

[0080] The estimation unit can adjust the estimation order based on the relevance of the tasks when estimating the difficulty of the tasks. For example, the estimation unit prioritizes highly relevant tasks. The estimation unit can also postpone tasks with low relevance. The estimation unit can also moderately prioritize tasks with medium relevance. This enables efficient difficulty estimation by adjusting the estimation order based on relevance. Some or all of the above-described processing in the estimation unit may be performed using AI, for example, or may be performed without using AI. For example, the estimation unit can input task relevance data to a generation AI and cause the generation AI to adjust the estimation order.

[0081] When estimating the difficulty of a task, the estimation unit can adjust the use of technical terminology in the estimation according to the user's level of expertise. For example, the estimation unit can provide an estimation result that uses a lot of technical terminology for a user with high level of expertise. The estimation unit can also provide an estimation result in simple language for a user with low level of expertise. The estimation unit can also provide an estimation result using appropriate technical terminology for a user with intermediate level of expertise. In this way, by adjusting the use of technical terminology according to the level of expertise, it is possible to provide an estimation result that is suitable for the user. Some or all of the above-mentioned processing in the estimation unit may be performed using, for example, AI, or may be performed without using AI. For example, the estimation unit can input the user's level of expertise data into the generation AI and cause the generation AI to adjust the use of technical terminology.

[0082] The selection unit can estimate the user's emotion and change the LLM model selection method based on the estimated user emotion. For example, when the user is relaxed, the selection unit performs a detailed analysis and selects the optimal LLM model. Furthermore, when the user is in a hurry, the selection unit can also suggest an LLM model that can be quickly selected. Furthermore, when the user is stressed, the selection unit can also provide a simple selection method. This allows the optimal LLM model to be selected by adjusting the selection method according to the user's emotion. Emotion estimation is realized using an emotion estimation function, for example, using an emotion engine or a generation AI. The generation AI may be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. Some or all of the above-described processing in the selection unit may be performed using, for example, an AI, or may be performed without using an AI. For example, the selection unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0083] When selecting an LLM model, the selection unit can adjust the level of detail of the selection based on the model's importance. For example, the selection unit may perform a detailed analysis for a task with high importance and select an optimal LLM model. The selection unit may also perform a simple analysis for a task with low importance and quickly select an LLM model. The selection unit may also perform an analysis with an appropriate level of detail for a task with medium importance and select an LLM model. This enables efficient model selection by adjusting the level of detail of the selection based on the model's importance. Some or all of the above-described processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit may input model importance data to a generation AI and cause the generation AI to adjust the level of detail of the selection.

[0084] When selecting an LLM model, the selection unit can apply different selection algorithms depending on the model category. For example, the selection unit can apply a specialized algorithm to select an LLM model for a technical task. The selection unit can also apply an algorithm that evaluates creativity to select an LLM model for a creative task. The selection unit can also apply an algorithm that evaluates efficiency to select an LLM model for an administrative task. This allows for more accurate model selection by applying a selection algorithm depending on the model category. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input model category data to the generation AI and cause the generation AI to apply the selection algorithm.

[0085] When selecting an LLM model, the selection unit can improve the accuracy of the selection by referring to the user's past selection results. For example, the selection unit selects the LLM model that is optimal for the current task based on the results of LLM models previously selected by the user. The selection unit can also analyze the user's past selection results and optimize the selection algorithm. The selection unit can also improve the accuracy of the selection by referring to the user's past selection results. In this way, the accuracy of the selection is improved by referring to the past selection results. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input past selection result data into the generation AI and cause the generation AI to improve the accuracy of the selection.

[0086] The selection unit can estimate the user's emotion and change the length of the selection of the LLM model based on the estimated user emotion. For example, if the user is in a hurry, the selection unit completes the selection in a short time. Furthermore, if the user is relaxed, the selection unit can perform a detailed analysis and take more time to complete the selection. Furthermore, if the user is stressed, the selection unit can perform a simple analysis and complete the selection quickly. This enables efficient model selection by adjusting the length of the selection according to the user's emotion. Emotion estimation is realized using an emotion estimation function, for example, an emotion engine or a generation AI. The generation AI may be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. Some or all of the above-described processing in the selection unit may be performed using, for example, an AI, or may be performed without using an AI. For example, the selection unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0087] When selecting an LLM model, the selection unit can determine the selection priority based on the submission date of the model. For example, the selection unit prioritizes the selection of tasks with upcoming deadlines. The selection unit can also postpone tasks with distant submission dates. The selection unit can also moderately prioritize tasks with medium submission dates. This enables efficient model selection by determining the selection priority based on the submission date. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input task submission date data into the generation AI and have the generation AI determine the selection priority.

[0088] When selecting an LLM model, the selection unit can adjust the selection order based on the relevance of the model. For example, the selection unit preferentially selects highly relevant tasks. The selection unit can also postpone tasks with low relevance. The selection unit can also moderately prioritize tasks with medium relevance. This enables efficient model selection by adjusting the selection order based on relevance. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input task relevance data to the generation AI and cause the generation AI to adjust the selection order.

[0089] When selecting an LLM model, the selection unit can adjust the use of technical terminology in the selection according to the user's level of expertise. For example, the selection unit can provide selection results using a lot of technical terminology to a user with high level of expertise. The selection unit can also provide selection results in simple language to a user with low level of expertise. The selection unit can also provide selection results using appropriate technical terminology to a user with medium level of expertise. In this way, by adjusting the use of technical terminology according to the level of expertise, selection results suitable for the user can be provided. Some or all of the above-mentioned processing in the selection unit may be performed using, for example, AI, or may be performed without using AI. For example, the selection unit can input the user's level of expertise data into the generation AI and cause the generation AI to adjust the use of technical terminology. === Hard Collateral 1-1 === Each of the multiple elements including the above-described reception unit, estimation unit, and selection unit is realized, for example, by at least one of the smart device 14 and the data processing device 12. For example, the reception unit can receive a user's input task via the reception device 38 of the smart device 14 or the communication I / F 26 of the data processing device 12. For example, the estimation unit is realized by the specific processing unit 290 of the data processing device 12 and analyzes the content of the task received from the reception unit and estimates the difficulty level. For example, the selection unit is realized by the specific processing unit 290 of the data processing device 12 and selects an optimal LLM model based on the difficulty level estimated by the estimation unit. === Hard Collateral 1-2 === Each of the multiple elements including the above-mentioned reception unit, estimation unit, and selection 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 can receive a user's input task via the microphone 238 of the smart glasses 214 or the communication I / F 26 of the data processing device 12. For example, the estimation unit is realized by the specific processing unit 290 of the data processing device 12 and analyzes the content of the task received from the reception unit and estimates the difficulty level. For example, the selection unit is realized by the specific processing unit 290 of the data processing device 12 and selects an optimal LLM model based on the difficulty level estimated by the estimation unit. === Hard Collateral 1-3 === Each of the multiple elements including the above-mentioned reception unit, estimation unit, and selection 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 can receive a task input by the user via the microphone 238 of the headset type terminal 314 or the communication I / F 26 of the data processing device 12. For example, the estimation unit is realized by the specific processing unit 290 of the data processing device 12 and analyzes the content of the task received from the reception unit and estimates the difficulty level. For example, the selection unit is realized by the specific processing unit 290 of the data processing device 12 and selects an optimal LLM model based on the difficulty level estimated by the estimation unit. === Hard Collateral 1-4 === Each of the multiple elements including the above-mentioned reception unit, estimation unit, and selection unit is realized, for example, by at least one of the robot 414 and the data processing device 12. For example, the reception unit can receive a task input by the user via the microphone 238 of the robot 414 or the communication I / F 26 of the data processing device 12. For example, the estimation unit is realized by the specific processing unit 290 of the data processing device 12 and analyzes the content of the task received from the reception unit and estimates the difficulty level. For example, the selection unit is realized by the specific processing unit 290 of the data processing device 12 and selects an optimal LLM model based on the difficulty level estimated by the estimation unit.

[0090] The system according to the embodiment is not limited to the above-described example, and various modifications are possible, for example, as follows.

[0091] The reception unit can dynamically change the priority of tasks based on the content of the tasks input by the user. For example, if the user inputs a highly urgent task, the reception unit processes that task with priority over other tasks. Also, if the user inputs a task related to a long-term project, the reception unit can postpone that task. Furthermore, if the user inputs multiple tasks simultaneously, the reception unit can set priorities based on the importance and urgency of each task. This enables efficient task management according to the content of the tasks input by the user.

[0092] The estimation unit can estimate not only the difficulty of a task but also the required time for the task based on the content of the task input by the user. For example, the estimation unit analyzes the content of the task input by the user and estimates the time required to complete the task. The estimation unit can also make an estimate based on past data and the required time for similar tasks. Furthermore, the estimation unit can provide appropriate advice to the user by taking into account both the difficulty and required time of the task. This allows the user to understand not only the difficulty of the task but also the required time, enabling efficient task management.

[0093] The selector can not only select an LLM model but also divide the task based on the content of the user's input task. For example, if the task entered by the user is very complex, the selector can divide the task into multiple smaller tasks and select an appropriate LLM model for each. The selector can also shorten the processing time for each task by dividing the task. Furthermore, the selector can monitor the progress of each task after dividing the task and perform re-division or re-selection as necessary. This enables efficient processing of complex tasks.

[0094] The reception unit can dynamically change the priority of tasks based on the content of the tasks input by the user. For example, if the user inputs a highly urgent task, the reception unit processes that task with priority over other tasks. Also, if the user inputs a task related to a long-term project, the reception unit can postpone that task. Furthermore, if the user inputs multiple tasks simultaneously, the reception unit can set priorities based on the importance and urgency of each task. This enables efficient task management according to the content of the tasks input by the user.

[0095] The reception unit can dynamically change the priority of tasks based on the content of the tasks input by the user. For example, if the user inputs a highly urgent task, the reception unit processes that task with priority over other tasks. Also, if the user inputs a task related to a long-term project, the reception unit can postpone that task. Furthermore, if the user inputs multiple tasks simultaneously, the reception unit can set priorities based on the importance and urgency of each task. This enables efficient task management according to the content of the tasks input by the user.

[0096] The reception unit can estimate the user's emotions and change the task reception method based on the estimated user emotions. For example, if the user is feeling stressed, the reception unit can provide a simple interface and minimize input steps. Alternatively, if the user is relaxed, the reception unit can provide detailed input options and suggest a customizable input method. Alternatively, if the user is in a hurry, the reception unit can prioritize voice input to enable quick task input. This improves the user experience by adjusting the task reception method according to the user's emotions. Emotion estimation is achieved using an emotion estimation function, 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 such examples. Some or all of the above-described processing in the reception unit may be performed using AI, or may be performed without AI. For example, the reception unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0097] The estimation unit can estimate the user's emotions and change the task difficulty estimation method based on the estimated user emotions. For example, if the user is relaxed, the estimation unit can perform a detailed analysis to accurately estimate the difficulty. If the user is in a hurry, the estimation unit can also perform a simple analysis to quickly estimate the difficulty. If the user is stressed, the estimation unit can also perform a simple analysis to estimate the difficulty. This allows for more accurate difficulty estimation by adjusting the difficulty estimation method according to the user's emotions. Emotion estimation is achieved using an emotion estimation function, such as 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. Some or all of the above-described processing in the estimation unit can be performed using, for example, an AI, or without an AI. For example, the estimation unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0098] The selection unit can estimate the user's emotion and change the selection method of the LLM model based on the estimated user's emotion. For example, if the user is relaxed, the selection unit performs a detailed analysis and selects the optimal LLM model. If the user is in a hurry, the selection unit can also suggest an LLM model that can be quickly selected. If the user is stressed, the selection unit can also provide a simple selection method. This allows the optimal LLM model to be selected by adjusting the selection method according to the user's emotion. Emotion estimation is achieved using an emotion estimation function, for example, an emotion engine or a generation AI. The generation AI may be a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to such examples. Some or all of the above-described processing in the selection unit may be performed using, for example, an AI, or may be performed without using an AI. For example, the selection unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0099] The estimation unit estimates the user's emotion and can change the length of the estimation based on the estimated user emotion. For example, if the user is in a hurry, the estimation unit can complete the estimation in a short time. Alternatively, if the user is relaxed, the estimation unit can perform a detailed analysis and take more time to estimate. Alternatively, if the user is stressed, the estimation unit can perform a simple analysis and quickly complete the estimation. This enables efficient estimation by adjusting the length of the estimation according to the user's emotion. Emotion estimation is realized using an emotion estimation function, for example, an emotion engine or a generation AI. The generation AI may be, but is not limited to, a text generation AI (e.g., LLM) or a multimodal generation AI. Some or all of the above-described processing in the estimation unit may be performed using, for example, an AI, or may be performed without using an AI. For example, the estimation unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0100] The selection unit can estimate the user's emotion and change the length of the selection of the LLM model based on the estimated user emotion. For example, if the user is in a hurry, the selection unit can complete the selection in a short time. Alternatively, if the user is relaxed, the selection unit can perform a detailed analysis and take more time to complete the selection. Alternatively, if the user is stressed, the selection unit can perform a simple analysis and complete the selection quickly. This enables efficient model selection by adjusting the length of the selection according to the user's emotion. Emotion estimation is realized using an emotion estimation function, for example, an emotion engine or a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, but is not limited to such examples. Some or all of the above-mentioned processing in the selection unit can be performed using AI, for example, or without AI. For example, the selection unit can input the user's facial expression data into the generation AI and have the generation AI perform emotion estimation.

[0101] The processing flow of the second embodiment will be briefly explained below.

[0102] Step 1: The reception unit receives a task input from the user. For example, if the user inputs a task in text format, the reception unit receives the text. The reception unit can also receive voice input and image input. Step 2: The estimation unit analyzes the content of the task received by the reception unit and estimates the difficulty. For example, the estimation unit analyzes the content of the task using natural language processing technology and estimates the difficulty using a machine learning model based on past data. The estimation unit performs morphological analysis, grammatical analysis, and semantic analysis to evaluate the complexity, processing time, and required resources of the task. Step 3: The selection unit selects the optimal LLM model based on the difficulty estimated by the estimation unit. For example, the selection unit selects an LLM model taking into account the balance between power consumption and performance. The selection unit uses a method of trying models in order of power consumption, or a method of selecting based on predefined criteria. This makes it possible to reduce excessive power consumption by selecting the optimal LLM model according to the difficulty of the user's input task. The selection unit can select an LLM model using an AI model that inputs the difficulty estimated by the estimation unit and outputs the optimal LLM model.

[0103] 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.

[0104] 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 generative AIs include the data generation model 58, such as a neural network model (e.g., a neural network model), and a neural network model (e.g., a neural network model). The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating speech, text data indicating text, and image data indicating an image is also input to the data generation model 58. 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 a data format such as voice data and text data. 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 can perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processes of each of the above-mentioned parts are performed by AI, the processes may be performed in part or entirely by AI, but are not limited to these examples. The processes performed by AI, including the generative AI, may be replaced with rule-based processes.

[0105] 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.

[0106] The correspondence between each part and the device or control part is not limited to the above example, and various modifications are possible.

[0107] [Second embodiment] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0108] 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0109] 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.

[0110] 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.

[0111] 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.

[0112] 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).

[0113] 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.

[0114] 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.

[0115] 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.

[0116] 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.

[0117] 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.

[0118] 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.

[0119] 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.

[0120] 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 containing an instruction, as well as inference data such as voice data representing speech, text data representing text, and image data representing an image. 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 a data format such as voice data and text 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 model 58 includes AI 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 can perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processes of each of the above-mentioned parts are performed by AI, the processes may be performed in part or entirely by AI, but are not limited to these examples. The processes performed by AI, including the generative AI, may be replaced with rule-based processes.

[0121] 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.

[0122] The correspondence between each part and the device or control part is not limited to the above example, and various modifications are possible.

[0123] [Third embodiment] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0124] 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.

[0125] 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.

[0126] 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.

[0127] 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.

[0128] 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).

[0129] 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.

[0130] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset 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.

[0131] 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.

[0132] 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.

[0133] 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.

[0134] 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.

[0135] 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.

[0136] 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 containing an instruction, as well as inference data such as voice data representing speech, text data representing text, and image data representing an image. 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 a data format such as voice data and text 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 model 58 includes AI 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 can perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processes of each of the above-mentioned parts are performed by AI, the processes may be performed in part or entirely by AI, but are not limited to these examples. The processes performed by AI, including the generative AI, may be replaced with rule-based processes.

[0137] 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.

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

[0139] [Fourth embodiment] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[0140] 7, a 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.

[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 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.

[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 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).

[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] 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.

[0147] 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.

[0148] 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.

[0149] 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.

[0150] 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.

[0151] 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.

[0152] 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.

[0153] 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 containing an instruction, as well as inference data such as voice data representing speech, text data representing text, and image data representing an image. 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 a data format such as voice data and text 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 model 58 includes AI 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 can perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processes of each of the above-mentioned parts are performed by AI, the processes may be performed in part or entirely by AI, but are not limited to these examples. The processes performed by AI, including the generative AI, may be replaced with rule-based processes.

[0154] 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.

[0155] The correspondence between each part and the device or control part is not limited to the above example, and various modifications are possible.

[0156] 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.

[0157] 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.

[0158] 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.

[0159] 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).

[0160] 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.

[0161] 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."

[0162] 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.

[0163] 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.

[0164] 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.

[0165] 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.

[0166] 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.

[0167] 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.

[0168] 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.

[0169] 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.

[0170] 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.

[0171] 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.

[0172] 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, in order to avoid confusion and to 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.

[0173] 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.

[0174] [Explanation of symbols]

[0175] 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 an input task from a user; an estimation unit that analyzes the content of the task accepted by the acceptance unit and estimates the difficulty level of the task; a selection unit that selects an LLM model based on a criterion based on the difficulty estimated by the estimation unit; A system characterized by:

2. The estimation unit Analyze the content of the task using natural language processing technology and estimate the difficulty level using a machine learning model based on past data 2. The system of claim 1.

3. The selection unit Select an LLM model based on criteria that considers the balance between power consumption and performance.

2. The system of claim 1.

4. The selection unit Use a method to try models with the lowest power consumption first or select based on predefined criteria 2. The system of claim 1.

5. The reception unit Provide criteria for determining which LLM model is appropriate depending on the type and content of the task the user inputs.

2. The system of claim 1.

6. The reception unit Estimate the user's emotions and change the task acceptance method based on the estimated user emotions.

2. The system of claim 1.

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

8. The reception unit Filter tasks based on your current projects and interests when accepting them 2. The system of claim 1.

Images