Information processing terminal, robot, information processing system, information processing method, control program, and recording medium

Abstract

To provide a configuration that allows determination of whether to offload a task.SOLUTION: An information processing terminal (12) includes: an acquisition unit (16) that obtains an information processing task; a determination unit (17) that determines whether to offload the information processing task to an external device (30) by using a learning model to which information about the information processing task is input; a communication processing unit (18) that offloads, when the determination unit (17) determines that the information processing task is to be offloaded to the external device (30), the information processing task to the external device (30); and a task processing unit (19) that executes, when the determination unit (17) determines that the information processing task is not to be offloaded to the external device (30), the information processing task.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an information processing terminal, a robot, an information processing system, an information processing method, a control program, and a recording medium. [Background technology]

[0002] There is known a technique for offloading a task occurring in a certain device to an external device such as a server, etc. Patent Document 1 discloses a portable camera that offloads image processing work. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-082773 Summary of the Invention [Problem to be solved by the invention]

[0004] In the conventional technology described above, in a configuration where offloading from an information processing terminal to an external device is possible, it is sometimes unclear whether a task is of a level of difficulty that can be executed by the information processing terminal. However, if all tasks are offloaded to an external device, the resources of the external device may be unnecessarily occupied, which may lead to, for example, an increase in costs.

[0005] One aspect of the present invention has been made in view of the above-mentioned problems, and aims to realize a configuration capable of determining whether or not a task should be offloaded. [Means for solving the problem]

[0006] In order to solve the above problem, an information processing terminal according to one embodiment of the present invention comprises an acquisition unit that acquires an information processing task, a judgment unit that determines whether or not the information processing task should be offloaded to an external device using a learning model to which information related to the information processing task is input, a communication processing unit that offloads the information processing task to the external device when the judgment unit determines that the information processing task should be offloaded to the external device, and a task processing unit that executes the information processing task when the judgment unit determines that the information processing task should not be offloaded to the external device.

[0007] Another aspect of the present invention provides an information processing system comprising an information processing terminal and a server device, wherein the information processing terminal comprises an acquisition unit that acquires an information processing task, a judgment unit that determines whether or not the information processing task should be offloaded to the server device using a learning model to which information related to the information processing task is input, a communication processing unit that offloads the information processing task to the server device when the judgment unit determines that the information processing task should be offloaded to the server device, and an information processing unit that executes the information processing task when the judgment unit determines that the information processing task should not be offloaded to the server device, and the server device executes the information processing task when the information processing task is offloaded from the information processing terminal and transmits the execution results to the information processing terminal.

[0008] An information processing method according to another aspect of the present invention is an information processing method executed by an apparatus, and includes an acquisition step of acquiring an information processing task, a determination step of determining whether or not the information processing task should be offloaded to an external device using a learning model to which information about the information processing task is input, a communication processing step of offloading the information processing task to the external device if it is determined in the determination step that the information processing task should be offloaded to the external device, and a task processing step of executing the information processing task if it is determined in the determination step that the information processing task should not be offloaded to the external device.

[0009] The present invention also encompasses a control program for an information processing terminal or server device that causes a computer to operate as each part (software element) of the information processing terminal or server device according to each aspect of the present invention, thereby realizing the information processing terminal or server device on the computer, and a computer-readable recording medium on which the program is recorded.

[0010] The control program may use various machine learning techniques in the process of causing a computer to operate as each of the above units or in other processes. In this case, the program using the machine learning technique may run on a server or device, or may run on another apparatus (e.g., an edge computer or a cloud server). [Brief explanation of the drawings]

[0011] [Figure 1] 1 is an example of a block diagram showing a functional configuration of an information processing system according to a first embodiment. [Figure 2] 1 illustrates an example of a conceptual diagram of a configuration including an information processing system. [Figure 3] 1 is a sequence diagram illustrating an example of a flow of an execution process of an information processing task according to the first embodiment. [Figure 4] FIG. 1 is a diagram illustrating an example of an image and a bounding box. [Figure 5] 10 shows an example of the results of object detection performed by a robot and the results of similar object detection performed by a server device. [Figure 6] 4 is an example of a sequence diagram showing the flow of a learning process according to the first embodiment. [Figure 7] FIG. 10 is a block diagram illustrating an example of a functional configuration of an information processing system according to a second embodiment. [Figure 8] 10 is an example of a sequence diagram showing the flow of a learning process according to the second embodiment. [Figure 9] FIG. 11 is an example of a block diagram showing a functional configuration of an information processing system according to a third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] [Embodiment 1] Hereinafter, one embodiment of the present invention will be described in detail.

[0013] [1. Example of configuration of information processing system 1] 1 is an example of a block diagram showing the functional configuration of an information processing system 1 according to this embodiment. The information processing system 1 includes a robot 10 and a server device (external device) 30, and is a system for executing information processing tasks assigned to the robot 10.

[0014] Here, an information processing task refers to a specific information processing that obtains a result from input such as sensing results or user instructions. Below are examples of information processing tasks, and the words in parentheses indicate typical methods used for the corresponding information processing tasks. ·Self-localization (AMCL:Adaptive Monte Carlo Localization) Path planning (DWA: Dynamic Window Approach) Object detection (CNN: Convolutional Neural Network, ViT: Vision Transformer) Manipulation (IK: Inverse Kinematics) ·Voice recognition (LSTM:Long Short Term Memory) General-purpose command processing (SayCan) ·Sound source localization ·Odor source localization (Music:MUltiple SIgnal Classification) The robot 10 is capable of recognizing users around it, conversing with them, and moving around. The information processing terminal 12 of the robot 10 includes a control unit 14, a storage unit 22, a communication unit 23, a photographing unit 24, a voice input unit 25, and a voice output unit 26.

[0015] The control unit 14 is a control device that controls the entire robot 10 , and includes an acquisition unit 16 , a determination unit 17 , a communication processing unit 18 , a task processing unit 19 , and a learning unit 20 .

[0016] The acquisition unit 16 acquires an information processing task, which includes receiving or identifying an input that is the target of the information processing task.

[0017] The determination unit 17 uses a trained learning model to determine whether or not an information processing task should be offloaded to the server device 30. Information about the information processing task is input to the learning model, and the determination unit 17 outputs information indicating whether or not the information processing task should be offloaded to the server device 30. Here, offloading means requesting an external device such as the server device 30 to execute a part or all of the target information processing task. The determination unit 17 also performs other determination processes, which will be described later.

[0018] The communication processing unit 18 controls communication processing by the communication unit 23. For example, when the determination unit 17 determines that an information processing task should be offloaded to the server device 30, the communication processing unit 18 offloads the information processing task to the server device 30.

[0019] The task processing unit 19 executes the information processing task when the determination unit 17 determines that the information processing task should not be offloaded to the server device 30, for example.

[0020] The learning unit 20 trains the learning model used by the determination unit 17 using, as training data, a set of information on an information processing task and information indicating whether the information processing task should be offloaded to the server device 30.

[0021] The storage unit 22 is a storage device that stores various information, for example, a parameter set that defines a learning model.

[0022] The communication unit 23 performs communication processing with an external device such as a server device 30 under the control of the communication processing unit 18. For example, the communication unit 23 transmits information indicating an information processing task and a request for its execution to the server device 30 via the network 3.

[0023] The photographing unit 24 is a camera that photographs the surroundings of the robot 10. The audio input unit 25 is a microphone that captures audio around the robot 10. The control unit 14 acquires the audio input to the audio input unit 25 and analyzes the content. The audio output unit 26 is a speaker that outputs the audio of the robot 10.

[0024] Note that the robot 10 may be configured to be able to give instructions via a button or touch panel (not shown), or the robot 10 may be configured to include other sensors.

[0025] The server device 30 is a device that functions as a server for the robot 10, which is a client, and includes a control unit 32, a storage unit , and a communication unit .

[0026] The control unit 32 is a control device that controls the entire server. For example, the control unit 32 executes information processing tasks offloaded from the robot 10. Furthermore, the control unit 32 has higher performance in executing information processing tasks than the task processing unit 19 of the robot 10.

[0027] The storage unit 34 is a storage device that stores various types of information, for example, information used to execute information processing tasks.

[0028] The communication unit 35 performs communication processing with the robot 10 and the like under the control of the control unit 32. For example, the communication unit 35 transmits the execution result of the information processing task to the robot 10 via the network 3.

[0029] Furthermore, the function of a single component included in the information processing system 1 may be realized by a plurality of other components, and the functions of a plurality of components included in the information processing system 1 may be realized by a single other component.

[0030] 2 shows an example of a conceptual diagram of a configuration including an information processing system 1. The robot 10 receives an information processing task from a user, inputs information about the information processing task into its own learning model, and determines whether or not the information processing task should be offloaded to the server device 30. Generally, if the information processing task is more complex than a certain standard, it is offloaded to the server device 30; otherwise, the robot 10 executes the information processing task. Note that the information processing task is not limited to being input by the user; for example, the robot 10 may execute the information processing task at regular intervals using sensing results as input.

[0031] [2. Processing example of information processing system 1] Next, an example of the flow of processing executed by the information processing system 1 will be described. In this example, the information processing task will be described as an object detection process in an image, but the same applies to other information processing tasks. Figure 3 is an example of a sequence diagram showing the flow of processing for executing an information processing task according to this embodiment.

[0032] In S101 (step S101), the photographing unit 24 photographs the surroundings of the robot 10. The acquiring unit 16 acquires the image photographed by the photographing unit 24 as a target for the information processing task.

[0033] In S102, the task processing unit 19 performs a process of detecting an object included in the image acquired by the acquisition unit 16, and derives the position, size, and confidence score of a bounding box (hereinafter simply referred to as a "box") surrounding the object. The position of the box indicates the area of the detected object.

[0034] FIG. 4 is a diagram showing an example of an image and a box. In FIG. 4, "Threshold (detection threshold)" indicates the low sensitivity of object detection. "TP (True Positive)" indicates a true positive, where an object in the image is correctly detected, and "TN (True Negative)" indicates a true negative, where an object not present in the image is not detected and a correct judgment is made. "FP (False Positive)" indicates a false positive, where an object not present in the image is incorrectly detected. "FN (False Negative)" indicates a false negative, where an object present in the image is not detected. The higher the confidence score of a box, the higher the probability that the judgment for that box is "TP".

[0035] In S103, the determination unit 17 inputs information indicating the position and size of each box and the confidence score of each box into the learning model as information related to the information processing task. Note that only one of the position and size of the box may be input into the learning model, and this also applies to the following embodiments. Next, the determination unit 17 determines whether or not the information processing task should be offloaded according to the output of the learning model. If the determination unit 17 determines that the information processing task should be offloaded to the server device 30 (S103: YES), the process of S104 is executed next. If the determination unit 17 determines that the information processing task should not be offloaded (S103: NO), the process of S107 is executed next.

[0036] In S104, the communication processing unit 18 offloads the object detection process, which is an information processing task, to the server device 30 via the communication unit 23. That is, the communication processing unit 18 transmits to the server device 30 an image and information indicating a request to execute a process for detecting an object in the image.

[0037] In S105, the control unit 32 of the server device 30 executes a process to detect an object in the image transmitted from the robot 10, and transmits the execution result to the robot 10 via the communication unit 35. In this way, when an information processing task is offloaded from the robot 10, the server device 30 executes the information processing task and transmits the execution result to the robot 10.

[0038] In S106, the communication processing unit 18 receives the execution result transmitted from the server device 30, and the task processing unit 19 determines the execution result as the result of the information processing task.

[0039] On the other hand, in S107, the task processing unit 19 determines the execution result of the object detection performed in S102 as the result of the information processing task.

[0040] The above has described an information processing method executed by an apparatus, which includes an acquisition step, a determination step, a communication processing step, and a task processing step. In the acquisition step, an information processing task is acquired, and in the determination step, whether or not the information processing task should be offloaded to the server apparatus 30 is determined using a learning model to which information related to the information processing task is input. If it is determined in the determination step that the information processing task should be offloaded to the server apparatus 30, a communication processing step is executed to offload the information processing task to the server apparatus 30. Furthermore, if it is determined in the determination step that the information processing task should not be offloaded to the server apparatus 30, a task processing step is executed to execute the information processing task.

[0041] According to the configuration of this example, the robot 10 determines whether or not to offload an information processing task, and if the information processing task is of an appropriate level of difficulty, the robot 10 itself executes it, and if not, the robot 10 offloads it to the server device 30.

[0042] [3. Learning process example] Next, the process of the learning unit 20 learning the learning model will be explained by assuming that Fig. 4 shows the object detection process by the task processing unit 19. Images 51a to 51c in Fig. 4 show the same image including a car, and images 52a to 52c show the same image including a car and a robot.

[0043] As shown in Figure 4, when the detection threshold is 0.8, the car is detected but the robot is not. When the detection threshold is 0.6 or 0.4, the car and the robot are detected, but objects that do not exist in the image are falsely detected.

[0044] In other words, even if the task processing unit 19 lowers the detection threshold from 0.8 to 0.6 in order to detect all objects in the image of FIG. 4, false detection occurs and the object cannot be detected correctly.

[0045] FIG. 5 shows an example of the results of object detection performed by the robot 10 on the aforementioned image 52 (52a to 52c) including a car and a robot, and the results of similar object detection performed by the server device 30. The object detection by the task processing unit 19 of the robot 10 resulted in erroneous detection and was not performed correctly, but when the control unit 32 of the server device 30 performed object detection on the same image, the object was correctly detected. In other words, it is preferable to offload object detection from the robot 10 to the server device 30 for this image 52. The above-mentioned "TP," "TN," "FP," and "FN" may be determined based on an Intersection over Union (IoU) threshold between the bounding boxes detected by the robot and the bounding boxes detected by the server device.

[0046] When image 52 is used for learning, learning unit 20 inputs, as training data, sets of information indicating the area of each object detected in image 52, i.e., the position of each box or the size of each box, as well as the confidence score of each box and information indicating that the corresponding object detection process should not be offloaded, into the learning model. The same applies to image 51 (51a to 51c) in Figure 4.

[0047] In a broad sense, the learning unit 20 compares the results of a certain information processing task executed by the robot 10 with the results of the server device 30, and if the difference between the results executed by the robot 10 and the server device 30 satisfies a predetermined condition, inputs training data corresponding to the certain information processing task into the learning model as training data for offloading to the server device 30. Here, the predetermined condition is, for example, a condition that is met when, in an object detection process within an image, the positions, sizes, or numbers of boxes derived by the robot 10 and boxes derived by the server device 30 differ within a predetermined detection threshold. On the other hand, if, for example, the positions, sizes, or numbers of boxes derived by the robot 10 and boxes derived by the server device 30 are the same within a predetermined detection threshold, the predetermined condition is not met, and these are used as training data that should not be offloaded.

[0048] In addition, when the information processing task is an object detection process within an image, the learning unit 20 trains the learning model using, as training data, information indicating the position of the box derived within the image or the size of the box, as well as a combination of a confidence score for each object detection and information indicating whether the information processing task should be offloaded to the server device 30.

[0049] Next, an example of the flow of the learning process according to this embodiment will be described with reference to FIG.

[0050] In S201, the task processing unit 19 performs a process of detecting objects included in a learning image, and derives the position, size, and confidence score of a box for each object.

[0051] In S202, the communication processing unit 18 transmits the image to the server device 30.

[0052] In S203, the control unit 32 of the server device 30 performs a process of detecting an object included in the image received from the robot 10, and derives the position, size, and reliability score of the box. The control unit 32 also transmits the derived position, size, and reliability score of the box to the robot 10.

[0053] In S204, the determination unit 17 of the robot 10 compares the result of the object detection process executed by the robot 10 with the result of the object detection process executed by the server device 30, and determines whether the above-mentioned predetermined condition is satisfied. In other words, the determination unit 17 determines whether the object detection process for the learning image should be offloaded to the server device 30.

[0054] In S205, the learning unit 20 inputs, as training data, a set of information indicating the position and size of the box, the confidence score of the box, and information indicating whether the object detection process should be offloaded to the server device 30 into the learning model. As a result, the values of the parameter set are updated, and the learning model is trained.

[0055] [Embodiment 2] A second embodiment of the present invention will be described below. For ease of explanation, components having the same functions as those described in the above embodiment will be denoted by the same reference numerals, and redundant explanations will not be repeated. In this embodiment, a case will be described in which a server device executes a learning process. This also applies to subsequent embodiments.

[0056] 1. Example of the configuration of the information processing system 1a FIG. 7 is an example of a block diagram showing the functional configuration of an information processing system 1a according to this embodiment. As shown in FIG. 7, the control unit 14a does not include a learning unit 20, and the control unit 32a includes a learning unit 37 that learns a learning model. The memory unit 34 according to this embodiment stores parameter sets and the like that define the learning model. The parameter sets stored in the memory unit 34 are also shared with the memory unit 22 of the robot 10a via the network as needed. In other words, the learning model held by the robot 10a is synchronized with the learning model held by the server device 30a as needed.

[0057] [2. Learning process example] Next, a description will be given of the process of learning the learning model by the learning unit 37. Fig. 8 is an example of a sequence diagram showing the flow of the learning process according to this embodiment.

[0058] In S201, similarly to FIG. 6, the task processing unit 19 performs processing to detect an object included in the learning image, and derives the position, size, and confidence score of a box.

[0059] In S302, the communication processing unit 18 transmits the image and the derived position, size and confidence score of the box to the server device 30a.

[0060] In S303, the control unit 32a of the server device 30a performs a process of detecting an object included in the image received from the robot 10a, and derives the position, size, and reliability score of the box.

[0061] In S304, the control unit 32a compares the result of the object detection process executed by the robot 10a with the result of the object detection process executed by the control unit 32a, and determines whether a predetermined condition is satisfied. In other words, the control unit 32a determines whether the object detection process for the learning image should be offloaded to the server device 30a.

[0062] In S305, the learning unit 37 inputs, as training data, a set of information indicating the position and size of the box, the confidence score of the box, and information indicating whether the object detection process should be offloaded from the robot 10a to the server device 30a, into the learning model. As a result, the parameter set is updated and the learning model is trained.

[0063] In S306, the control unit 32a transmits the values of the updated parameter set to the robot 10a.

[0064] In S307, the control unit 14a of the robot 10a overwrites the values of the parameter set stored in the storage unit 22 with the values of the parameter set received from the server device 30a, thereby synchronizing the learning model of the robot 10a with the learning model of the server device 30a.

[0065] [Embodiment 3] A third embodiment of the present invention will be described below, which deals with a case where there are multiple server devices that are candidates for offloading work from a robot.

[0066] [1. Example of configuration of information processing system 1b] Fig. 9 is an example of a block diagram showing the functional configuration of an information processing system 1b according to this embodiment. As shown in Fig. 9, the information processing system 1b includes a robot 10 and a plurality of server devices 30. The performance of some or all of the server devices 30 may differ from one another.

[0067] [2. Processing example of information processing system 1b] 3, the determination unit 17 according to this embodiment uses a learning model to determine to which server device 30 the object detection process should be offloaded or not offloaded. In addition, in a step corresponding to S104, the communication processing unit 18 offloads the object detection process to the server device 30 determined to be offloaded.

[0068] Furthermore, the determination unit 17 may determine to which server device 30 the work should be offloaded, based on at least one of the performance, processing load, and communication load of each server device 30. The control unit 14 acquires information on the performance and the like from the server device 30 via the communication unit 23. Furthermore, the performance may include the cost incurred when using the server device 30.

[0069] For example, when the output of the learning model indicates that offloading should be performed to a first server device 30, if the processing load on the first server device 30 is above a threshold, the judgment unit 17 may determine that offloading should be performed to a second server device 30 that has higher performance than the first server device 30.

[0070] According to the configuration of this example, it is possible to offload information processing tasks to a suitable server device 30 depending on the level of difficulty of the information processing task.

[0071] [3. Learning process example] The learning unit 20 in this embodiment inputs a set of information indicating the position and size of the box, the box's reliability score, and information indicating to which server device 30 the information processing task should be offloaded or not offloaded into the learning model as training data.

[0072] Here, which server device 30 the information processing task should be offloaded to is determined in advance by the determination unit 17, for example, as follows.

[0073] The determination unit 17 compares the execution result of the information processing task by the robot 10, the execution result of the information processing task by the first server device 30, and the execution result of the information processing task by a second server device 30 that has higher performance than the first server device 30. If the execution results are all identical, the determination unit 17 determines that offloading is not necessary and that the information processing task should not be offloaded. If the execution result by the second server device 30 is different from the other execution results, the determination unit 17 determines that the information processing task should be offloaded to the second server device 30. If the execution result by the first server device 30 and the execution result by the second server device 30 are identical but different from the execution result by the robot 10, the determination unit 17 determines that the information processing task should be offloaded to the first server device 30.

[0074] In the configuration of this embodiment, the server device 30a may also be configured to include the learning unit 37 as in the second embodiment.

[0075] [Modifications of Embodiments 1 to 3] The information processing system 1 (1a, 1b) may include a plurality of robots 10 (10a). In particular, when the performance of each robot 10 is the same, a learning model updated in one of the robots 10 or the server device 30a may be synchronized with the other robots 10.

[0076] Furthermore, it is not essential that the robot 10 be able to talk or move. The information processing terminal 12 (12a) may be a personal computer, a smartphone, a tablet, or the like.

[0077] As described above, the information processing task is not limited to the object detection process. For example, if the result of the self-location estimation performed by the robot 10 indicates that the robot 10 is in a location where the robot 10 cannot exist, the information processing task may be offloaded to the server device 30.

[0078] Alternatively, offloading may be performed in multiple stages, i.e., the robot 10 may offload to a first server device 30 based on the output of a learning model, and the first server device 30 may offload to a second server device 30 based on the output of another learning model.

[0079] Furthermore, the amount of information used by the robot 10 for the information processing task does not have to be the same as the amount of information used by the server device 30 for the information processing task. For example, when offloading to the server device 30, the robot 10 may be configured to transmit to the server device 30 parameters that are not used by the robot 10 for the information processing task but are used by the server device 30 for the information processing task.

[0080] [Software implementation example] The functions of the robot 10 and the server device 30 (hereinafter referred to as the "device") can be realized by a program that causes a computer to function as the device, and a program that causes a computer to function as each control block of the device (particularly each part included in the control unit 14 or the control unit 32).

[0081] In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The control device and storage device execute the program, thereby realizing the functions described in each of the above embodiments.

[0082] The program may be non-transitory and may be recorded on one or more computer-readable recording media. The recording media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

[0083] Furthermore, some or all of the functions of the control blocks can be realized by logic circuits. For example, an integrated circuit in which a logic circuit that functions as each of the control blocks is formed is also included in the scope of the present invention. In addition, the functions of the control blocks can also be realized by, for example, a quantum computer.

[0084] Furthermore, each process described in each of the above embodiments may be executed by AI (Artificial Intelligence). In this case, the AI may run on the control device or on another device (for example, an edge computer or a cloud server).

[0085] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention. [Explanation of symbols]

[0086] 1, 1a, 1b Information Processing System 10, 10a Robot 12, 12a Information processing terminal 14, 14a, 32, 32a Control unit 16 Acquisition Department 17 Judgment section 18 Communication processing unit 19 Task Processing Unit 20, 37 Learning Department 22, 34 Storage section 23, 35 Communications Department 24 Filming Department 25 Audio input section 26 Audio output section 30, 30a Server device (external device)

Claims

1. an acquisition unit that acquires an information processing task; a determination unit that determines whether or not the information processing task should be offloaded to an external device using a learning model to which information related to the information processing task is input; a communication processing unit that offloads the information processing task to the external device when the determination unit determines that the information processing task should be offloaded to the external device; a task processing unit that executes the information processing task when the determination unit determines that the information processing task should not be offloaded to the external device; and Equipped with the information processing task is an information processing task for detecting an object in an image; The determination unit As information related to the information processing task, information indicating the area of each object detected in the image or the size of the area, and a confidence score for each object detection are input to the learning model. An information processing terminal characterized by:

2. The determination unit determining to which of the two or more external devices the information processing task should be offloaded, based on at least one of the performance, processing load, and communication load of each external device; 2. The information processing terminal according to claim 1, wherein:

3. A robot comprising the information processing terminal according to claim 1 or 2.

4. An information processing system including an information processing terminal and a server device, The information processing terminal an acquisition unit that acquires an information processing task; a determination unit that determines whether or not the information processing task should be offloaded to the server device using a learning model to which information related to the information processing task is input; a communication processing unit that offloads the information processing task to the server device when the determination unit determines that the information processing task should be offloaded to the server device; an information processing unit that executes the information processing task when the determination unit determines that the information processing task should not be offloaded to the server device; and Equipped with The server device When the information processing task is offloaded from the information processing terminal, the information processing task is executed and a result of the execution is transmitted to the information processing terminal; The information processing terminal or the server device a learning unit that causes the learning model to learn using, as training data, a set of information about the information processing task and information indicating whether the information processing task should be offloaded to the server device; the information processing task is an information processing task for detecting an object in an image; The learning unit The learning model is trained using, as training data, a set of information indicating the area of each object detected in the image or the size of the area, a confidence score for each object detection, and information indicating whether the information processing task of detecting the object in the image should be offloaded to the server device. An information processing system comprising:

5. An information processing system comprising an information processing terminal and a server device, The information processing terminal an acquisition unit that acquires an information processing task; a determination unit that determines whether or not the information processing task should be offloaded to the server device using a learning model to which information related to the information processing task is input; a communication processing unit that offloads the information processing task to the server device when the determination unit determines that the information processing task should be offloaded to the server device; an information processing unit that executes the information processing task when the determination unit determines that the information processing task should not be offloaded to the server device; and Equipped with The server device When the information processing task is offloaded from the information processing terminal, the information processing task is executed and a result of the execution is transmitted to the information processing terminal; The information processing terminal or the server device a learning unit that causes the learning model to learn using, as training data, a set of information about the information processing task and information indicating whether the information processing task should be offloaded to the server device; The learning unit A result of execution of a certain information processing task for learning by the information processing terminal is compared with a result of execution by the server device, and if a difference between the result of execution by the information processing terminal and the result of execution by the server device satisfies a predetermined condition, training data corresponding to the certain information processing task is input into the learning model as training data for when the certain information processing task should be offloaded to the server device. An information processing system comprising:

6. 1. An information processing method performed by an apparatus, comprising: an acquisition step of acquiring an information processing task; a determining step of determining whether the information processing task should be offloaded to an external device using a learning model to which information about the information processing task is input; a communication processing step of offloading the information processing task to the external device when it is determined in the determining step that the information processing task should be offloaded to the external device; a task processing step of executing the information processing task when it is determined in the determination step that the information processing task should not be offloaded to the external device; Including, the information processing task is an information processing task for detecting an object in an image; In the determining step, As information related to the information processing task, information indicating the area of each object detected in the image or the size of the area, and a confidence score for each object detection are input to the learning model. An information processing method comprising:

7. 2. A control program for causing a computer to function as the information processing terminal according to claim 1, the control program causing a computer to function as said acquisition section, said determination section, said communication processing section, and said task processing section.

8. A computer-readable recording medium on which the control program according to claim 7 is recorded.

Images