Information processing device and non-transitory computer-readable medium

Abstract

An information processing device and a non-transitory computer-readable medium. An information processing device includes: a receiving unit that receives learning data; and a changing unit that changes a learning process of an artificial intelligence based on information attached to the learning data and indicating whether the artificial intelligence is allowed to learn the learning data.

Description

Technical Field

[0001] The present disclosure relates to an information processing device and a non-transitory computer-readable medium. Background Art

[0002] Typically, artificial intelligence learns from learning data.

[0003] Japanese Unexamined Patent Application Publication No. 2010-223824 describes an apparatus for preventing repeated learning of erroneous new roads. Summary of the Invention

[0004] Sometimes, certain learning data is not desired to be learned by artificial intelligence.

[0005] Therefore, an object of the present disclosure is to prevent artificial intelligence from learning learning data that is not desired to be learned by the artificial intelligence.

[0006] According to a first aspect of the present disclosure, there is provided an information processing device comprising: a receiving unit that receives learning data; and a changing unit that changes a learning process of an artificial intelligence based on information attached to the learning data and indicating whether the artificial intelligence is allowed to learn the learning data.

[0007] According to a second aspect of the present disclosure, the information processing apparatus according to the first aspect of the present disclosure is configured such that, when the information indicates that the artificial intelligence is not allowed to learn the learning data, the changing unit prohibits the artificial intelligence from learning the learning data.

[0008] According to the third aspect of the present disclosure, the information processing device according to the second aspect of the present disclosure is configured so that when information indicates that artificial intelligence is not allowed to learn the learning data, the change unit allows artificial intelligence to learn the learning data if the purpose of using the learning data for learning meets the standard.

[0009] According to a fourth aspect of the present disclosure, the information processing device according to the second aspect or the third aspect of the present disclosure is configured to further include an output unit that outputs a warning when information indicates that artificial intelligence is not allowed to learn learning data and the learning data is about to be changed without permission.

[0010] According to the fifth aspect of the present disclosure, the information processing device according to the second aspect or the third aspect of the present disclosure is configured to also include a deletion unit, which deletes the learning data when the information indicates that artificial intelligence is not allowed to learn the learning data and the learning data is about to be changed without permission.

[0011] According to a sixth aspect of the present disclosure, the information processing apparatus according to the first aspect of the present disclosure is configured such that, if the information indicates that the artificial intelligence is allowed to learn the learning data, the changing unit allows the artificial intelligence to learn the learning data.

[0012] According to a seventh aspect of the present disclosure, the information processing device according to the sixth aspect of the present disclosure is configured so that the information further includes information indicating a time period during which the artificial intelligence is allowed to learn the learning data.

[0013] According to an eighth aspect of the present disclosure, the information processing apparatus according to the sixth aspect of the present disclosure is configured so that the information further includes information indicating a limit on the number of times learning is performed on the learning data.

[0014] According to a ninth aspect of the present disclosure, the information processing device according to any one of the first to eighth aspects of the present disclosure is configured so that the information further includes information indicating whether artificial intelligence is allowed to learn the learning data according to the purpose of the learning data.

[0015] According to a tenth aspect of the present disclosure, the information processing device according to any one of the first to ninth aspects of the present disclosure is configured so that the information further includes information indicating whether the artificial intelligence is allowed to learn the learning data according to the system of the artificial intelligence.

[0016] According to an eleventh aspect of the present disclosure, the information processing device according to any one of the first to tenth aspects of the present disclosure is configured so that the information includes information indicating whether the artificial intelligence is allowed to learn the learning data for each use field of the artificial intelligence.

[0017] According to the twelfth aspect of the present disclosure, an information processing device is provided, which includes: a receiving unit that receives learning data; and a prohibiting unit that prohibits artificial intelligence from learning the learning data when the learning data is attached with information indicating that artificial intelligence is prohibited from learning the learning data.

[0018] According to the thirteenth aspect of the present disclosure, an information processing device is provided, which includes: a receiving unit that receives learning data; and a allowing unit that allows artificial intelligence to learn the learning data if the learning data is attached with information indicating that artificial intelligence is allowed to learn the learning data.

[0019] According to a fourteenth aspect of the present disclosure, there is provided an information processing apparatus including: an appending unit that appends information indicating whether use of the content data as learning data for artificial intelligence is permitted to content data.

[0020] According to the fifteenth aspect of the present disclosure, a non-temporary computer-readable medium is provided, which stores a program that causes a computer to perform processing for image processing, the processing including: receiving learning data; and changing the learning processing of artificial intelligence based on information attached to the learning data and indicating whether the artificial intelligence is allowed to learn the learning data.

[0021] According to a sixteenth aspect of the present disclosure, there is provided a non-transitory computer-readable medium storing a program causing a computer to execute a process for image processing, the process including appending information indicating whether the content data is permitted to be used as learning data for artificial intelligence to content data.

[0022] According to the first aspect, the second aspect, and the twelfth to the sixteenth aspects of the present disclosure, it is possible to prevent artificial intelligence from learning learning data that is not desired to be learned by the artificial intelligence.

[0023] According to the third aspect of the present disclosure, even in a case where artificial intelligence is not allowed to learn the learning data, artificial intelligence is allowed to learn the learning data if the learning purpose satisfies the criteria.

[0024] According to the fourth aspect of the present disclosure, it is possible to notify a user or the like that learning data is about to be changed without permission.

[0025] According to the fifth aspect of the present disclosure, in a case where learning data is about to be changed without permission, the change can be prevented.

[0026] According to the sixth aspect of the present disclosure, artificial intelligence is allowed to learn from learning data.

[0027] According to the seventh aspect of the present disclosure, artificial intelligence is allowed to learn the learning data only during a learning-allowed time period.

[0028] According to the eighth aspect of the present disclosure, artificial intelligence is allowed to learn learning data only a limited number of times.

[0029] According to the ninth aspect of the present disclosure, artificial intelligence is allowed to learn the learning data according to the usage of the learning data.

[0030] According to a tenth aspect of the present disclosure, a system based on artificial intelligence allows artificial intelligence to learn learning data.

[0031] According to the eleventh aspect of the present disclosure, artificial intelligence is allowed to learn learning data according to the use field of artificial intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Exemplary embodiments of the present disclosure will be described in detail based on the following drawings, in which:

[0033] Figure 1is a block diagram showing a configuration of an information processing apparatus according to a first exemplary embodiment;

[0034] Figure 2 shows learning data according to the first exemplary embodiment;

[0035] Figure 3 shows content data according to the first exemplary embodiment;

[0036] Figure 4 is a block diagram showing a configuration of an information processing apparatus according to a second exemplary embodiment;

[0037] Figure 5 shows learning data according to the second exemplary embodiment;

[0038] Figure 6 shows learning data according to the second exemplary embodiment;

[0039] Figure 7 shows learning data according to the second exemplary embodiment;

[0040] Figure 8 A database showing determination results;

[0041] Figure 9 shows learning data according to a first variation of the second exemplary embodiment;

[0042] Figure 10 shows learning data according to a first variation of the second exemplary embodiment;

[0043] Figure 11 shows learning data according to a first variation of the second exemplary embodiment;

[0044] Figure 12 A database showing determination results;

[0045] Figure 13 shows learning data according to a second variation of the second exemplary embodiment; and

[0046] Figure 14 Learning data according to a second modification of the second exemplary embodiment is shown. DETAILED DESCRIPTION

[0047] First Exemplary Embodiment

[0048] Reference Figure 1 An information processing apparatus according to a first exemplary embodiment of the present disclosure is described. Figure 1 An example of the information processing apparatus according to the first exemplary embodiment is shown.

[0049] The information processing apparatus 10 according to the first exemplary embodiment is configured to receive learning data and change a learning process of artificial intelligence (ie, AI) according to information attached to the learning data and indicating whether to allow artificial intelligence (ie, AI) to learn the learning data.

[0050] The information processing device 10 is, for example, a personal computer (hereinafter referred to as "PC"), a tablet PC, a smartphone, a mobile phone, or any other type of device (for example, a multifunction printer having functions such as a printing function). Needless to say, the information processing device 10 may be a device other than these devices.

[0051] The algorithm used for artificial intelligence is not particularly limited and can be any algorithm. The algorithm is, for example, machine learning. Machine learning can be supervised learning, or can be unsupervised learning, or can be reinforcement learning. Specifically, deep learning (for example, multilayer perceptron, convolutional neural network, recursive neural network, autoencoder, restricted Boltzmann machine), perceptron, back propagation, associator, support vector machine, decision tree, k- nearest neighbor algorithm, linear regression, self-organizing map, Boltzmann machine, principal component analysis, cluster analysis or Q-learning etc. can be used. Genetic algorithm or hill climbing method etc. can be used as algorithms other than machine learning. Needless to say, algorithms other than these algorithms can be used.

[0052] Learning data is data used for artificial intelligence learning. Learning data can be data that includes correct judgments (i.e., answers) as learning data for supervised learning, or can be data that does not include correct judgments as learning data for unsupervised learning. Learning data is, for example, document data (e.g., text data), image data (e.g., still image data or moving image data), music data, audio data, or a combination thereof, and its type, data format, and content are not particularly limited.

[0053] Attribute information is attached to learning data. Attribute information attached to learning data is information indicating whether artificial intelligence is allowed to learn from the learning data. The state in which attribute information is attached to learning data indicates that the attribute information accompanies the learning data as accompanying information, or that the attribute information is embedded in the learning data itself. Attribute information can be added to learning data as accompanying information, or can be associated with the learning data as data separate from the learning data. The state in which attribute information is embedded in the learning data itself is, for example, a state in which attribute information is set in an image represented by image data, or a state in which attribute information is set in a document represented by document data.

[0054] Artificial intelligence may be installed in the information processing device 10, or may be installed in a device other than the information processing device 10 (for example, a server or a PC). That is, a program for realizing artificial intelligence may be stored in the information processing device 10, or may be stored in a device other than the information processing device 10.

[0055] The configuration of the information processing device 10 is described in detail below.

[0056] The communication unit 12 is a communication interface and has a function of sending information to another device and a function of receiving information from another device. The communication unit 12 may have a wireless communication function or a wired communication function. The communication unit 12 can communicate with another device via a communication path by using wireless communication or wired communication. The communication path is, for example, a network such as a local area network (LAN) or the Internet. The communication unit 12 can communicate with another device without a communication path by using, for example, short-range wireless communication. Short-range wireless communication is, for example, Bluetooth (registered trademark), radio frequency identifier (RFID), or NFC.

[0057] For example, when learning data is transmitted to information processing device 10 from a device other than information processing device 10, communication unit 12 receives the learning data. When attribute information is attached to the learning data, communication unit 12 also receives the attribute information. In addition, communication unit 12 can transmit learning data to another device.

[0058] The UI unit 14 is a user interface and includes a display and an operating unit. The display is a display device such as a liquid crystal display. The operating unit is an input unit such as a keyboard, input keys, or an operating panel. The UI unit 14 can be a UI unit such as a touch panel that serves as both a display and an operating unit.

[0059] The storage unit 16 is one or more storage areas that store various information. Each storage area is composed of, for example, one or more storage devices (e.g., physical drives such as hard disk drives or memories) provided in the information processing device 10. Learning data can be stored in the storage unit 16.

[0060] The receiving unit 18 is configured to receive learning data. For example, in the case where the communication unit 12 receives learning data sent from a device other than the information processing device 10, the receiving unit 18 receives the learning data received by the communication unit 12. In the case where the communication unit 12 receives attribute information attached to the learning data, the receiving unit 18 also receives the attribute information. In addition, in the case where learning data is provided to the information processing device 10 through the UI unit 14, or in the case where learning data is provided to the information processing device 10 by using a storage device (for example, a hard disk drive, a USB memory, a CD, a DVD, or any other portable storage medium), the receiving unit 18 receives the provided learning data. In the case where attribute information is attached to the provided learning data, the receiving unit 18 also receives the attribute information.

[0061] The changing unit 20 is configured to change the learning process of artificial intelligence according to the attribute information attached to the learning data.

[0062] If the attribute information attached to the learning data indicates that the artificial intelligence is permitted to learn the learning data, the change unit 20 permits the artificial intelligence to learn the learning data. For example, if the user instructs the artificial intelligence to learn the learning data by operating the UI unit 14, the controller 28 causes the artificial intelligence to learn the learning data. The user can specify the artificial intelligence to learn the learning data and cause the artificial intelligence to learn the learning data.

[0063] If the attribute information attached to the learning data indicates that the artificial intelligence is not allowed to learn the learning data, the change unit 20 prohibits the artificial intelligence from learning the learning data. In other words, the change unit 20 does not allow the artificial intelligence to learn the learning data. For example, even if the user instructs the artificial intelligence to learn the learning data by operating the UI unit 14, the controller 28 does not cause the artificial intelligence to learn the learning data.

[0064] Learning can be allowed for a portion of the learning data, and can be prohibited for the rest of the learning data. Alternatively, learning can be allowed or prohibited for all learning data. This is indicated by attribute information.

[0065] When the attribute information attached to the learning data indicates that the artificial intelligence is not allowed to learn the learning data and the purpose of learning with the learning data satisfies a predetermined standard, the change unit 20 may allow the artificial intelligence to learn the learning data. For example, when the learning purpose is a highly public purpose such as a medical purpose or an educational purpose, the change unit 20 may allow the artificial intelligence to learn the learning data.

[0066] The output unit 22 is configured to output a warning when the attribute information attached to the learning data indicates that the artificial intelligence is not allowed to learn the learning data and the learning data is about to be changed without permission. The situation where the learning data is about to be changed without permission is a situation where a user who has no authority to change the learning data attempts to change the learning data.

[0067] For example, user identification information (e.g., user name, user ID, password) for identifying a user who has the authority to change the learning data is associated with the learning data. In a case where the user identification information of the user attempting to change the learning data is associated with the learning data, the controller 28 allows the user to change the learning data. Specifically, in a case where the user performs an operation to change the learning data by operating the UI unit 14, the controller 28 allows the operation and changes the learning data according to the user's operation. In a case where the user identification information of the user attempting to change the learning data is not associated with the learning data, the controller 28 prohibits the user from changing the learning data. Specifically, in a case where the user performs an operation to change the learning data by operating the UI unit 14, the controller 28 prohibits the operation and does not change the learning data. Note that the operation of changing the learning data can be performed by a device other than the information processing device 10 (e.g., a PC).

[0068] For example, in a case where learning data is to be changed (for example, in a case where the user specifies learning data by operating the UI unit 14), the controller 28 causes an input screen for inputting user identification information of a user who has the authority to change the learning data to be displayed on the display of the UI unit 14. In a case where user identification information of a user who has the authority to change the learning data is input on the input screen, the controller 28 allows the learning data to be changed. In another example, in a case where user identification information is input when the user logs into the information processing device 10, the controller 28 may allow the learning data to be changed if the user identification information is the user identification information of a user who has the authority to change the learning data.

[0069] For example, the output unit 22 can cause information indicating a warning to be displayed on the display of the UI unit 14, can send information indicating a warning to a device (e.g., a PC) that has sent learning data to the information processing device 10, can cause a speaker to produce a warning sound, or can send information indicating a warning to a pre-registered device (e.g., a PC used by an administrator, etc.).

[0070] The deletion unit 24 is configured to delete the learning data when the attribute information attached to the learning data indicates that the artificial intelligence is not allowed to learn the learning data and the learning data is about to be changed without permission. As described above, the situation in which the learning data is about to be changed without permission is a situation in which a user who has no authority to change the learning data attempts to change the learning data.

[0071] For example, in a case where the learning data to be changed is stored in the storage unit 16 without permission, the deletion unit 24 deletes the learning data from the storage unit 16. In a case where the learning data to be changed is stored in a device other than the information processing device 10 (e.g., a PC or a server) without permission, the deletion unit 24 may delete the learning data from the device other than the information processing device 10.

[0072] Furthermore, in the case where the receiving unit 18 receives learning data that does not allow artificial intelligence learning, the deleting unit 24 may delete the learning data.

[0073] The attaching unit 26 is configured to attach attribute information to the content data, indicating whether the content data is permitted to be used as learning data for artificial intelligence. The content data may be, for example, document data (e.g., text data), image data (e.g., still image data or moving image data), music data, audio data, or a combination thereof, and its type, data format, or content is not particularly limited. Content data to which attribute information is attached is an example of learning data to which attribute information is attached.

[0074] The attaching unit 26 may be provided in another device different from the information processing device 10 provided with the changing unit 20. In this case, the processing of the attaching unit 26 is performed by the other device.

[0075] The controller 28 is configured to control the operation of each unit of the information processing device 10 .

[0076] Refer to the following Figure 2 The processing of the information processing apparatus 10 is described. Figure 2 An example of learning data according to the first exemplary embodiment is shown.

[0077] Attribute information 32 is attached to learning data 30. Attribute information 32 is information indicating that artificial intelligence is permitted to learn learning data 30. Attribute information 36 is attached to learning data 34. Attribute information 36 is information indicating that artificial intelligence is not permitted to learn learning data 34.

[0078] In the learning data 30 will be artificial intelligence ( Figure 2 In a case where the artificial intelligence (“AI” in the example above) is allowed to learn the learning data 30 (for example, when the user gives an instruction to cause the artificial intelligence to learn the learning data 30), the changing unit 20 allows the artificial intelligence to learn the learning data 30 or prohibits the artificial intelligence from learning the learning data 30 based on the attribute information 32 attached to the learning data 30. Since the attribute information 32 is information indicating that the artificial intelligence is allowed to learn the learning data 30, the changing unit 20 allows the artificial intelligence to learn the learning data 30.

[0079] Similarly, in the case where learning data 34 is to be learned by artificial intelligence, the changing unit 20 allows the artificial intelligence to learn the learning data 34 or prohibits the artificial intelligence from learning the learning data 34 based on the attribute information 36 attached to the learning data 34. Since the attribute information 36 is information indicating that the artificial intelligence is not permitted to learn the learning data 34, the changing unit 20 prohibits the artificial intelligence from learning the learning data 34. This can prevent the artificial intelligence from learning learning data that is not desired to be learned by the artificial intelligence.

[0080] Refer to the following Figure 3 Describe the process for creating learning data. Figure 3 An example of content data according to the first exemplary embodiment is shown.

[0081] Adding unit 26 adds attribute information to content data 38. Content data 38 to which attribute information has been added is an example of learning data to which attribute information has been added. Attribute information 40 indicates that artificial intelligence is not permitted to learn from content data 38. Content data 38 to which attribute information 40 has been added is prohibited from being learned by artificial intelligence. Attribute information 42 indicates that artificial intelligence is permitted to learn from content data 38. Content data 38 to which attribute information 42 has been added is permitted to be learned by artificial intelligence.

[0082] For example, the user can specify the content data 38 to which attribute information is to be attached by operating the UI unit 14, and specify whether to allow the artificial intelligence to learn the content data 38. If the user allows the artificial intelligence to learn the content data 38, the attachment unit 26 attaches attribute information 42 indicating that the artificial intelligence is allowed to learn the content data 38 to the content data 38 specified by the user. If the user does not allow the artificial intelligence to learn the content data 38, the attachment unit 26 attaches attribute information 40 indicating that the artificial intelligence is not allowed to learn the content data 38 to the content data 38 specified by the user. Needless to say, attribute information can be attached to the learning data by a device other than the information processing device 10 (e.g., a PC or a server).

[0083] The user who specifies whether artificial intelligence is allowed to learn content data 38 may be the user who created content data 38, the user who provided content data 38, the user who purchased content data 38, the user who has the right to use content data 38, the administrator who manages content data 38, or a user other than these users.

[0084] For example, a user who created content data 38 can specify whether to allow artificial intelligence to learn from the content data 38 created by that user. This also applies to other users. For example, when content data 38 is created, appending unit 26 appends attribute information to content data 38. Appending unit 26 may append attribute information to content data 38 when the user instructs it to do so, or may append attribute information to content data 38 upon completion of content data 38 without requiring a user instruction.

[0085] Attachment unit 26 can attach attribute information to content data 38 without requiring user instruction. For example, if content data 38 includes an incorrect answer, attachment unit 26 attaches attribute information 40 to content data 38, indicating that artificial intelligence is not permitted to learn from content data 38. Examples of situations in which content data 38 includes an incorrect answer include situations in which content data 38 used as learning data for translation includes incorrect translation data, situations in which content data 38 used as learning data for image recognition includes information indicating an object different from the object represented by content data 38, or situations in which content data 38 used as learning data for character recognition includes information indicating a character different from the character represented by content data 38.

[0086] The appending unit 26 determines whether the content data 38 includes an incorrect answer by analyzing the content data 38, and, if the content data 38 includes an incorrect answer, appends attribute information 40 indicating that the artificial intelligence is not allowed to learn the content data 38 to the content data 38. The appending unit 26 may append attribute information 40 indicating that the artificial intelligence is not allowed to learn the content data 38 to the content data 38 if the ratio of incorrect answers or the number of occurrences of incorrect answers is equal to or greater than a predetermined threshold.

[0087] If the content data 38 does not include an incorrect answer, the appending unit 26 appends attribute information 42 indicating that the artificial intelligence is allowed to learn the content data 38 to the content data 38. The appending unit 26 may append attribute information 42 indicating that the artificial intelligence is allowed to learn the content data 38 to the content data 38 if the proportion of incorrect answers or the number of occurrences of incorrect answers is less than a predetermined threshold.

[0088] If content data 38 includes information related to security, appending unit 26 may append attribute information 40 to content data 38, indicating that artificial intelligence is not permitted to learn from content data 38. If content data 38 does not include information related to security, appending unit 26 appends attribute information 42 to content data 38, indicating that artificial intelligence is permitted to learn from content data 38. Examples of information related to security include personal information, information related to trade secrets, or information designated as undisclosed by a public institution (e.g., a regulatory agency). Examples of personal information include a specific individual's name, date of birth, user ID, and password. Appending unit 26 analyzes content data 38 to determine whether content data 38 includes information related to security. If content data 38 includes information related to security, appending attribute information 40 to indicate that artificial intelligence is not permitted to learn from content data 38 prevents leakage of information related to security due to artificial intelligence learning from content data 38.

[0089] The attribute information attached to the learning data may include permitted time period information indicating the permitted time period during which the artificial intelligence is permitted to learn the learning data. That is, the time period during which the learning data is permitted to be learned can be limited. Time periods other than the permitted time period are time periods during which the artificial intelligence is prohibited from learning the learning data. The permitted time period is determined, for example, by a date, time zone, or time. If the permitted time period includes the time point (e.g., date or time) at which the artificial intelligence is to learn the learning data, the change unit 20 allows the artificial intelligence to learn the learning data. If the permitted time period does not include the time point at which the artificial intelligence is to learn the learning data, the change unit 20 prohibits the artificial intelligence from learning the learning data. For example, the time period other than the permitted time period may be set to a time period during which the learning data is unknown, and the permitted time period may be set to a time period during which the learning data is known. According to this setting, the artificial intelligence is permitted to learn the learning data during the time period during which the learning data is known, and is prohibited from learning the learning data during the time period during which the learning data is unknown. Needless to say, the permitted time period can be set based on different reasons. The permitted time period can be set by a user who has the authority to set the permitted time period. For example, when creating learning data, attribute information including the allowed time period information is added to the learning data. This process can be performed by the adding unit 26.

[0090] The attribute information attached to the learning data may include number information indicating a limit on the number of times the learning data is learned. The controller 28 counts the number of times the artificial intelligence learns the learning data and attaches the information indicating the number of times to the learning data. In a case where the number of times the learning exceeds the limit on the number of times the learning is performed, the learning of the learning data is prohibited. Even in a case where the user gives an instruction to cause the artificial intelligence to learn the learning data whose number of times of learning exceeds the limit, the controller 28 will not cause the artificial intelligence to learn the learning data. Even in a case where a device other than the information processing device 10 is used for the learning of the artificial intelligence, the learning data whose number of times of learning exceeds the limit will not be learned by the artificial intelligence. The number of times of learning can be set by a user who has the right to set a limit on the number of times of learning. Note that the number of times the learning data whose number of times of learning exceeds the limit can be set to 0 by a user who has the right to set a limit on the number of times of learning. For example, when creating learning data, attribute information including number information is attached to the learning data. This processing can be performed by the attaching unit 26.

[0091] The attribute information attached to learning data may include information indicating whether an AI is permitted to learn from the learning data, based on the intended use of the learning data. For example, if the learning data is data related to or used in the medical field, a predetermined AI is permitted to learn from the learning data, while AIs other than the predetermined AI are prohibited from learning from the learning data. Predetermined AIs are, for example, AIs installed in the information processing device 10 or AIs installed in predetermined devices (e.g., PCs). Predetermined devices are, for example, medical facilities (e.g., buildings at hospitals or medical schools). AIs installed in devices other than these are prohibited from learning from the learning data. Because data related to or used in the medical field sometimes includes personal information, only predetermined AIs are permitted to learn from the learning data, while AIs other than the predetermined AIs are prohibited from learning from the learning data. This can prevent or suppress the leakage of personal information. This also applies to cases where the learning data is data in fields other than medicine. The AIs permitted to learn from the learning data and the devices on which the AIs are installed can be set by a user authorized to make such settings. For example, when creating learning data, attribute information including information indicating the AIs permitted to learn from the learning data and attribute information including information indicating the devices on which the AIs are installed are attached to the learning data. This processing may be performed by the additional unit 26 .

[0092] The attribute information attached to the learning data may include information indicating whether the AI ​​is permitted to learn the learning data, based on the AI's system. An AI system is an algorithm used by the AI. Learning data suitable for learning may vary depending on the AI's algorithm. For example, when an AI learns from learning data, the AI's performance may improve in some cases and decrease in others, depending on the AI's algorithm. Furthermore, the AI's performance may remain unchanged in some cases. Learning data that improves the AI's performance through learning is learning data suitable for the AI ​​to learn. An AI whose performance improves by learning from the learning data is permitted to learn from the learning data. Meanwhile, an AI whose performance does not improve by learning from the learning data (i.e., an AI whose performance degrades or whose performance remains unchanged) is prohibited from learning from the learning data. For example, whether the performance of each AI improves through learning is determined based on the AI's learning history for each piece of learning data. The attribute information attached to the learning data includes information indicating the AI's algorithm whose performance improves by learning from the learning data as information indicating the AI's algorithm permitted to learn from the learning data. For example, when creating learning data, attribute information including information indicating an algorithm of an artificial intelligence whose performance is improved by learning the learning data is attached to the learning data. This processing can be performed by the attaching unit 26. In the case where the algorithm of the artificial intelligence that learns the learning data is an algorithm of artificial intelligence whose performance is improved by learning the learning data, the changing unit 20 allows the artificial intelligence to learn the learning data. In the case where the algorithm of the artificial intelligence that learns the learning data is an algorithm of artificial intelligence whose performance is not improved by learning the learning data, the changing unit 20 prohibits the artificial intelligence from learning the learning data. This allows the artificial intelligence to learn the learning data that improves the performance of the artificial intelligence. Learning data that has not changed the performance of the artificial intelligence through learning can be set as learning data that the artificial intelligence is allowed to learn.

[0093] The attribute information attached to the learning data may include information indicating, for each field of use of the AI, whether the AI ​​is permitted to learn from the learning data. For example, the attribute information may include information indicating the fields in which learning is permitted. Examples of such fields include medicine, education, character recognition, translation, or business. Needless to say, fields other than these may be fields of use for the AI. For example, if information indicating the medical field is included in the attribute information as information indicating the fields in which learning is permitted, the learning data to which this attribute information is attached is learning data permitted to be learned by the AI ​​in the medical field. If the AI ​​learning the learning data is AI for the medical field, the modification unit 20 allows the AI ​​to learn from the learning data. If the AI ​​learning the learning data is AI for a field other than the medical field, the modification unit 20 prohibits the AI ​​from learning from the learning data. This allows the AI ​​to learn from learning data appropriate for the field of use of the AI. For example, when creating learning data, attribute information indicating the fields in which learning is permitted is attached to the learning data. This processing may be performed by the attachment unit 26. Information indicating multiple fields may be included in the attribute information. In this case, artificial intelligence for at least one of the plurality of fields may be allowed to learn the learning data, or artificial intelligence for all of the plurality of fields may be allowed to learn the learning data.

[0094] The changing unit 20 can allow artificial intelligence to learn the learning data or prohibit artificial intelligence from learning the learning data based on the user who uses the learning data. For example, the attribute information attached to the learning data includes the user identification information of the user who is allowed to make the artificial intelligence learn the learning data. In the case where the user identification information of the user who has given the instruction to learn the learning data is included in the attribute information attached to the learning data, the changing unit 20 allows the artificial intelligence to learn the learning data. In the case where the user identification information of the user who has given the instruction to learn the learning data is not included in the learning data, the changing unit 20 prohibits artificial intelligence from learning the learning data. For example, when creating learning data, attribute information including the user identification information of the user who is allowed to make the artificial intelligence learn the learning data is attached to the learning data. This processing can be performed by the attaching unit 26.

[0095] In a first exemplary embodiment, in a case where artificial intelligence is prohibited from learning the learning data, the controller 28 may send information to the creator who created the learning data, asking whether artificial intelligence is allowed to learn the learning data. For example, the attribute information attached to the learning data includes information indicating the email address of the creator who created the learning data or the address of the device used by the creator. The controller 28 sends the query information to the email address or the address of the device. In a case where the creator allows artificial intelligence to learn the learning data, the information indicating permission is sent to the information processing device 10 via email, or is sent from the device used by the creator to the information processing device 10. In this case, the change unit 20 allows artificial intelligence to learn the learning data. In a case where the creator prohibits artificial intelligence from learning the learning data, the information indicating prohibition is sent to the information processing device 10 via email, or is sent from the device used by the creator to the information processing device 10. In this case, the change unit 20 prohibits artificial intelligence from learning the learning data.

[0096] In the first exemplary embodiment, when attribute information indicating prohibition of artificial intelligence from learning the learning data is attached to the learning data received by receiving unit 18, changing unit 20 may prohibit artificial intelligence from learning the learning data. When attribute information indicating prohibition of artificial intelligence from learning the learning data is not attached to the learning data received by receiving unit 18, changing unit 20 allows artificial intelligence to learn the learning data. In this example, changing unit 20 is an example of a prohibiting unit.

[0097] In another example, when attribute information indicating that the artificial intelligence is allowed to learn the learning data is attached to the learning data received by the receiving unit 18, the changing unit 20 may allow the artificial intelligence to learn the learning data. If attribute information indicating that the artificial intelligence is allowed to learn the learning data is not attached to the learning data received by the receiving unit 18, the changing unit 20 prohibits the artificial intelligence from learning the learning data. In this example, the changing unit 20 is an example of a allowing unit.

[0098] The processing of the changing unit 20, the output unit 22, the deleting unit 24, and the adding unit 26 can be performed by artificial intelligence. For example, artificial intelligence instructed to learn the learning data can perform these types of processing.

[0099] Second Exemplary Embodiment

[0100] Reference Figure 4 An information processing apparatus according to a second exemplary embodiment of the present disclosure is described. Figure 4 An example of an information processing apparatus according to the second exemplary embodiment is shown.

[0101] The information processing device 50 according to the second exemplary embodiment is configured to classify learning data learned by artificial intelligence into influential learning data that affects the performance of artificial intelligence and non-influential learning data that does not affect the performance of artificial intelligence. The information processing device 50 can classify the learning data into influential learning data or non-influential learning data and then record the learning data in a recording unit.

[0102] The information processing device 50 may classify the influential learning data learned by the artificial intelligence into improving learning data that improves the performance of the artificial intelligence or degrading learning data that degrades the performance of the artificial intelligence. The information processing device 50 may classify the influential learning data into improving learning data or degrading learning data, and then record the influential learning data in the recording unit.

[0103] The information processing device 50 may classify the learning data learned by the artificial intelligence into improved learning data that improves the performance of the artificial intelligence or non-improved learning data that does not improve the performance of the artificial intelligence. The information processing device 50 may classify the learning data into improved learning data or non-improved learning data, and then record the learning data in the recording unit.

[0104] The information processing device 50 is, for example, a PC, a tablet PC, a smartphone, a mobile phone, or any other type of device (eg, a multifunction printer having functions such as a print function). Needless to say, devices other than these may be the information processing device 50 .

[0105] As in the first exemplary embodiment, the algorithm used for artificial intelligence is not particularly limited, and the algorithm described in the first exemplary embodiment may be used.

[0106] Learning data is data used for artificial intelligence learning. As in the first exemplary embodiment, the learning data can be learning data for supervised learning or learning data for unsupervised learning. The type, data format, and content of the learning data are not particularly limited.

[0107] Artificial intelligence may be installed in the information processing device 50, or may be installed in a device other than the information processing device 50 (for example, a server or a PC). That is, a program for realizing artificial intelligence may be stored in the information processing device 50, or may be stored in a device other than the information processing device 50.

[0108] The configuration of the information processing device 50 is described in detail below.

[0109] The communication unit 52 is a communication interface and has a function of sending information to another device and a function of receiving information from another device. The communication unit 52 may have a wireless communication function or a wired communication function. The communication unit 52 can communicate with another device via a communication path using wireless communication or wired communication. The communication path is, for example, a network such as a LAN or the Internet. The communication unit 52 can communicate with another device without a communication path by using, for example, short-range wireless communication. Examples of short-range wireless communication include Bluetooth (registered trademark), RFID, or NFC.

[0110] For example, in the case where learning data is transmitted from a device other than the information processing device 50 to the information processing device 50, the communication unit 52 receives the learning data. In addition, the communication unit 52 may transmit the learning data to another device.

[0111] The UI unit 54 is a user interface and includes a display and an operating unit. The display is a display device such as a liquid crystal display. The operating unit is an input unit such as a keyboard, input keys, or an operating panel. The UI unit 54 can be a UI unit such as a touch panel that serves as both a display and an operating unit.

[0112] The storage unit 56 is one or more storage areas that store various information. Each storage area is composed of, for example, one or more storage devices (e.g., physical drives such as hard disk drives or memories) provided in the information processing device 50. Learning data can be stored in the storage unit 56.

[0113] The receiving unit 58 is configured to receive learning data. For example, in the case where the communication unit 52 receives learning data transmitted from a device other than the information processing device 50, the receiving unit 58 receives the learning data received by the communication unit 52. In addition, in the case where learning data is provided to the information processing device 50 through the UI unit 54, or in the case where learning data is provided to the information processing device 50 by using a storage device (e.g., a hard disk drive, a USB memory, a CD, a DVD, or any other portable storage medium), the receiving unit 58 receives the provided learning data.

[0114] The determination unit 60 is configured to determine whether the learning data learned by the artificial intelligence affects the performance of the artificial intelligence, and classify the learning data as influential learning data that affects the performance of the artificial intelligence or non-influential learning data that does not affect the performance of the artificial intelligence. This clearly indicates whether the learning data affects the performance of the artificial intelligence. The learning data classified as influential learning data or non-influential learning data can be recorded in the recording unit.

[0115] The concept of influential learning data includes both learning data that improves the performance of the AI ​​and learning data that degrades the performance of the AI. Non-impact learning data is learning data that does not change the performance of the AI ​​through learning. This clearly indicates whether the learning data improves or degrades the performance of the AI. If the difference in the performance of the AI ​​before and after learning the learning data is less than a threshold, the determination unit 60 may classify the learning data as non-impact learning data.

[0116] The determining unit 60 may classify the influential learning data learned by the artificial intelligence into improving learning data that improves the performance of the artificial intelligence or degrading learning data that degrades the performance of the artificial intelligence. The learning data classified as improving learning data or degrading learning data may be recorded in the recording unit.

[0117] The performance of artificial intelligence is the performance of the functions of artificial intelligence, and is, for example, the performance of a character recognition function, the performance of an image recognition function, the performance of a voice recognition function, the performance of an object recognition function, the performance of a translation function, the performance of creativity (for example, creativity in fields such as business or art), or the performance of problem-solving ability (the ability to solve problems in fields such as business). Needless to say, these functions are merely examples of the functions of artificial intelligence, and performance other than these types of performance may be included within the scope of the concept of the performance of artificial intelligence.

[0118] For each artificial intelligence, the determination unit 60 may classify the learning data into influential learning data or non-influential learning data. In addition, for each artificial intelligence, the determination unit 60 may classify the learning data into improving learning data or reducing learning data.

[0119] For each function of artificial intelligence, the determining unit 60 may classify the learning data into impact learning data or non-impact learning data. In addition, for each function of artificial intelligence, the determining unit 60 may classify the learning data into improvement learning data or reduction learning data.

[0120] The determination unit 60 performs a test to determine whether the learning data to be determined affects the performance of the artificial intelligence. For example, the determination unit 60 causes the artificial intelligence to learn the learning data to be determined, and determines whether the performance of the artificial intelligence changes before and after the learning.

[0121] For example, in the case of determining whether the learning data to be determined affects the performance of the character recognition function, the determination unit 60 first gives predetermined document data (for example, document data for testing) to an artificial intelligence that has not yet learned the learning data to be determined, so that the artificial intelligence recognizes the characters included in the document data and calculates the character recognition accuracy (for example, the character recognition rate).

[0122] Then, the determination unit 60 causes the artificial intelligence to learn the learning data to be determined. Then, the determination unit 60 gives the document data to the artificial intelligence that has learned the learning data to be determined, causes the artificial intelligence to recognize characters included in the document data, and calculates character recognition accuracy.

[0123] The determination unit 60 determines whether the character recognition accuracy changes before and after learning. In the case where the character recognition accuracy changes before and after learning, the determination unit 60 determines that the learning data to be determined is learning data that affects the performance of the artificial intelligence regarding the character recognition function, and in the case where the character recognition accuracy does not change before and after learning, the determination unit 60 determines that the learning data to be determined is learning data that does not affect the performance of the artificial intelligence regarding the character recognition function. In the case where the character recognition accuracy changes before and after learning, the determination unit 60 classifies the learning data to be determined as influential learning data, and in the case where the character recognition accuracy does not change before and after learning, the determination unit 60 classifies the learning data to be determined as non-influential learning data.

[0124] In addition, in a case where the character recognition accuracy after learning is higher than the character recognition accuracy before learning, the determination unit 60 determines that the learning data to be determined is learning data that improves the performance of the artificial intelligence regarding the character recognition function, and in a case where the character recognition accuracy after learning is lower than the character recognition accuracy before learning, the determination unit 60 determines that the learning data to be determined is learning data that reduces the performance of the artificial intelligence regarding the character recognition function. In a case where the character recognition accuracy after learning is higher than the character recognition accuracy before learning, the determination unit 60 classifies the learning data to be determined as improving learning data, and in a case where the character recognition accuracy after learning is lower than the character recognition accuracy before learning, the determination unit 60 classifies the learning data to be determined as not improving learning data.

[0125] As in the case of the performance of the character recognition function, it is also determined whether the learning data to be determined affects other types of performance of the artificial intelligence. For example, before and after the artificial intelligence learns the image data as learning data, the image recognition accuracy (for example, the image recognition rate) of the image data for testing is calculated, and the learning data is classified based on the calculation result. For example, it is assumed that the artificial intelligence that has not yet learned the image data presenting a cat and cannot recognize the cat by analyzing the image data presenting a cat learns the image data presenting a cat as learning data. In the case where the artificial intelligence becomes able to recognize the cat by analyzing the image data presenting a cat through learning, this indicates that the performance of the image recognition function has been improved. In the case where the artificial intelligence fails to recognize the cat by analyzing the image data presenting a cat through learning, this indicates that the performance of the image recognition function has not been improved.

[0126] The determination unit 60 can determine whether the learning data learned by the artificial intelligence improves the performance of the artificial intelligence, and classify the learning data into improved learning data that improves the performance of the artificial intelligence or non-improved learning data that does not improve the performance of the artificial intelligence. This clearly shows whether the learning data improves the performance of the artificial intelligence. The learning data classified as improved learning data or non-improved learning data can be recorded in the recording unit. When the difference between the performance of the artificial intelligence before learning the learning data and the performance of the artificial intelligence after learning the learning data is less than a threshold, the determination unit 60 can classify the learning data as non-improved learning data. For each artificial intelligence, the determination unit 60 can determine whether the learning data learned by the artificial intelligence improves the performance of the artificial intelligence, and for each function of the artificial intelligence, classify the learning data as improved learning data or non-improved learning data.

[0127] Note that the learning data to be determined may be learning data designated by the user, or may be learning data designated in advance as learning data to be determined.

[0128] The appending unit 62 is configured to append determination result information indicating the determination result of the determination unit 60 to the learning data to be determined. The appending unit 62 may append the determination result information to the learning data to be determined as accompanying information, or may embed the determination result information in the learning data to be determined. Note that the appending unit 62 may append the attribute information according to the first exemplary embodiment to the learning data.

[0129] The controller 64 is configured to control the operation of each unit of the information processing device 50. The controller 64 includes a recording controller 66 and a learning controller 68.

[0130] The recording controller 66 is configured to record learning data with determination result information attached in a recording unit. For example, the recording controller 66 classifies the learning data to be determined as influential learning data or non-influential learning data, and then records the learning data to be determined in the recording unit. The recording controller 66 may classify influential learning data as improving learning data or reducing learning data, and then record the influential learning data in the recording unit. The recording controller 66 may classify learning data as improving learning data or non-improving learning data, and then record the learning data in the recording unit. The recording unit for recording learning data may be provided in the information processing device 50, or may be provided in a device other than the information processing device 50 (e.g., a server or a PC). The recording unit for recording learning data may be specified by a user or may be predetermined.

[0131] The recording controller 66 may record learning history information indicating the learning history of the AI ​​in association with the learning data to be determined in the recording unit. The learning history information may be attached to the learning data to be determined. The learning history information associated with the learning data to be determined is information indicating the learning data previously learned by the AI ​​that has already learned the learning data to be determined. The learning history information is information indicating, for each AI, the correspondence between AI identification information used to identify the AI ​​and learning data identification information used to identify the learning data previously learned by the AI. Furthermore, the learning history information may include information indicating the order in which the AI ​​learned each piece of learning data. For example, learning data identification information used to identify the learning data previously learned by the AI ​​is associated with the AI, and the recording controller 66 associates the learning data identification information associated with the AI ​​that has learned the learning data to be determined with the learning data to be determined. For example, if AI α learns learning data A, B, and C in the order A, B, and C, information indicating that AI α learned learning data A, B, and C in the order A, B, and C is associated with learning data C as learning history information.

[0132] The learning controller 68 is configured to cause the artificial intelligence to learn learning data. For example, the learning controller 68 causes the artificial intelligence to learn influential learning data or non-influential learning data that has not yet been learned by the artificial intelligence, based on the purpose of the artificial intelligence learning. More specifically, the learning controller 68 causes another artificial intelligence, different from the artificial intelligence that has already learned the learning data to be determined, to learn influential learning data or non-influential learning data classified according to the learning results of the artificial intelligence, based on the purpose of the other artificial intelligence. Specifically, the learning controller 68 causes the other artificial intelligence to learn improved learning data or reduced learning data classified according to the learning results of the artificial intelligence, based on the learning purpose of the other artificial intelligence. The learning controller 68 can cause the other artificial intelligence to learn improved learning data or non-improved learning data classified according to the learning results of the artificial intelligence, based on the learning purpose of the other artificial intelligence. For example, the other artificial intelligence can be specified by the user or can be pre-specified.

[0133] The processing of learning controller 68 will be described using a specific example. For example, if the performance of function a of artificial intelligence α has been improved by learning from learning data A, and another artificial intelligence β is learning to improve function a of artificial intelligence β, learning controller 68 causes artificial intelligence β to learn from learning data A. Specifically, since learning data A is classified as improved learning data capable of improving the performance of function a, learning controller 68 causes artificial intelligence β to learn from learning data A, which is improved learning data for improving the performance of function a. In other words, if the learning goal of artificial intelligence β is to bring the performance of artificial intelligence β closer to that of artificial intelligence α, artificial intelligence β is caused to learn from learning data A, which improves the performance of artificial intelligence α.

[0134] Meanwhile, if the purpose of learning artificial intelligence β is to prevent the performance of artificial intelligence β from approaching that of artificial intelligence α, that is, to make the performance of artificial intelligence β different from that of artificial intelligence α, learning controller 68 causes artificial intelligence β to learn degrading learning data that degrades the performance of artificial intelligence α, learning data that has no effect learned by artificial intelligence α, or learning data that has no improvement learned by artificial intelligence α. In this way, artificial intelligence β having different performance from artificial intelligence α is created.

[0135] The learning objective of the AI ​​is, for example, specified by the user. For example, the user specifies the AI ​​(e.g., AI β) to be trained on learning data and the learning objective of the AI ​​by operating the UI unit 54. The learning controller 68 causes the AI ​​specified by the user to learn the learning data that matches the learning objective specified by the user.

[0136] The learning controller 68 can cause other artificial intelligences to learn from the influential learning data classified by learning an artificial intelligence having a function corresponding to the function of the other artificial intelligence. The function corresponding to the function of the other artificial intelligence can be the same function as the function of the other artificial intelligence, or can be a function similar to the function of the other artificial intelligence (for example, a function whose difference from the function of the other artificial intelligence is equal to or less than a threshold value).

[0137] For example, if the performance of function a of artificial intelligence α is improved by learning learning data A, learning controller 68 causes another artificial intelligence β having function a to learn learning data A. In other words, learning controller 68 causes artificial intelligence β to learn the improved learning data that improved the performance of artificial intelligence α. For example, artificial intelligences α and β can be specified by the user or can be specified in advance.

[0138] In addition, the learning controller 68 can cause other artificial intelligences to learn from influential learning data classified by learning of an artificial intelligence having a learning history corresponding to the learning history of the other artificial intelligence. The learning history corresponding to the learning history of the other artificial intelligence can be a learning history that is the same as the learning history of the other artificial intelligence, or can be a learning history that is similar to the learning history of the other artificial intelligence (for example, a learning history whose difference from the learning history of the other artificial intelligence is equal to or less than a threshold). The learning history is the history of learning data that the artificial intelligence has learned in the past.

[0139] Refer to the following Figures 5 to 7 The processing of the information processing device 50 is described. Figures 5 to 7 An example of learning data according to the second exemplary embodiment is shown.

[0140] like Figure 5 As shown, for example, the learning data A is designated as the learning data to be determined, and the artificial intelligence α ( Figure 5 The AI ​​(α) in the figure is designated as the artificial intelligence that learns the learning data A. For example, when the user gives an instruction to determine the learning data A by operating the UI unit 54, the determination unit 60 causes the artificial intelligence α to learn the learning data A to be determined, and performs a test for determining whether the learning data A affects the performance of the artificial intelligence α. That is, the determination unit 60 determines whether the performance of the artificial intelligence α changes before and after learning the learning data A. If the performance of the artificial intelligence α changes, the determination unit 60 determines that the learning data A is influential learning data that affects the performance of the artificial intelligence α. If the performance of the artificial intelligence α does not change, the determination unit 60 determines that the learning data A is non-influential learning data that does not affect the performance of the artificial intelligence α.

[0141] In the case where the learning data A is classified as influential learning data, the appending unit 62 appends determination result information 70 indicating that the learning data A is influential learning data to the learning data A, such as Figure 5 In the case where the learning data A is classified as non-influence learning data, the appending unit 62 appends determination result information 72 indicating that the learning data A is non-influence learning data to the learning data A, as shown. Figure 5 shown.

[0142] The recording controller 66 may cause the learning data A classified as the influential learning data or the non-influential learning data to be recorded in the recording unit.

[0143] The determination unit 60 can determine whether the performance of the artificial intelligence α is improved before and after learning the learning data A. In the case where the performance of the artificial intelligence α after learning the learning data A is improved from the performance of the artificial intelligence α before learning the learning data A, the determination unit 60 determines that the learning data A is improved learning data that improves the performance of the artificial intelligence α. In this case, Figure 6 As shown, the appending unit 62 appends determination result information 74 indicating that the learning data A is improved learning data to the learning data A. In a case where the performance of the artificial intelligence α after learning the learning data A is lower than the performance of the artificial intelligence α before learning the learning data A, the determination unit 60 determines that the learning data A is reduced learning data that reduces the performance of the artificial intelligence α. In this case, Figure 6 As shown, the adding unit 62 adds determination result information 76 indicating that the learning data A is reduced learning data to the learning data A.

[0144] The recording controller 66 may cause the learning data A classified as improvement learning data or reduction learning data to be recorded in the recording unit.

[0145] Furthermore, in the case where the learning data A is classified as improvement learning data, the learning controller 68 may make the designated other artificial intelligence β ( Figure 6 The AI ​​(β) in the learning controller 68 learns the learning data A. This can improve the performance of the artificial intelligence β. At the same time, if the learning data A is classified as degraded learning data, the learning controller 68 can prohibit another artificial intelligence β from learning the learning data A. This prevents the performance of the artificial intelligence β from being degraded.

[0146] In the case where the performance of the artificial intelligence α after learning the learning data A does not improve from the performance of the artificial intelligence α before learning the learning data A, the determination unit 60 may determine that the learning data A is non-improvement learning data that does not improve the performance of the artificial intelligence α. In the case where the performance of the artificial intelligence α after learning the learning data A improves from the performance of the artificial intelligence α before learning the learning data A, the determination unit 60 determines that the learning data A is improved learning data that improves the performance of the artificial intelligence α. In the case of classifying the learning data A as improved learning data, the appending unit 62 appends the determination result information 78 indicating that the learning data A is improved learning data to the learning data A, such as Figure 7 In the case where the learning data A is classified as non-improvement learning data, the appending unit 62 appends the determination result information 80 indicating that the learning data A is non-improvement learning data to the learning data A, as shown. Figure 7 shown.

[0147] The recording controller 66 may cause the learning data A classified as the improvement learning data or the non-improvement learning data to be recorded in the recording unit.

[0148] Furthermore, in the case where the learning data A is classified as improvement learning data, the learning controller 68 may make another artificial intelligence β ( Figure 7 The AI ​​(β) in the learning controller 68 learns the learning data A. This can improve the performance of the artificial intelligence β. At the same time, if the learning data A is classified as non-improvement learning data, the learning controller 68 prohibits another artificial intelligence β from learning the learning data A. This prevents the performance of the artificial intelligence β from being degraded.

[0149] For each function of each AI, the determination unit 60 can determine the impact of the learning data on the AI ​​and create management information (e.g., a database) for managing the determination results. The management information can be stored in the storage unit 56 or in a device other than the information processing device 50.

[0150] Figure 8 An example of a database is shown as an example of management information. Figure 8 The database shown is a database indicating the determination results on the influence given to artificial intelligence by learning data A. In this database, information indicating the determination results on the influence given to each artificial intelligence by learning data A is managed for each function of the artificial intelligence.

[0151] For example, AI α ( Figure 8 AI (α)) and artificial intelligence β ( Figure 8 Each of the AI ​​(β) in

[15] has functions such as character recognition, translation, creativity, and problem-solving. Determination result A indicates a significant improvement in performance. Determination result B indicates a slight improvement in performance. Determination result C indicates no change in performance. Determination result D indicates a decrease in performance.

[0152] As a result of learning data A, the character recognition rate of artificial intelligence α significantly improved, the translation accuracy of artificial intelligence α slightly improved, the creativity of artificial intelligence α remained unchanged, and the problem-solving ability of artificial intelligence α decreased. In other words, the performance of the character recognition function of artificial intelligence α significantly improved, the performance of the translation function of artificial intelligence α slightly improved, the creativity of artificial intelligence α remained unchanged, and the problem-solving ability of artificial intelligence α decreased.

[0153] As a result of learning the learning data A, the character recognition rate, translation accuracy, and creativity of artificial intelligence β did not change, and the problem-solving ability of artificial intelligence β decreased. In other words, the performance of the character recognition function, translation function, and creativity of artificial intelligence β did not change, and the performance of the problem-solving ability of artificial intelligence β decreased.

[0154] Since the determination results are managed as described above, it is possible to evaluate the influence of the learning data on the performance of each artificial intelligence. Figure 8 In the example shown, the comparison between AI α and AI β shows that, although both AI α and AI β learned the same learning data A, the performance of AI β did not improve compared to the performance of AI α. In other words, the comparison shows that the performance of AI α improved compared to the performance of AI β.

[0155] Because there may be differences in learning history between AI α and AI β, it can be said that the impact of learning data A on the performance of AI cannot necessarily be judged based solely on the determination results. However, the determination results can be used as an indicator for evaluating the impact of learning data A on the performance of AI.

[0156] Furthermore, when managing the learning history of AI, it is possible to estimate the learning history of AI that can improve its performance by learning from learning data A. For example, if AI α and AI β use the same algorithm, it is possible to estimate the learning history of AI that can improve its performance by learning from learning data A.

[0157] In addition, when the artificial intelligence α and the artificial intelligence β use different algorithms, it is possible to estimate the algorithm that improves the performance of the artificial intelligence by learning the learning data A. Figure 8 In the example shown, the performance of artificial intelligence α is improved compared to the performance of artificial intelligence β, so it can be estimated that the algorithm that improves the performance of artificial intelligence by learning the learning data A is the algorithm used by artificial intelligence α.

[0158] In addition, when the algorithm used by the artificial intelligence, as well as the use start time, learning start time, and learning time period of the artificial intelligence, are managed as a database for each artificial intelligence, it is possible to estimate whether the factor that affects the artificial intelligence through learning the learning data is the learning data or reasons other than the learning data (for example, the algorithm or learning history).

[0159] Modifications of the second exemplary embodiment are described below.

[0160] First Modification of the Second Exemplary Embodiment

[0161] The following describes a first variation of the second exemplary embodiment. In the first variation of the second exemplary embodiment, determination unit 60 causes the AI ​​to learn a plurality of different learning data combinations and determines whether the combination of the plurality of learning data affects the performance of the AI. Determination unit 60 then classifies the combination of the plurality of learning data as either a combination that affects the performance of the AI ​​or a combination that does not affect the performance of the AI.

[0162] The determination unit 60 may determine whether the combination of the plurality of learning data improves the performance of artificial intelligence, and classify the combination of the plurality of learning data as a combination that improves the performance of artificial intelligence or a combination that degrades the performance of artificial intelligence.

[0163] The determination unit 60 may determine whether the combination of the plurality of learning data improves the performance of artificial intelligence, and classify the combination of the plurality of learning data into a combination that improves the performance of artificial intelligence or a combination that does not improve the performance of artificial intelligence.

[0164] The adding unit 62 adds, to the combination of the plurality of pieces of learning data, determination result information indicating the determination result of the determining unit 60. The recording controller 66 may cause the combination to which the determination result information is added to be recorded in the recording unit.

[0165] The combination of enabling artificial intelligence to learn multiple learning data means giving multiple learning data to artificial intelligence at the same time and enabling artificial intelligence to learn multiple learning data, or giving each learning data included in the multiple learning data to artificial intelligence sequentially and enabling artificial intelligence to learn each learning data sequentially.

[0166] Refer to the following Figures 9 to 11 A process according to a first modification of the second exemplary embodiment is described. Figures 9 to 11 An example of learning data according to the first modification of the second exemplary embodiment is shown.

[0167] like Figure 9 As shown, for example, a combination of learning data A and B is designated as learning data to be determined, and artificial intelligence α ( Figure 9The AI(α) in is designated as an artificial intelligence that learns the combination of learning data A and B. For example, when the user gives an instruction to determine the combination of learning data A and B by operating the UI unit 54, the determination unit 60 causes the artificial intelligence α to learn the combination of learning data A and B to be determined, and performs a test for determining whether the combination of learning data A and B affects the performance of the artificial intelligence α. That is, the determination unit 60 determines whether the performance of the artificial intelligence α changes before and after learning the combination of learning data A and B. In the case where the performance of the artificial intelligence α changes, the determination unit 60 determines that the combination of learning data A and B is a combination that affects the performance of the artificial intelligence α. In the case where the performance of the artificial intelligence α does not change, the determination unit 60 determines that the combination of learning data A and B is a combination that does not affect the performance of the artificial intelligence α.

[0168] In the case where the combination of the learning data A and B is classified as a combination that affects the performance of the artificial intelligence α, the appending unit 62 appends determination result information 82 indicating that the combination of the learning data A and B is a combination that affects the performance of the artificial intelligence α to the combination of the learning data A and B, such as Figure 9 In the case where the combination of the learning data A and B is classified as a combination that does not affect the performance of the artificial intelligence α, the appending unit 62 appends determination result information 84 indicating that the combination of the learning data A and B is a combination that does not affect the performance of the artificial intelligence α to the combination of the learning data A and B, as shown. Figure 9 shown.

[0169] The recording controller 66 may cause the combination of the learning data A and B classified as a combination that affects the performance of the artificial intelligence α or a combination that does not affect the performance of the artificial intelligence α to be recorded in the recording unit.

[0170] The determination unit 60 may determine whether the performance of the artificial intelligence α is improved before and after learning the combination of the learning data A and B. In a case where the performance of the artificial intelligence α after learning the combination of the learning data A and B is improved from the performance of the artificial intelligence α before learning the combination of the learning data A and B, the determination unit 60 determines that the combination of the learning data A and B is a combination that improves the performance of the artificial intelligence α. In this case, Figure 10 As shown, the adding unit 62 adds the determination result information 86 indicating that the combination of the learning data A and B is a combination that improves the performance of the artificial intelligence α to the combination of the learning data A and B. In a case where the performance of the artificial intelligence α after learning the combination of the learning data A and B is lower than the performance of the artificial intelligence α before learning the combination of the learning data A and B, the determining unit 60 determines that the combination of the learning data A and B is a combination that reduces the performance of the artificial intelligence α. In this case, Figure 10As shown, the adding unit 62 adds determination result information 88 indicating that the combination of the learning data A and B is a combination that degrades the performance of the artificial intelligence α to the combination of the learning data A and B.

[0171] The recording controller 66 may cause the combination of the learning data A and B classified as a combination that improves the performance of the artificial intelligence α or a combination that reduces the performance of the artificial intelligence α to be recorded in the recording unit.

[0172] If the combination of learning data A and B improves the performance of artificial intelligence α, the learning controller 68 may cause another designated artificial intelligence (e.g., artificial intelligence β) to learn the combination of learning data A and B. Meanwhile, if the combination of learning data A and B degrades the performance of artificial intelligence α, the learning controller 68 may prohibit another artificial intelligence from learning the combination of learning data A and B.

[0173] In a case where the performance of the artificial intelligence α after learning the combination of the learning data A and B does not improve from the performance of the artificial intelligence α before learning the combination of the learning data A and B, the determination unit 60 may determine that the combination of the learning data A and B is a combination that does not improve the performance of the artificial intelligence α. In a case where the performance of the artificial intelligence α after learning the combination of the learning data A and B improves from the performance of the artificial intelligence α before learning the combination of the learning data A and B, the determination unit 60 determines that the combination of the learning data A and B is a combination that improves the performance of the artificial intelligence α. In a case where the combination of the learning data A and B is classified as a combination that improves the performance of the artificial intelligence α, the appending unit 62 appends determination result information 90 indicating that the combination of the learning data A and B is a combination that improves the performance of the artificial intelligence α to the combination of the learning data A and B, as Figure 11 In the case where the combination of learning data A and B is classified as a combination that does not improve the performance of artificial intelligence α, the appending unit 62 appends determination result information 92 indicating that the combination of learning data A and B is a combination that does not improve the performance of artificial intelligence α to the combination of learning data A and B, as shown. Figure 11 shown.

[0174] The recording controller 66 may cause the combination of the learning data A and B classified as a combination that improves the performance of the artificial intelligence α or a combination that does not improve the performance of the artificial intelligence α to be recorded in the recording unit.

[0175] In the case where the combination of learning data A and B is classified as a combination that does not improve the performance of artificial intelligence α, the learning controller 68 can prohibit another artificial intelligence from learning the combination of learning data A and B.

[0176] The determination unit 60 can determine the influence of the combination of multiple learning data on each function of each artificial intelligence and create management information (e.g., a database) for managing the determination results. The management information can be recorded in the storage unit 56 or in a device other than the information processing device 10.

[0177] Figure 12 An example of a database is shown as an example of management information. Figure 12 The database shown is a database indicating the determination results of the influence of the combination of learning data A and B on artificial intelligence. In this database, for each function of artificial intelligence, information indicating the determination results of the influence of the combination of learning data A and B on each artificial intelligence is managed. The meanings of the determination results A, B, C, and D are the same as Figure 8 The determination results shown have the same meaning.

[0178] As a result of learning the combination of learning data A and B, the character recognition rate of artificial intelligence α significantly improved, the translation accuracy of artificial intelligence α slightly improved, the creativity of artificial intelligence α remained unchanged, and the problem-solving ability of artificial intelligence α decreased. In other words, the performance of the character recognition function of artificial intelligence α significantly improved, the performance of the translation function of artificial intelligence α slightly improved, the creativity of artificial intelligence α remained unchanged, and the problem-solving ability of artificial intelligence α decreased.

[0179] As a result of learning the combination of learning data A and B, the character recognition rate, translation accuracy, and creativity of artificial intelligence β did not change, and the problem-solving ability of artificial intelligence β decreased. In other words, the performance of the character recognition function, translation function, and creativity of artificial intelligence β did not change, and the performance of the problem-solving ability of artificial intelligence β decreased.

[0180] Since the determination results are managed as described above, it is possible to evaluate the influence of the combination of multiple pieces of learning data on the performance of each artificial intelligence. Figure 12 In the example shown, the comparison between AI α and AI β shows that, although both AI α and AI β learned the same combination of learning data A and B, the performance of AI β did not improve compared to the performance of AI α. In other words, the comparison shows that the performance of AI α improved compared to the performance of AI β.

[0181] In the case where the artificial intelligence sequentially learns the learning data A and B, the determination results obtained when the order is changed can be managed in the database. That is, the determination results obtained when the artificial intelligence learns the learning data A and B in the order of learning data A and B and the determination results obtained when the artificial intelligence learns the learning data A and B in the order of learning data B and A can be managed in the database.

[0182] Although the determination unit 60 causes the artificial intelligence to learn a combination of two items of learning data in the above example, the determination unit 60 may cause the artificial intelligence to learn a combination of three or more items of learning data and determine the influence of the learning.

[0183] The pieces of learning data included in the combination of multiple learning data may be learning data of the same type or the same format, or may be learning data of different types or different formats. For example, a combination of multiple pieces of document data or a combination of multiple pieces of image data may be used as the combination of multiple learning data. Alternatively, a combination of document data and image data may be used as the combination of multiple learning data. These combinations are merely examples, and the pieces of learning data included in the combination of multiple learning data may be specified by the user.

[0184] Second Modification of the Second Exemplary Embodiment

[0185] A second variation of the second exemplary embodiment is described below. In this second variation of the second exemplary embodiment, determination unit 60 determines the lifespan of the AI. As used herein, "lifespan" means that even when the AI ​​learns from learning data, the performance of the AI ​​does not improve, or even if the performance of the AI ​​improves, the improvement is less than a threshold. Determining that the AI ​​has reached its lifespan is used as a criterion for prompting the user to replace the AI ​​or change the algorithm.

[0186] For example, Figure 13 As shown, determination unit 60 causes artificial intelligence α to learn learning data C that has not yet been learned by artificial intelligence α, and determines whether the performance of artificial intelligence α has improved due to the learning. If the performance of artificial intelligence α has improved, determination unit 60 determines that artificial intelligence α has not yet reached the end of its lifespan. If the performance of artificial intelligence α has not improved, determination unit 60 determines that artificial intelligence α has reached the end of its lifespan. Even if the performance of artificial intelligence α has improved, determination unit 60 may determine that artificial intelligence α has reached the end of its lifespan if the improvement is less than a threshold value.

[0187] If it is determined that the AI ​​α has reached the end of its lifespan, the controller 64 may output information such as a recommendation to replace the AI ​​or a recommendation to change the AI ​​algorithm. For example, the controller 64 may display the information indicating the recommendation on the display of the UI unit 54 or output a sound.

[0188] Note that the determination unit 60 may cause the artificial intelligence α to learn the same learning data a plurality of times continuously or at predetermined time intervals, and determine the life end of the artificial intelligence α based on the learning results.

[0189] The determination unit 60 may cause the artificial intelligence α to learn a plurality of different learning data and determine the lifespan of the artificial intelligence α based on the learning results. For example, if the number of each learning data item that improves the performance of the artificial intelligence α is less than a threshold, the determination unit 60 determines that the artificial intelligence α has reached the lifespan end. If the number of each learning data item that improves the performance of the artificial intelligence α is equal to or greater than the threshold, the determination unit 60 determines that the artificial intelligence α has not yet reached the lifespan end.

[0190] Figure 14 Another example is shown. In this other example, Figure 14 As shown, determination unit 60 causes each of AIs α and β to learn learning data C that has not yet been learned by either AI α or β, and compares the learning results of AI α with the learning results of AI β. If the difference between the learning results of AI α and AI β is less than a threshold, determination unit 60 determines that neither AI α or β has reached the end of its lifespan. If the difference between the learning results of AI α and AI β is equal to or greater than a threshold, determination unit 60 determines that the AI ​​with the lowest learning effectiveness among AIs α and β has reached the end of its lifespan. An AI with the lowest learning effectiveness is one whose performance has not improved compared to the performance of other AIs, whose performance has not improved while the performance of other AIs has improved, or whose performance has declined more than that of other AIs despite the decline in the performance of other AIs.

[0191] Artificial intelligences α and β may be artificial intelligences with identical or similar learning histories, or may be artificial intelligences with learning histories that are neither identical nor similar. A case where the learning histories are similar is a case where the difference between the learning histories of artificial intelligence α and artificial intelligence β is less than a threshold.

[0192] The processes of the determination unit 60, the addition unit 62, the recording controller 66, and the learning controller 68 may be performed by artificial intelligence. For example, artificial intelligence instructed to learn the learning data may perform these types of processes.

[0193] The functions of the units of the information processing devices 10 and 50 are implemented, for example, by the cooperation of hardware and software. Specifically, the information processing devices 10 and 50 have one or more processors such as a CPU (not shown). One or more processors read out and execute the program stored in the storage device (not shown), thereby implementing the functions of the units of the information processing devices 10 and 50. The program is stored in the storage device via a recording medium such as a CD or DVD or a communication path such as a network. In another example, the functions of the units of the information processing devices 10 and 50 can be implemented by hardware resources such as a processor, an electronic circuit or an application-specific integrated circuit (ASIC). The functions of the units of the information processing devices 10 and 50 can be implemented using a device such as a memory. In yet another example, the functions of the units of the information processing devices 10 and 50 can be implemented, for example, by a digital signal processor (DSP) or a field programmable gate array (FPGA).

[0194] The foregoing description of exemplary embodiments of the present disclosure has been provided for the purposes of illustration and description and is not intended to be exhaustive or to limit the disclosure to the precise forms disclosed. Obviously, many variations and modifications will be apparent to those skilled in the art. The above embodiments have been selected and described in order to best explain the principles of the present disclosure and its practical application, so that others skilled in the art can understand the disclosure of the various embodiments and the various modifications suitable for the intended specific use. The scope of the present disclosure is intended to be defined by the appended claims and their equivalents.

Claims

1. An information processing device, comprising: a receiving unit, wherein the receiving unit receives learning data; a changing unit that changes a learning process of the artificial intelligence based on information that is appended to the learning data and indicates whether the artificial intelligence is permitted to learn the learning data; a determining unit, when the information indicates that the artificial intelligence is permitted to learn the learning data, configured to determine that the learning data is improved learning data that improves the performance of the artificial intelligence, if the performance of the artificial intelligence after learning the learning data is improved from the performance of the artificial intelligence before learning the learning data; as well as A learning controller enables another artificial intelligence to learn the improved learning data.

2. The information processing device according to claim 1, wherein In a case where the information indicates that the artificial intelligence is not allowed to learn the learning data, the changing unit prohibits the artificial intelligence from learning the learning data.

3. The information processing device according to claim 2, wherein In a case where the information indicates that the artificial intelligence is not allowed to learn the learning data, the changing unit allows the artificial intelligence to learn the learning data if a purpose of learning using the learning data satisfies a criterion.

4. The information processing device according to claim 1 further includes an output unit that outputs a warning when the information indicates that the artificial intelligence is not allowed to learn the learning data and the learning data will be changed without permission.

5. The information processing device according to claim 1 further includes a deletion unit, which deletes the learning data when the information indicates that the artificial intelligence is not allowed to learn the learning data and the learning data will be changed without permission. The information processing device according to claim 1 , wherein: In a case where the information indicates that the artificial intelligence is permitted to learn the learning data, the changing unit permits the artificial intelligence to learn the learning data.

7. The information processing device according to claim 6, wherein The information also includes information indicating a time period during which the artificial intelligence is allowed to learn the learning data.

8. The information processing device according to claim 6, wherein The information also includes information indicating a limit on the number of times learning is performed on the learning data.

9. The information processing apparatus according to any one of claims 1 to 8, wherein: The information also includes information indicating whether the artificial intelligence is allowed to learn the learning data according to the purpose of the learning data.

10. The information processing apparatus according to any one of claims 1 to 8, wherein The information also includes information indicating whether the artificial intelligence is allowed to learn the learning data according to the system indication of the artificial intelligence.

11. The information processing apparatus according to any one of claims 1 to 8, wherein The information includes information indicating whether the artificial intelligence is allowed to learn the learning data for each use field of the artificial intelligence.

12. An information processing device, comprising: a receiving unit, wherein the receiving unit receives learning data; a allowing unit configured to allow the artificial intelligence to learn the learning data if information indicating that the artificial intelligence is allowed to learn the learning data is appended to the learning data; a determining unit configured to determine that the learning data is improved learning data that improves the performance of the artificial intelligence when the performance of the artificial intelligence after learning the learning data is improved from the performance of the artificial intelligence before learning the learning data; as well as A learning controller enables another artificial intelligence to learn the improved learning data.

13. An information processing device, comprising: an appending unit that appends information indicating whether use of the content data as learning data for artificial intelligence is permitted to the content data; a determining unit configured to determine that the learning data is improved learning data that improves the performance of the artificial intelligence when the performance of the artificial intelligence after learning the learning data is improved from the performance of the artificial intelligence before learning the learning data; as well as A learning controller enables another artificial intelligence to learn the improved learning data.

14. A non-transitory computer-readable medium storing a program for causing a computer to execute a process for image processing, the process comprising: receiving learning data; changing a learning process of the artificial intelligence according to information attached to the learning data and indicating whether the artificial intelligence is permitted to learn the learning data; When the information indicates that the artificial intelligence is permitted to learn the learning data, if the performance of the artificial intelligence after learning the learning data is improved from the performance of the artificial intelligence before learning the learning data, determining that the learning data is improved learning data that improves the performance of the artificial intelligence; and Another artificial intelligence is enabled to learn from the improved learning data.

15. A non-transitory computer-readable medium storing a program for causing a computer to execute a process for image processing, the process comprising: appending information indicating whether use of the content data as learning data for artificial intelligence is permitted to the content data; determining that the learning data is improved learning data that improves the performance of the artificial intelligence if the performance of the artificial intelligence after learning the learning data is improved from the performance of the artificial intelligence before learning the learning data; and Another artificial intelligence is enabled to learn from the improved learning data.

16. A computer program product comprising computer instructions, wherein when the computer instructions are executed by a processor, a method for image processing is performed, the method comprising the following steps: receiving learning data; changing a learning process of the artificial intelligence according to information attached to the learning data and indicating whether the artificial intelligence is permitted to learn the learning data; When the information indicates that the artificial intelligence is permitted to learn the learning data, if the performance of the artificial intelligence after learning the learning data is improved from the performance of the artificial intelligence before learning the learning data, determining that the learning data is improved learning data that improves the performance of the artificial intelligence; and Another artificial intelligence is enabled to learn from the improved learning data.

17. A computer program product comprising computer instructions, wherein when the computer instructions are executed by a processor, a method for image processing is performed, the method comprising the following steps: appending information indicating whether use of the content data as learning data for artificial intelligence is permitted to the content data; determining that the learning data is improved learning data that improves the performance of the artificial intelligence if the performance of the artificial intelligence after learning the learning data is improved from the performance of the artificial intelligence before learning the learning data; and Another artificial intelligence is enabled to learn from the improved learning data.

Images