WO2017175434A1 - 情報処理装置、情報処理方法および情報提供方法 - Google Patents
情報処理装置、情報処理方法および情報提供方法 Download PDFInfo
- Publication number
- WO2017175434A1 WO2017175434A1 PCT/JP2017/000441 JP2017000441W WO2017175434A1 WO 2017175434 A1 WO2017175434 A1 WO 2017175434A1 JP 2017000441 W JP2017000441 W JP 2017000441W WO 2017175434 A1 WO2017175434 A1 WO 2017175434A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- learning
- learning setting
- user
- setting
- information processing
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Definitions
- the present disclosure relates to an information processing apparatus, an information processing method, and an information providing method.
- the learning setting according to the information related to the past learning process in which the similarity with the information related to the learning process specified by the user is higher than the predetermined similarity is learned as a recommendation target for the user.
- An information processing apparatus includes a data acquisition unit that is acquired as a setting and a display control unit that controls display according to the learning setting of the recommendation target.
- the learning setting according to the information related to the past learning process in which the similarity with the information related to the learning process specified by the user is higher than the predetermined similarity is learned as a recommendation target for the user.
- an information processing method including obtaining as a setting and controlling a display according to the learning setting of the recommendation object by a processor.
- the learning setting according to the information related to the past learning process in which the similarity with the information related to the learning process specified by the user is higher than the predetermined similarity is learned as a recommendation target for the user.
- an information providing method including searching as a setting and controlling transmission of the learning setting of the recommendation target by a processor.
- a hardware configuration of the information processing apparatus will be described. A hardware configuration of the information providing apparatus will be described.
- a plurality of constituent elements having substantially the same functional configuration may be distinguished by adding different numerals after the same reference numerals. However, when it is not necessary to particularly distinguish each of a plurality of constituent elements having substantially the same functional configuration, only the same reference numerals are given.
- Deep Learning the selection of learning settings is important. However, in order to obtain an index for selecting a learning setting, it is generally necessary to execute a learning process that takes a long time. Therefore, in this specification, a technique that can shorten the time required for selecting the learning setting will be mainly described. In this specification, it is mainly assumed that Deep Learning is used as learning. However, the learning mode is not particularly limited to Deep Learning.
- FIG. 1 is a diagram illustrating a configuration example of an information processing system according to an embodiment of the present disclosure.
- the information processing system 1 according to the present embodiment includes an information processing device 10 and an information providing device 20.
- the information processing apparatus 10 and the information providing apparatus 20 can communicate via the communication network 931.
- the information providing apparatus 20 provides a cloud service for learning processing to the information processing apparatus 10
- this cloud service it is possible to execute learning processing and acquire learning results from a browser displayed by the information processing apparatus 10.
- the present embodiment is not limited to such an example.
- the cloud service provided to the information processing apparatus 10 by the information providing apparatus 20 may be performed by the information processing apparatus 10 by itself instead of the information providing apparatus 20.
- the learning process is performed by selecting a data set and a learning setting and executing a parameter optimization algorithm on the data set and the learning setting.
- the learning result includes learning settings and optimized parameters.
- the learning setting includes a network structure. Also, as shown below, learning settings can also include optimization algorithms, error functions, regularization, mini-batch numbers, input data preprocessing, and the like.
- hyper parameter means a parameter selected before the optimization algorithm is executed
- parameter means a parameter optimized by the optimization algorithm, and depends on each layer of the network. Retained.
- the network structure includes a graph structure showing a connection relation of each layer constituting the network, a type of each layer constituting the network, a shape of an output from each layer, a hyper parameter of each layer, and the like.
- the graph structure indicating the connection relationship of each layer is a graph structure in which a layer (Affine layer or the like) is an edge, and a numerical group (for example, a vector, a tensor, or the like) that is an input to the layer and an output from the layer is a node. Equivalent to.
- the maxout layer includes a 50-dimensional vector as an example of the shape of the output, and the number of values that take max in each dimension of the output as an example of the hyperparameter.
- Optimized algorithm includes optimization algorithm type and hyperparameters.
- An example of an optimization algorithm is adagrad.
- An example of a hyperparameter of the optimization algorithm is a learning coefficient.
- the error function includes the error function type and hyperparameter.
- An example of the type of error function is a square error.
- Regularization includes regularization types and hyperparameters.
- An example of the type of regularization is L1 regularization.
- Examples of regularization hyperparameters include regularization term coefficients.
- the number of mini-batches corresponds to the number of data used in a single mini-batch when mini-batch learning is performed by optimization.
- the input data preprocessing includes the type of input data preprocessing and hyper parameters. Examples of the types of input data preprocessing include normalization processing and pre-learning by Auto Encoder.
- a search history is a set of search history trees.
- the search history tree includes the same data set ID and various information corresponding thereto (for example, the learning setting, the performance of the learning setting obtained by the optimization algorithm, the execution time of the optimization algorithm, and the optimization algorithm executed immediately before. Set of learned learning settings, etc.).
- Prediction accuracy is calculated by using the evaluation data set to calculate the value of the error function after performing the training process using the parameter training data set when parameter training (for learning) and evaluation are provided. It may be an average value of error function values calculated and calculated by each data sample.
- the learning setting performance is not limited to the prediction accuracy.
- the performance of the learning setting may be the number of parameters included in the learning setting (the smaller the number of parameters, the higher the performance), or the calculation amount from input to output in the network structure (the smaller the calculation amount, the lower the calculation amount). High performance).
- the learning setting performance may be a memory size used from input to output in the network structure (the smaller the memory size, the higher the performance).
- the learning setting performance may be any combination of the prediction accuracy, the number of parameters, the calculation amount, and the memory size.
- the search history is accumulated in the cloud service.
- the search history may include a search history for learning settings. Further, the search history may include a search history by a learning process executed based on the user's own operation and a search history of a learning process executed based on another user's operation.
- the user specifies a data set corresponding to the problem to be solved.
- the learning setting recommendation is provided to the user based on the search history.
- the form of the information processing apparatus 10 is not particularly limited.
- the information processing apparatus 10 may be a game machine, a smartphone, a mobile phone, a tablet terminal, or a PC (Personal Computer). May be.
- the information providing apparatus 20 is assumed to be a computer such as a server.
- FIG. 2 is a block diagram illustrating a functional configuration example of the information processing apparatus 10 according to the present embodiment.
- the information processing apparatus 10 includes an operation unit 110, a control unit 120, a communication unit 130, a storage unit 140, and a display unit 150.
- these functional blocks included in the information processing apparatus 10 will be described.
- the operation unit 110 has a function of accepting user operations.
- the operation unit 110 may include an input device such as a mouse and a keyboard.
- the operation part 110 since it should just have the function to receive a user's operation, it may also include a touch panel.
- the method employed by the touch panel is not particularly limited, and may be a capacitance method, a resistance film method, an infrared method, or an ultrasonic method.
- the operation unit 110 may include a camera.
- the control unit 120 executes control of each unit of the information processing apparatus 10. As illustrated in FIG. 2, the control unit 120 includes an operation acquisition unit 121, a transmission control unit 122, a data acquisition unit 123, and a display control unit 124. Details of these functional blocks included in the control unit 120 will be described later.
- the control part 120 may be comprised with CPU (Central Processing Unit; Central processing unit) etc., for example.
- CPU Central Processing Unit
- the control unit 120 is configured by a processing device such as a CPU, the processing device may be configured by an electronic circuit.
- the communication unit 130 has a function of performing communication with the information providing apparatus 20.
- the communication unit 130 is configured by a communication interface.
- the communication unit 130 can communicate with the information providing apparatus 20 via the communication network 931 (FIG. 1).
- the storage unit 140 is a recording medium that stores a program executed by the control unit 120 and stores data necessary for executing the program.
- the storage unit 140 temporarily stores data for calculation by the control unit 120.
- the storage unit 140 may be a magnetic storage unit device, a semiconductor storage device, an optical storage device, or a magneto-optical storage device.
- the display unit 150 has a function of displaying various information.
- the display unit 150 may be a liquid crystal display, an organic EL (Electro-Luminescence) display, or an HMD (Head Mount Display).
- the display unit 150 may be a display of another form as long as it has a function of displaying various types of information.
- FIG. 3 is a block diagram illustrating a functional configuration example of the information providing apparatus 20 according to the present embodiment.
- the information providing apparatus 20 includes a control unit 220, a communication unit 230, and a storage unit 240.
- these functional blocks provided in the information providing apparatus 20 will be described.
- the control unit 220 executes control of each unit of the information providing apparatus 20. As illustrated in FIG. 3, the control unit 220 includes an acquisition unit 221, a learning processing unit 222, a search processing unit 223, and a transmission control unit 224. Details of these functional blocks included in the control unit 220 will be described later.
- the control unit 220 may be configured by, for example, a CPU (Central Processing Unit). When the control unit 220 is configured by a processing device such as a CPU, the processing device may be configured by an electronic circuit.
- the communication unit 230 has a function of performing communication with the information processing apparatus 10.
- the communication unit 230 is configured by a communication interface.
- the communication unit 230 can communicate with the information processing apparatus 10 via the communication network 931 (FIG. 1).
- the storage unit 240 is a recording medium that stores a program executed by the control unit 220 and stores data necessary for executing the program.
- the storage unit 240 temporarily stores data for calculation by the control unit 220.
- the storage unit 240 may be a magnetic storage unit device, a semiconductor storage device, an optical storage device, or a magneto-optical storage device.
- FIG. 4 is a diagram illustrating an example of a database stored by the storage unit 240 of the information providing apparatus 20.
- the storage unit 240 includes a data set database 260 and a learning setting search history database 270.
- the storage unit 240 also stores a user database 280.
- the data set database 260 stores information in which the data set ID 261 and the data set 262 are associated with each other.
- the learning setting search history database 270 stores information in which the data set ID 271, the learning setting 272, the accuracy 273, and the user ID 274 are associated with each other.
- the user database 280 stores information in which the group ID 281 and the user ID 282 are associated with each other.
- the user needs to register his / her user ID and his / her group ID in the user database 280. That is, the user database 280 stores information in which the user ID of the user is associated with the group ID to which the user belongs.
- the user ID may be information that can uniquely identify the user, and may be a user account or the like.
- information indicating the relationship with other users may be further registered.
- a dataset is required to execute the learning process. Therefore, the user designates a data set to be used for the learning process when executing the learning process.
- the data set designated by the user is uploaded to the information providing apparatus 20, the data set is associated with the data set ID 261 and stored as the data set 262.
- learning settings are required to execute the learning process. Therefore, the user specifies learning settings when executing the learning process.
- the learning setting designated by the user is uploaded to the information providing apparatus 20.
- a learning process is executed by the learning processing unit 222 based on the data set and learning settings specified by the user. Further, in the information providing apparatus 20, after the learning process is executed, the accuracy of the learning setting obtained by the learning process is calculated, and the data set ID, the learning setting, the accuracy, and the user ID obtained by the learning process are calculated. , Data set ID 271, learning setting 272, accuracy 273, and user ID 274. The information providing apparatus 20 recommends learning settings to the user using various information stored in the database.
- This embodiment may be applied to any scene.
- the present embodiment can be applied to a situation in which a user registers a data set for image classification in the information providing apparatus 20 and causes the information providing apparatus 20 to execute a learning process to solve an image classification problem.
- a part factory identifies whether a part is defective or not by image classification.
- the data set includes photographs of a plurality of parts and labels indicating whether or not each photograph is defective.
- the operation acquisition unit 121 acquires the operation.
- Information related to the learning process specified by the user is controlled to be transmitted to the information providing apparatus 20 by the transmission control unit 122.
- information related to the learning process is acquired by the acquisition unit 221, and related to the past learning process in which the similarity with the information related to the learning process is higher than a predetermined similarity by the search processing unit 223.
- the learning setting corresponding to the information to be searched is searched.
- the learning setting obtained by the search processing unit 223 is transmitted to the information processing apparatus 10 by the transmission control unit 224.
- the data acquisition unit 123 acquires the learning setting received by the communication unit 130 as the learning setting of the recommendation target for the user.
- the display control part 124 controls the display according to the learning setting of recommendation object. According to such a configuration, it is possible to shorten the time required for selecting the learning setting.
- FIG. 5 is a diagram illustrating an example of a recommendation target learning setting display screen.
- the learning setting display screen G ⁇ b> 10 displayed and controlled by the display control unit 124 displays the learning setting for the recommendation target (in the example illustrated in FIG. 5, the network structure, the optimization algorithm, the error function, the regularization, The number of mini-batches and input data pre-processing) and the accuracy of the learning setting for the recommendation target are included.
- the specific contents of the information and search history related to the learning process specified by the user are not limited.
- the information related to the learning process specified by the user includes a data set specified by the user, and the information related to the past learning process includes the data set used in the past learning process. May include.
- the information related to the learning process may include a learning setting specified by the user, and the information related to the past learning process may include a learning setting in which the learning process has been performed in the past. .
- the search processing unit 223 searches for a learning setting in which learning processing has been performed in the past using a data set whose similarity with a data set specified by the user is higher than a predetermined similarity. Obtained from the history, the transmission control unit 224 may control transmission of the learning setting acquired by the search processing unit 223 to the information processing apparatus 10. At this time, the data acquisition unit 123 may acquire the learning setting received from the information providing device 20 as the learning setting for the recommendation target.
- the calculation of the similarity between the data sets may be performed in any way.
- the similarity between the data sets may be calculated based on the similarity between the feature information of the data sets, may be calculated based on the similarity between the statistics of the data sets, or may be calculated by both. (For example, it may be calculated by the sum of both).
- the feature information and statistics of the dataset may be registered in the dataset database 260 along with the dataset.
- FIG. 6 is a diagram illustrating an example of a data set registration screen.
- the data set registration screen G30 that is displayed and controlled by the display control unit 124 includes a directory G31 in which the data set is stored, and feature information of the data set (in the example shown in FIG. 6, the data type G32). , Task G33, data set description G34) and registration button G35.
- the directory G31 in which the data set is stored is designated by the user, one of the items of the data type G32 is selected, one of the items of the task G33 is selected, and is freely described in the data set description G34. Assume that G35 is selected.
- the statistics of the data set (for example, the number of learning samples, the average value of image sizes included in the learning data, the label bias value, etc.) is calculated by the learning processing unit 222, and the statistics of the data set and the data set
- the feature information and the data set are registered in the data set database 260.
- the similarity between the feature information of the data sets may be calculated in any way (the similarity between the statistics of the data sets may be calculated in the same manner as the similarity between the feature information of the data sets).
- the similarity between the feature information of the data sets may be represented by discrete values. At this time, if the feature information of the data sets match, the similarity between the feature information of the data sets may be “1: similar”, and if the feature information of the data sets do not match, the data The similarity between sets of feature information may be “0: not similar”.
- the similarity between the feature information of the data sets may be represented by continuous values.
- the similarity between the feature information of the data sets may be expressed by exp ( ⁇ (Euclidean distance)) using the Euclidean distance between the feature information of the data sets.
- the similarity between the feature information of the data sets may be represented by the cosine similarity of the bag of words vector converted from the text.
- data other than the similarity between data sets may be additionally considered.
- the search processing unit 223 has an accuracy higher than a predetermined accuracy among learning settings related to a learning process using a data set whose similarity with a data set specified by the user is higher than a predetermined similarity.
- You may acquire the learning setting which has as a learning setting of recommendation object from search history.
- the data acquisition part 123 may acquire the learning setting received from the information provision apparatus 20 as a learning setting of recommendation object.
- the data acquisition unit 123 may acquire a plurality of learning settings received from the information providing apparatus 20 as a plurality of recommendation target learning settings.
- the display control unit 124 may control the display according to the learning setting of the plurality of recommendation objects according to at least one of the similarity and the accuracy.
- FIG. 7 is a diagram showing another example of the recommendation target learning setting display screen.
- the learning setting display screen G ⁇ b> 40 displayed and controlled by the display control unit 124 has a network structure as an example of the recommendation target learning setting, the accuracy of the learning setting of the recommendation target, and the similarity between the data sets. It is included.
- the learning settings may be arranged in a predetermined direction (for example, from top to bottom) in descending order of accuracy. Further, as shown in FIG. 7, the learning settings may be arranged in a predetermined direction (for example, from top to bottom) in descending order of the data set similarity.
- the display of the learning setting for the recommendation target may be selectable by the user.
- the display control unit 124 may control the display of the details of the learning setting for the recommendation target (such as the learning setting display screen G10 illustrated in FIG. 5).
- the display control unit 124 displays a learning setting search history tree including the learning setting of the recommendation target (such as the learning setting search history tree display screen G50 illustrated in FIG. 9). The display may be controlled.
- the information related to the learning process includes the learning setting specified by the user
- the information related to the past learning process includes the learning setting in which the learning process has been performed in the past. Including.
- the search processing unit 223 acquires from the search history a learning setting whose similarity with the learning setting specified by the user is higher than a predetermined similarity, and the transmission control unit 224 The transmission of the learning setting acquired by the unit 223 to the information processing apparatus 10 may be controlled.
- the data acquisition unit 123 may acquire the learning setting received from the information providing device 20 as the learning setting for the recommendation target.
- the calculation of the similarity between the learning settings may be performed in any manner.
- the similarity between learning settings may be calculated by the sum of the similarities between corresponding elements in the two learning settings.
- the similarity between elements may be calculated when the types of information included in the elements match (for example, the types of information included in the optimization algorithm are the type of learning algorithm and the value of the hyperparameter).
- the calculation of the similarity between elements can be performed in the same manner as the calculation of the similarity between data sets.
- the graph kernel technique described in the following reference can be used to calculate the similarity of the graph structure in the network structure.
- the search processing unit 223 is a learning setting in which the similarity to the learning setting specified by the user is higher than a predetermined similarity, and the learning setting having higher accuracy than the learning setting specified by the user. You may acquire as learning setting of recommendation object from search history. And a data acquisition part may acquire the learning setting received from the information provision apparatus 20 as a learning setting of recommendation object.
- the search processing unit 223 is a learning setting in which the similarity with the learning setting specified by the user is higher than a predetermined similarity, and the learning setting with the highest frequency of appearing in the learning setting search history database 270.
- You may acquire as a learning setting of recommendation object.
- a data acquisition part may acquire the learning setting received from the information provision apparatus 20 as a learning setting of recommendation object.
- the search processing unit 223 is a learning setting in which the similarity to the learning setting specified by the user is higher than a predetermined similarity, and according to the accuracy and the frequency of appearance in the learning setting search history database 270.
- the learning setting (for example, the learning setting that maximizes the product of the accuracy and the frequency of appearance in the learning setting search history database 270) may be acquired as the learning setting for the recommendation target.
- a data acquisition part may acquire the learning setting received from the information provision apparatus 20 as a learning setting of recommendation object.
- the display control unit 124 may control the display of the learning setting for the recommendation target.
- the display control unit 124 may control the display of the learning setting display screen G10 as illustrated in FIG.
- the display control part 124 may control the display of the difference of the learning setting of the recommendation object with respect to the learning setting designated by the user. Note that the display of the difference may be limited to a case where the similarity between learning settings is higher than a predetermined similarity.
- FIG. 8 is a diagram showing another example of the recommended setting learning setting display screen.
- the difference D1 indicates that “Tanh layer” included in the learning setting designated by the user is changed to “Relu layer”. It is displayed.
- the learning setting display screen G20 displays a difference D2 indicating that the learning coefficient included in the learning setting designated by the user is changed to “0.1”.
- the prediction accuracy is displayed for each difference on the learning setting display screen G20.
- the prediction accuracy is calculated by calculating an average value of the past accuracy increase in the recommended change and applying the average value to the accuracy of the learning setting specified by the user.
- the learning setting element to be recommended may be designated by the user. In this case, a change of only the designated element may be recommended.
- the display of the recommended change may be selectable by the user.
- the display control unit 124 may control the display of the details of the learning setting for the recommendation target (such as the learning setting display screen G10 illustrated in FIG. 5).
- the display control unit 124 displays a learning setting search history tree (such as the learning setting search history tree display screen G50 illustrated in FIG. 9) including the learning setting of the recommendation target. May be controlled.
- FIG. 9 is a diagram illustrating an example of a learning setting search history tree display screen. As illustrated in FIG. 9, the display control unit 124 can control the display of the learning setting search history tree display screen G50 as a learning setting search history tree including the learning setting of the recommendation target.
- the learning setting of the recommendation target may be acquired. That is, the search processing unit 223 obtains the most accurate learning setting from the past learning setting search history tree whose similarity with the learning setting search history tree executed based on the user's operation is higher than the predetermined similarity. To do. And the data acquisition part 123 may acquire the learning setting received from the information provision apparatus 20 as a learning setting of recommendation object.
- Similarity between learning setting search history trees may be calculated in any way.
- the similarity between learning setting search history trees may be calculated by the sum of the similarities between corresponding learning settings in two learning setting search history trees. The calculation of the similarity between learning settings is as described above.
- the similarity of the similarity calculation target is calculated by the sum of the similarities of the corresponding elements included in the similarity calculation target.
- the influence of each element may be considered. That is, a weight scalar value is assigned to each element, and the similarity of the similarity calculation target is calculated by the weighted sum of the similarities between corresponding elements included in the similarity calculation target. May be.
- the user can execute the learning process using the learning setting for the recommendation target.
- a part selected from the learning setting for the recommendation target may be used for executing the learning process, or a plurality of learning settings for the recommendation target may be used in combination.
- a parameter included in the recommendation target learning setting may be used as an initial value.
- a group can be created by a user, and the created user becomes a host of the group.
- the host can define users belonging to the group by inviting other users to the group.
- the learning setting search history tree is referred to, but some users may not want to disclose their learning setting search history tree to all other users. In that case, the user performs an operation for designating the disclosure range of the learning setting search history tree.
- the search processing unit 223 discloses the user's learning setting search history tree only to other users belonging to the same group as the user. Thereby, the access right to the learning setting search history tree can be controlled.
- the learning setting search history tree accessible only to a user who has registered as a friend.
- the user can make a setting such that a part of his / her learning setting search history tree is disclosed, a part is disclosed only to the group, and a part is not disclosed to other users.
- FIG. 10 is a block diagram illustrating a hardware configuration example of the information processing apparatus 10 according to the embodiment of the present disclosure.
- the information processing apparatus 10 includes a CPU (Central Processing unit) 801, a ROM (Read Only Memory) 803, and a RAM (Random Access Memory) 805.
- the information processing apparatus 10 may also include a host bus 807, a bridge 809, an external bus 811, an interface 813, an input device 815, an output device 817, a storage device 819, a drive 821, a connection port 823, and a communication device 825.
- the information processing apparatus 10 may include an imaging device 833 and a sensor 835 as necessary.
- the information processing apparatus 10 may have a processing circuit called DSP (Digital Signal Processor) or ASIC (Application Specific Integrated Circuit) instead of or together with the CPU 801.
- DSP Digital Signal Processor
- ASIC Application Specific Integrated Circuit
- the CPU 801 functions as an arithmetic processing device and a control device, and controls all or a part of the operation in the information processing device 10 according to various programs recorded in the ROM 803, the RAM 805, the storage device 819, or the removable recording medium 827.
- the ROM 803 stores programs used by the CPU 801, calculation parameters, and the like.
- the RAM 805 temporarily stores programs used in the execution of the CPU 801, parameters that change as appropriate during the execution, and the like.
- the CPU 801, the ROM 803, and the RAM 805 are connected to each other by a host bus 807 configured by an internal bus such as a CPU bus. Further, the host bus 807 is connected to an external bus 811 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 809.
- PCI Peripheral Component Interconnect / Interface
- the input device 815 is a device operated by the user, such as a mouse, a keyboard, a touch panel, a button, a switch, and a lever.
- the input device 815 may include a microphone that detects the user's voice.
- the input device 815 may be, for example, a remote control device using infrared rays or other radio waves, or may be an external connection device 829 such as a mobile phone corresponding to the operation of the information processing device 10.
- the input device 815 includes an input control circuit that generates an input signal based on information input by the user and outputs the input signal to the CPU 801. The user operates the input device 815 to input various data to the information processing device 10 or instruct a processing operation.
- An imaging device 833 which will be described later, can also function as an input device by imaging a user's hand movement, a user's finger, and the like. At this time, the pointing position may be determined according to the movement of the hand or the direction of the finger.
- the output device 817 is configured by a device that can notify the user of the acquired information visually or audibly.
- the output device 817 includes, for example, an LCD (Liquid Crystal Display), a PDP (Plasma Display Panel), an organic EL (Electro-Luminescence) display, a display device such as a projector, a hologram display device, a sound output device such as a speaker and headphones, As well as a printer device.
- the output device 817 outputs the result obtained by the processing of the information processing device 10 as a video such as text or an image, or outputs it as a sound or sound.
- the output device 817 may include a light or the like for brightening the surroundings.
- the storage device 819 is a data storage device configured as an example of a storage unit of the information processing device 10.
- the storage device 819 includes, for example, a magnetic storage device such as an HDD (Hard Disk Drive), a semiconductor storage device, an optical storage device, or a magneto-optical storage device.
- the storage device 819 stores programs executed by the CPU 801, various data, various data acquired from the outside, and the like.
- the drive 821 is a reader / writer for a removable recording medium 827 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and is built in or externally attached to the information processing apparatus 10.
- the drive 821 reads information recorded on the mounted removable recording medium 827 and outputs the information to the RAM 805. Further, the drive 821 writes a record in the attached removable recording medium 827.
- the connection port 823 is a port for directly connecting a device to the information processing apparatus 10.
- the connection port 823 can be, for example, a USB (Universal Serial Bus) port, an IEEE 1394 port, a SCSI (Small Computer System Interface) port, or the like.
- the connection port 823 may be an RS-232C port, an optical audio terminal, an HDMI (registered trademark) (High-Definition Multimedia Interface) port, or the like.
- the communication device 825 is a communication interface configured with a communication device for connecting to the communication network 931, for example.
- the communication device 825 can be, for example, a communication card for wired or wireless LAN (Local Area Network), Bluetooth (registered trademark), or WUSB (Wireless USB).
- the communication device 825 may be a router for optical communication, a router for ADSL (Asymmetric Digital Subscriber Line), or a modem for various communication.
- the communication device 825 transmits and receives signals and the like using a predetermined protocol such as TCP / IP with the Internet and other communication devices, for example.
- the communication network 931 connected to the communication device 825 is a wired or wireless network, such as the Internet, home LAN, infrared communication, radio wave communication, or satellite communication.
- the image pickup apparatus 833 uses various members such as an image pickup element such as a CCD (Charge Coupled Device) or a CMOS (Complementary Metal Oxide Semiconductor), and a lens for controlling the formation of a subject image on the image pickup element. It is an apparatus that images a real space and generates a captured image.
- the imaging device 833 may capture a still image, or may capture a moving image.
- the sensor 835 is, for example, various sensors such as an acceleration sensor, a gyro sensor, a geomagnetic sensor, an optical sensor, and a sound sensor.
- the sensor 835 acquires information about the state of the information processing apparatus 10 itself, such as the attitude of the information processing apparatus 10, and information about the surrounding environment of the information processing apparatus 10, such as brightness and noise around the information processing apparatus 10. To do.
- the sensor 835 may include a GPS sensor that receives a GPS (Global Positioning System) signal and measures the latitude, longitude, and altitude of the apparatus.
- GPS Global Positioning System
- FIG. 11 is a block diagram illustrating a hardware configuration example of the information providing apparatus 20 according to the embodiment of the present disclosure.
- the information providing apparatus 20 includes a CPU (Central Processing unit) 901, a ROM (Read Only Memory) 903, and a RAM (Random Access Memory) 905. Further, the information providing apparatus 20 may include a host bus 907, a bridge 909, an external bus 911, an interface 913, a storage apparatus 919, a drive 921, a connection port 923, and a communication apparatus 925.
- the information processing apparatus 10 may include a processing circuit called a DSP (Digital Signal Processor) or ASIC (Application Specific Integrated Circuit) instead of or in addition to the CPU 901.
- DSP Digital Signal Processor
- ASIC Application Specific Integrated Circuit
- the CPU 901 functions as an arithmetic processing device and a control device, and controls all or a part of the operation within the information providing device 20 according to various programs recorded in the ROM 903, the RAM 905, the storage device 919, or the removable recording medium 927.
- the ROM 903 stores programs and calculation parameters used by the CPU 901.
- the RAM 905 temporarily stores programs used in the execution of the CPU 901, parameters that change as appropriate during the execution, and the like.
- the CPU 901, the ROM 903, and the RAM 905 are connected to each other by a host bus 907 configured by an internal bus such as a CPU bus. Further, the host bus 907 is connected to an external bus 911 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 909.
- PCI Peripheral Component Interconnect / Interface
- the storage device 919 is a data storage device configured as an example of a storage unit of the information providing device 20.
- the storage device 919 includes, for example, a magnetic storage device such as an HDD (Hard Disk Drive), a semiconductor storage device, an optical storage device, or a magneto-optical storage device.
- the storage device 919 stores programs executed by the CPU 901, various data, various data acquired from the outside, and the like.
- the drive 921 is a reader / writer for a removable recording medium 927 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and is built in or externally attached to the information providing apparatus 20.
- the drive 921 reads information recorded on the attached removable recording medium 927 and outputs the information to the RAM 905.
- the drive 921 writes a record in the attached removable recording medium 927.
- the connection port 923 is a port for directly connecting a device to the information providing apparatus 20.
- the connection port 923 can be, for example, a USB (Universal Serial Bus) port, an IEEE 1394 port, a SCSI (Small Computer System Interface) port, or the like.
- the connection port 923 may be an RS-232C port, an optical audio terminal, an HDMI (registered trademark) (High-Definition Multimedia Interface) port, or the like.
- the communication device 925 is a communication interface configured with, for example, a communication device for connecting to the communication network 931.
- the communication device 925 can be, for example, a communication card for wired or wireless LAN (Local Area Network), Bluetooth (registered trademark), or WUSB (Wireless USB).
- the communication device 925 may be a router for optical communication, a router for ADSL (Asymmetric Digital Subscriber Line), or a modem for various communication.
- the communication device 925 transmits and receives signals and the like using a predetermined protocol such as TCP / IP with the Internet and other communication devices, for example.
- the communication network 931 connected to the communication device 925 is a wired or wireless network, such as the Internet, a home LAN, infrared communication, radio wave communication, or satellite communication.
- the learning setting according to the similarity between the information related to the learning process specified by the user and the search history in the past learning process is set as a recommendation target for the user.
- An information processing apparatus includes a data acquisition unit that is acquired as a learning setting, and a display control unit that controls display according to the learning setting of the recommendation target.
- a high-performance learning setting is automatically determined.
- the position of each component is not particularly limited.
- the example in which the learning processing unit 222, the data set database 260, and the learning setting search history database 270 are provided in the information providing apparatus 20 has been described above.
- some or all of the learning processing unit 222, the data set database 260, and the learning setting search history database 270 may be provided in the information processing apparatus 10.
- the information processing system 1 may not include the information providing apparatus 20.
- the following configurations also belong to the technical scope of the present disclosure.
- (1) Data acquisition for acquiring learning settings corresponding to information related to past learning processes whose similarity to information related to learning processes specified by the user is higher than a predetermined similarity as learning settings for a user to be recommended
- a display control unit that controls display according to the learning setting of the recommendation target;
- An information processing apparatus comprising: (2) The data acquisition unit acquires the performance of learning setting of the recommendation target, The display control unit controls display of the performance; The information processing apparatus according to (1).
- Information related to the learning process specified by the user includes a data set specified by the user;
- the information related to the past learning process includes a data set used for the past learning process.
- the data acquisition unit acquires, as the recommendation target learning setting, a learning setting in which learning processing has been performed in the past using a data set whose similarity with a data set specified by the user is higher than a predetermined similarity.
- the data acquisition unit is higher than a predetermined performance among learning settings in which learning processing has been performed in the past using a data set whose similarity with a data set specified by the user is higher than a predetermined similarity Acquiring a learning setting having performance as a learning setting of the recommendation target; The information processing apparatus according to (4).
- the display control unit controls display according to the learning setting of the plurality of recommendation targets according to at least one of the similarity and the performance when the learning setting of the recommendation target is acquired by the data acquisition unit. , The information processing apparatus according to (4) or (5). (7) The similarity is calculated based on the similarity of at least one of the feature information and statistics of the data sets, The information processing apparatus according to any one of (4) to (6). (8) The display control unit controls display of the similarity; The information processing apparatus according to any one of (4) to (7). (9) The information related to the learning process includes a learning setting specified by the user, The information related to the past learning process includes a learning setting in which the learning process has been performed in the past. The information processing apparatus according to (1) or (2).
- the data acquisition unit acquires a learning setting whose similarity with a learning setting specified by the user is higher than a predetermined similarity as the learning setting of the recommendation target;
- the data acquisition unit is a learning setting having a higher degree of similarity to a learning setting specified by the user than a predetermined similarity and having a higher performance than the learning setting specified by the user. Obtained as a learning setting for the recommendation object, The information processing apparatus according to (9) or (10).
- (12) The data acquisition unit is a learning setting whose similarity with a learning setting specified by the user is higher than a predetermined similarity, and which has the highest frequency of appearing in the learning setting search history. Get as learning settings for The information processing apparatus according to (9) or (10).
- the data acquisition unit is a learning setting whose similarity with the learning setting specified by the user is higher than a predetermined similarity, and the learning setting according to the performance and the frequency of appearance in the learning setting search history Obtained as a learning setting for the recommendation object,
- the information processing apparatus according to (9) or (10).
- the display control unit controls the display of the learning setting of the recommendation target or the display of the difference of the learning setting of the recommendation target with respect to the learning setting specified by the user;
- the information processing apparatus according to any one of (9) to (13).
- the display control unit is configured to display a learning setting search history tree including the learning setting of the recommendation object or learning of the recommendation object when display of the learning setting of the recommendation object or display of the difference is selected by a user.
- the display control unit controls display of a learning setting search history tree including learning settings of the recommendation target; The information processing apparatus according to any one of (1) to (15).
- the data acquisition unit acquires a learning setting with the highest performance from a past learning setting search history tree whose similarity with a learning setting search history tree executed based on the user's operation is higher than a predetermined similarity.
- the information processing apparatus includes: An operation acquisition unit that acquires an operation for specifying a disclosure range of a learning setting search history executed based on the user's operation; The information processing apparatus according to (1).
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Library & Information Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Algebra (AREA)
- Operations Research (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- User Interface Of Digital Computer (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Description
0.背景
1.本開示の実施形態
1.1.システム構成例
1.2.機能構成例
1.3.情報処理システムの機能詳細
1.4.ハードウェア構成例
2.むすび
ニューラルネットワークを用いた学習に関する技術として様々な技術が存在する(例えば、特開平5-135000号公報参照)。ニューラルネットワークは大きく三つの層(入力層、中間層および出力層)に分けられる。このうち、中間層を複数有するネットワークを用いた学習は、Deep Learningと呼ばれている。
[1.1.システム構成例]
まず、図面を参照しながら本開示の一実施形態に係る情報処理システムの構成例について説明する。図1は、本開示の一実施形態に係る情報処理システムの構成例を示す図である。図1に示したように、本実施形態に係る情報処理システム1は、情報処理装置10および情報提供装置20を備える。情報処理装置10および情報提供装置20は、通信ネットワーク931を介して通信を行うことが可能である。
続いて、本実施形態に係る情報処理装置10の機能構成例について説明する。図2は、本実施形態に係る情報処理装置10の機能構成例を示すブロック図である。図2に示したように、情報処理装置10は、操作部110、制御部120、通信部130、記憶部140および表示部150を備える。以下、情報処理装置10が備えるこれらの機能ブロックについて説明する。
続いて、情報処理システム1の機能詳細について説明する。図4は、情報提供装置20の記憶部240によって記憶されるデータベースの例を示す図である。図4に示すように、記憶部240は、データセットデータベース260と学習設定探索履歴データベース270とを含む。また、記憶部240は、ユーザデータベース280を記憶する。
次に、図10を参照して、本開示の実施形態に係る情報処理装置10のハードウェア構成について説明する。図10は、本開示の実施形態に係る情報処理装置10のハードウェア構成例を示すブロック図である。
以上説明したように、本開示の実施形態によれば、ユーザによって指定される学習処理に関連する情報と過去の学習処理における探索履歴との類似度に応じた学習設定をユーザへの推薦対象の学習設定として取得するデータ取得部と、前記推薦対象の学習設定に応じた表示を制御する表示制御部と、を備える、情報処理装置が提供される。
(1)
ユーザによって指定される学習処理に関連する情報との類似度が所定の類似度よりも高い過去の学習処理に関連する情報に応じた学習設定をユーザへの推薦対象の学習設定として取得するデータ取得部と、
前記推薦対象の学習設定に応じた表示を制御する表示制御部と、
を備える、情報処理装置。
(2)
前記データ取得部は、前記推薦対象の学習設定の性能を取得し、
前記表示制御部は、前記性能の表示を制御する、
前記(1)に記載の情報処理装置。
(3)
前記ユーザによって指定される学習処理に関連する情報は、前記ユーザによって指定されるデータセットを含み、
前記過去の学習処理に関連する情報は、過去の学習処理に用いられたデータセットを含む、
前記(1)または(2)に記載の情報処理装置。
(4)
前記データ取得部は、前記ユーザによって指定されるデータセットとの類似度が所定の類似度よりも高いデータセットを用いた過去に学習処理が実施された学習設定を前記推薦対象の学習設定として取得する、
前記(3)に記載の情報処理装置。
(5)
前記データ取得部は、前記ユーザによって指定されるデータセットとの類似度が所定の類似度よりも高いデータセットを用いた過去に学習処理が実施された学習設定のうち、所定の性能よりも高い性能を有する学習設定を前記推薦対象の学習設定として取得する、
前記(4)に記載の情報処理装置。
(6)
前記表示制御部は、前記データ取得部によって前記推薦対象の学習設定が複数取得された場合、前記類似度および性能の少なくともいずれか一方に従って当該複数の推薦対象の学習設定に応じた表示を制御する、
前記(4)または(5)に記載の情報処理装置。
(7)
前記類似度は、データセット同士の特徴情報および統計量の少なくともいずれか一方の類似度に基づいて算出される、
前記(4)~(6)のいずれか一項に記載の情報処理装置。
(8)
前記表示制御部は、前記類似度の表示を制御する、
前記(4)~(7)のいずれか一項に記載の情報処理装置。
(9)
前記学習処理に関連する情報は、前記ユーザによって指定される学習設定を含み、
前記過去の学習処理に関連する情報は、過去に学習処理が実施された学習設定を含む、
前記(1)または(2)に記載の情報処理装置。
(10)
前記データ取得部は、前記ユーザによって指定される学習設定との類似度が所定の類似度よりも高い学習設定を前記推薦対象の学習設定として取得する、
前記(9)に記載の情報処理装置。
(11)
前記データ取得部は、前記ユーザによって指定される学習設定との類似度が所定の類似度よりも高い学習設定であり、かつ、前記ユーザによって指定される学習設定よりも高い性能を有する学習設定を前記推薦対象の学習設定として取得する、
前記(9)または(10)に記載の情報処理装置。
(12)
前記データ取得部は、前記ユーザによって指定される学習設定との類似度が所定の類似度よりも高い学習設定であり、かつ、学習設定探索履歴に登場する頻度が最も高い学習設定を前記推薦対象の学習設定として取得する、
前記(9)または(10)に記載の情報処理装置。
(13)
前記データ取得部は、前記ユーザによって指定される学習設定との類似度が所定の類似度よりも高い学習設定であり、かつ、性能と学習設定探索履歴に登場する頻度とに応じた学習設定を前記推薦対象の学習設定として取得する、
前記(9)または(10)に記載の情報処理装置。
(14)
前記表示制御部は、前記推薦対象の学習設定の表示、または、前記ユーザによって指定される学習設定に対する前記推薦対象の学習設定の差分の表示を制御する、
前記(9)~(13)のいずれか一項に記載の情報処理装置。
(15)
前記表示制御部は、前記推薦対象の学習設定の表示、または、前記差分の表示がユーザによって選択された場合、前記推薦対象の学習設定を含む学習設定探索履歴ツリー、または、前記推薦対象の学習設定の詳細の表示を制御する、
前記(14)に記載の情報処理装置。
(16)
前記表示制御部は、前記推薦対象の学習設定を含む学習設定探索履歴ツリーの表示を制御する、
前記(1)~(15)のいずれか一項に記載の情報処理装置。
(17)
前記データ取得部は、前記ユーザの操作に基づいて実行された学習設定探索履歴ツリーとの類似度が所定の類似度よりも高い過去の学習設定探索履歴ツリーから最も性能の高い学習設定を取得する、
前記(1)に記載の情報処理装置。
(18)
前記情報処理装置は、
前記ユーザの操作に基づいて実行された学習設定探索履歴の公開範囲を指定するための操作を取得する操作取得部を備える、
前記(1)に記載の情報処理装置。
(19)
ユーザによって指定される学習処理に関連する情報との類似度が所定の類似度よりも高い過去の学習処理に関連する情報に応じた学習設定をユーザへの推薦対象の学習設定として取得することと、
プロセッサにより、前記推薦対象の学習設定に応じた表示を制御することと、
を含む、情報処理方法。
(20)
ユーザによって指定される学習処理に関連する情報との類似度が所定の類似度よりも高い過去の学習処理に関連する情報に応じた学習設定をユーザへの推薦対象の学習設定として検索することと、
プロセッサにより、前記推薦対象の学習設定の送信を制御することと、
を含む、情報提供方法。
10 情報処理装置
110 操作部
120 制御部
121 操作取得部
122 送信制御部
123 データ取得部
124 表示制御部
130 通信部
140 記憶部
150 表示部
20 情報提供装置
220 制御部
221 取得部
222 学習処理部
223 検索処理部
224 送信制御部
230 通信部
240 記憶部
260 データセットデータベース
262 データセット
270 学習設定探索履歴データベース
272 学習設定
273 精度
280 ユーザデータベース
Claims (20)
- ユーザによって指定される学習処理に関連する情報との類似度が所定の類似度よりも高い過去の学習処理に関連する情報に応じた学習設定をユーザへの推薦対象の学習設定として取得するデータ取得部と、
前記推薦対象の学習設定に応じた表示を制御する表示制御部と、
を備える、情報処理装置。 - 前記データ取得部は、前記推薦対象の学習設定の性能を取得し、
前記表示制御部は、前記性能の表示を制御する、
請求項1に記載の情報処理装置。 - 前記ユーザによって指定される学習処理に関連する情報は、前記ユーザによって指定されるデータセットを含み、
前記過去の学習処理に関連する情報は、過去の学習処理に用いられたデータセットを含む、
請求項1に記載の情報処理装置。 - 前記データ取得部は、前記ユーザによって指定されるデータセットとの類似度が所定の類似度よりも高いデータセットを用いて過去に学習処理が実施された学習設定を前記推薦対象の学習設定として取得する、
請求項3に記載の情報処理装置。 - 前記データ取得部は、前記ユーザによって指定されるデータセットとの類似度が所定の類似度よりも高いデータセットを用いて過去に学習処理が実施された学習設定のうち、所定の性能よりも高い性能を有する学習設定を前記推薦対象の学習設定として取得する、
請求項4に記載の情報処理装置。 - 前記表示制御部は、前記データ取得部によって前記推薦対象の学習設定が複数取得された場合、前記類似度および性能の少なくともいずれか一方に従って当該複数の推薦対象の学習設定に応じた表示を制御する、
請求項4に記載の情報処理装置。 - 前記類似度は、データセット同士の特徴情報および統計量の少なくともいずれか一方の類似度に基づいて算出される、
請求項4に記載の情報処理装置。 - 前記表示制御部は、前記類似度の表示を制御する、
請求項4に記載の情報処理装置。 - 前記学習処理に関連する情報は、前記ユーザによって指定される学習設定を含み、
前記過去の学習処理に関連する情報は、過去に学習処理が実施された学習設定を含む、
請求項1に記載の情報処理装置。 - 前記データ取得部は、前記ユーザによって指定される学習設定との類似度が所定の類似度よりも高い学習設定を前記推薦対象の学習設定として取得する、
請求項9に記載の情報処理装置。 - 前記データ取得部は、前記ユーザによって指定される学習設定との類似度が所定の類似度よりも高い学習設定であり、かつ、前記ユーザによって指定される学習設定よりも高い性能を有する学習設定を前記推薦対象の学習設定として取得する、
請求項9に記載の情報処理装置。 - 前記データ取得部は、前記ユーザによって指定される学習設定との類似度が所定の類似度よりも高い学習設定であり、かつ、学習設定探索履歴に登場する頻度が最も高い学習設定を前記推薦対象の学習設定として取得する、
請求項9に記載の情報処理装置。 - 前記データ取得部は、前記ユーザによって指定される学習設定との類似度が所定の類似度よりも高い学習設定であり、かつ、性能と学習設定探索履歴に登場する頻度とに応じた学習設定を前記推薦対象の学習設定として取得する、
請求項9に記載の情報処理装置。 - 前記表示制御部は、前記推薦対象の学習設定の表示、または、前記ユーザによって指定される学習設定に対する前記推薦対象の学習設定の差分の表示を制御する、
請求項9に記載の情報処理装置。 - 前記表示制御部は、前記推薦対象の学習設定の表示、または、前記差分の表示がユーザによって選択された場合、前記推薦対象の学習設定を含む学習設定探索履歴ツリー、または、前記推薦対象の学習設定の詳細の表示を制御する、
請求項14に記載の情報処理装置。 - 前記表示制御部は、前記推薦対象の学習設定を含む学習設定探索履歴ツリーの表示を制御する、
請求項1に記載の情報処理装置。 - 前記データ取得部は、前記ユーザの操作に基づいて実行された学習設定探索履歴ツリーとの類似度が所定の類似度よりも高い過去の学習設定探索履歴ツリーから最も性能の高い学習設定を取得する、
請求項1に記載の情報処理装置。 - 前記情報処理装置は、
前記ユーザの操作に基づいて実行された学習設定探索履歴の公開範囲を指定するための操作を取得する操作取得部を備える、
請求項1に記載の情報処理装置。 - ユーザによって指定される学習処理に関連する情報との類似度が所定の類似度よりも高い過去の学習処理に関連する情報に応じた学習設定をユーザへの推薦対象の学習設定として取得することと、
プロセッサにより、前記推薦対象の学習設定に応じた表示を制御することと、
を含む、情報処理方法。 - ユーザによって指定される学習処理に関連する情報との類似度が所定の類似度よりも高い過去の学習処理に関連する情報に応じた学習設定をユーザへの推薦対象の学習設定として検索することと、
プロセッサにより、前記推薦対象の学習設定の送信を制御することと、
を含む、情報提供方法。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP17778813.0A EP3441912A4 (en) | 2016-04-06 | 2017-01-10 | INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING METHOD, AND INFORMATION PROVIDING METHOD |
US16/076,396 US11593635B2 (en) | 2016-04-06 | 2017-01-10 | Information processing device, method of processing information, and method of providing information |
JP2018510231A JP6897673B2 (ja) | 2016-04-06 | 2017-01-10 | 情報処理装置、情報処理方法および情報提供方法 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2016076340 | 2016-04-06 | ||
JP2016-076340 | 2016-04-06 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2017175434A1 true WO2017175434A1 (ja) | 2017-10-12 |
Family
ID=60001003
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2017/000441 WO2017175434A1 (ja) | 2016-04-06 | 2017-01-10 | 情報処理装置、情報処理方法および情報提供方法 |
Country Status (4)
Country | Link |
---|---|
US (1) | US11593635B2 (ja) |
EP (1) | EP3441912A4 (ja) |
JP (1) | JP6897673B2 (ja) |
WO (1) | WO2017175434A1 (ja) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20200027089A (ko) * | 2018-08-30 | 2020-03-12 | 한양대학교 산학협력단 | 하이브리드 아이템 추천 방법 및 장치 |
WO2020157939A1 (ja) * | 2019-01-31 | 2020-08-06 | 富士通株式会社 | 情報処理装置、学習履歴管理システム及び学習履歴管理プログラム |
JP2021526063A (ja) * | 2018-09-04 | 2021-09-30 | 深▲セン▼先進技術研究院 | 不整脈検出方法、装置、電子装置およびコンピュータ記憶媒体 |
JP7260704B1 (ja) | 2022-08-29 | 2023-04-18 | 株式会社博報堂Dyホールディングス | 情報処理システム、コンピュータプログラム、及び情報処理方法 |
WO2023175947A1 (ja) * | 2022-03-18 | 2023-09-21 | 日本電気株式会社 | 情報処理装置、制御方法及び記憶媒体 |
JP7552996B2 (ja) | 2018-10-09 | 2024-09-18 | 株式会社Preferred Networks | ハイパーパラメータチューニング方法、プログラム、ユーザプログラム、装置、方法 |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021188354A1 (en) * | 2020-03-14 | 2021-09-23 | DataRobot, Inc. | Automated and adaptive design and training of neural networks |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06124196A (ja) * | 1992-10-12 | 1994-05-06 | Hitachi Ltd | モデル作成支援システム |
JP2001067223A (ja) * | 1999-08-25 | 2001-03-16 | Sumitomo Metal Ind Ltd | 問題解決方法、問題解決システム、及び記録媒体 |
Family Cites Families (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5598509A (en) * | 1992-08-28 | 1997-01-28 | Hitachi, Ltd. | Method of configuring a neural network and a diagnosis/recognition system using the same |
JP4487517B2 (ja) * | 2003-08-28 | 2010-06-23 | ソニー株式会社 | 情報提供装置及び情報提供方法、並びにコンピュータ・プログラム |
JP2005352900A (ja) * | 2004-06-11 | 2005-12-22 | Canon Inc | 情報処理装置、情報処理方法、パターン認識装置、及びパターン認識方法 |
WO2008047835A1 (fr) * | 2006-10-19 | 2008-04-24 | Nec Corporation | Système, procédé et programme d'étude active |
JP4577410B2 (ja) * | 2008-06-18 | 2010-11-10 | ソニー株式会社 | 画像処理装置、画像処理方法およびプログラム |
US8712930B1 (en) * | 2010-08-09 | 2014-04-29 | Google Inc. | Encoding digital content based on models for predicting similarity between exemplars |
US8645298B2 (en) * | 2010-10-26 | 2014-02-04 | Microsoft Corporation | Topic models |
US20120158623A1 (en) * | 2010-12-21 | 2012-06-21 | Microsoft Corporation | Visualizing machine learning accuracy |
JP6471934B2 (ja) * | 2014-06-12 | 2019-02-20 | パナソニックIpマネジメント株式会社 | 画像認識方法、カメラシステム |
JP6090286B2 (ja) * | 2014-10-31 | 2017-03-08 | カシオ計算機株式会社 | 機械学習装置、機械学習方法、分類装置、分類方法、プログラム |
US10248653B2 (en) * | 2014-11-25 | 2019-04-02 | Lionbridge Technologies, Inc. | Information technology platform for language translation and task management |
JP5965557B1 (ja) * | 2016-01-29 | 2016-08-10 | 株式会社リクルートホールディングス | 類似度学習システム及び類似度学習方法 |
-
2017
- 2017-01-10 EP EP17778813.0A patent/EP3441912A4/en not_active Withdrawn
- 2017-01-10 JP JP2018510231A patent/JP6897673B2/ja active Active
- 2017-01-10 US US16/076,396 patent/US11593635B2/en active Active
- 2017-01-10 WO PCT/JP2017/000441 patent/WO2017175434A1/ja active Application Filing
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06124196A (ja) * | 1992-10-12 | 1994-05-06 | Hitachi Ltd | モデル作成支援システム |
JP2001067223A (ja) * | 1999-08-25 | 2001-03-16 | Sumitomo Metal Ind Ltd | 問題解決方法、問題解決システム、及び記録媒体 |
Non-Patent Citations (4)
Title |
---|
FEURER, MATTHIAS ET AL.: "Initializing Bayesian Hyperparameter Optimization via Meta-Learning,", PROCEEDINGS OF THE TWENTY-NINTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, 16 February 2015 (2015-02-16), pages 1128 - 1135, XP055429085, Retrieved from the Internet <URL:http:www.aaai.org/ocs/index.php/AAAI/AAAI15/paper/view/10029/9349> [retrieved on 20170202] * |
See also references of EP3441912A4 * |
SMITH, MICHAEL R. ET AL., AN EASY TO USE REPOSITORY FOR COMPARING AND IMPROVING MACHINE LEARNING ALGORITHM USAGE, 5 June 2014 (2014-06-05), pages 1 - 7, XP055429070, Retrieved from the Internet <URL:https:arxiv.org/pdf/1405.7292v2.pdf> [retrieved on 20170202] * |
SMITH, MICHAEL R. ET AL.: "Recommending Learning Algorithms and Their Associated Hyperparameters", 7 July 2014 (2014-07-07), pages 1 - 2, XP055429064, Retrieved from the Internet <URL:https:arxiv.org/pdf/1407.1890v1.pdf> [retrieved on 20170202] * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20200027089A (ko) * | 2018-08-30 | 2020-03-12 | 한양대학교 산학협력단 | 하이브리드 아이템 추천 방법 및 장치 |
KR102195894B1 (ko) * | 2018-08-30 | 2020-12-29 | 한양대학교 산학협력단 | 하이브리드 아이템 추천 방법 및 장치 |
JP2021526063A (ja) * | 2018-09-04 | 2021-09-30 | 深▲セン▼先進技術研究院 | 不整脈検出方法、装置、電子装置およびコンピュータ記憶媒体 |
JP7304901B2 (ja) | 2018-09-04 | 2023-07-07 | 深▲セン▼先進技術研究院 | 不整脈検出方法、装置、電子装置およびコンピュータ記憶媒体 |
JP7552996B2 (ja) | 2018-10-09 | 2024-09-18 | 株式会社Preferred Networks | ハイパーパラメータチューニング方法、プログラム、ユーザプログラム、装置、方法 |
WO2020157939A1 (ja) * | 2019-01-31 | 2020-08-06 | 富士通株式会社 | 情報処理装置、学習履歴管理システム及び学習履歴管理プログラム |
JPWO2020157939A1 (ja) * | 2019-01-31 | 2021-10-14 | 富士通株式会社 | 情報処理装置、学習履歴管理システム及び学習履歴管理プログラム |
WO2023175947A1 (ja) * | 2022-03-18 | 2023-09-21 | 日本電気株式会社 | 情報処理装置、制御方法及び記憶媒体 |
JP7260704B1 (ja) | 2022-08-29 | 2023-04-18 | 株式会社博報堂Dyホールディングス | 情報処理システム、コンピュータプログラム、及び情報処理方法 |
JP2024032488A (ja) * | 2022-08-29 | 2024-03-12 | 株式会社博報堂Dyホールディングス | 情報処理システム、コンピュータプログラム、及び情報処理方法 |
Also Published As
Publication number | Publication date |
---|---|
EP3441912A1 (en) | 2019-02-13 |
US20190050730A1 (en) | 2019-02-14 |
JPWO2017175434A1 (ja) | 2019-02-14 |
EP3441912A4 (en) | 2019-02-13 |
US11593635B2 (en) | 2023-02-28 |
JP6897673B2 (ja) | 2021-07-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6897673B2 (ja) | 情報処理装置、情報処理方法および情報提供方法 | |
US11908176B2 (en) | Data recognition model construction apparatus and method for constructing data recognition model thereof, and data recognition apparatus and method for recognizing data thereof | |
CN102708120B (zh) | 生活流式传输 | |
JP7154678B2 (ja) | 目標の位置取得方法、装置、コンピュータ機器及びコンピュータプログラム | |
WO2019120032A1 (zh) | 模型构建方法、拍照方法、装置、存储介质及终端 | |
US11317018B2 (en) | Camera operable using natural language commands | |
US9269011B1 (en) | Graphical refinement for points of interest | |
US11887246B2 (en) | Generating ground truth datasets for virtual reality experiences | |
US11657085B1 (en) | Optical devices and apparatuses for capturing, structuring, and using interlinked multi-directional still pictures and/or multi-directional motion pictures | |
WO2017168922A1 (ja) | 情報処理装置、情報処理方法および情報提供方法 | |
JP2015532739A (ja) | ユーザインターフェースコントローラなどの有形オブジェクトの拡張 | |
KR20180049786A (ko) | 데이터 인식 모델 구축 장치 및 이의 데이터 인식 모델 구축 방법과, 데이터 인식 장치 및 이의 데이터 인식 방법 | |
WO2020259522A1 (zh) | 一种内容查找方法、相关设备及计算机可读存储介质 | |
WO2020221121A1 (zh) | 视频查询方法、装置、设备及存储介质 | |
CN111797850A (zh) | 视频分类方法、装置、存储介质及电子设备 | |
CN117413296A (zh) | 用于密集估计的具有相关表征的体积采样 | |
US20240311966A1 (en) | Apparatus and methods for augmenting vision with region-of-interest based processing | |
US20240153184A1 (en) | Real-time hand-held markerless human motion recording and avatar rendering in a mobile platform | |
CN111797867A (zh) | 系统资源优化方法、装置、存储介质及电子设备 | |
CN111797873A (zh) | 场景识别方法、装置、存储介质及电子设备 | |
KR102605355B1 (ko) | 이미지 기반 정보 제공 방법 및 장치 | |
WO2020207297A1 (zh) | 信息处理方法、存储介质及电子设备 | |
WO2017199505A1 (ja) | 情報処理装置、情報処理方法およびプログラム | |
US20150154210A1 (en) | Computer ecosystem with automatically curated content | |
CN118173088A (zh) | 智能设备的控制命令触发方法、装置、设备和存储介质 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 2018510231 Country of ref document: JP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2017778813 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 2017778813 Country of ref document: EP Effective date: 20181106 |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 17778813 Country of ref document: EP Kind code of ref document: A1 |