Information processing device, information processing server, information processing method, and computer-readable non-temporary storage medium

Abstract

To achieve both protection of privacy and high inference accuracy.SOLUTION: Provided is an information processing device comprising a learning unit and a communication unit. The learning unit clusters hierarchical data on the basis of a plurality of inference models distributed from an information processing server, and performs learning using the inference model corresponding to each cluster. The communication unit transmits intermediate results generated for each cluster in the learning by the learning unit to the information processing server. The hierarchical data includes information for identifying a main element, and a log collected or generated in association with the main element.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] This disclosure relates to an information processing device, an information processing server, an information processing method, and a computer-readable non-temporary storage medium. [Background technology]

[0002] In recent years, models have been developed that perform some form of inference based on collected data. Furthermore, techniques to improve the accuracy of such inferences have been proposed. For example, Patent Document 1 discloses a technique for clustering the data used for inference. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-154825 [Overview of the Initiative] [Problems that the invention aims to solve]

[0004] However, the technology disclosed in Patent Document 1 may not be sufficiently effective depending on the structure of the data used for inference. [Means for solving the problem]

[0005] According to one aspect of this disclosure, an information processing device is provided, comprising: a learning unit that clusters hierarchical data based on a plurality of inference models distributed from an information processing server and performs learning using the corresponding inference model for each cluster; and a communication unit that transmits intermediate results generated for each cluster in the learning by the learning unit to the information processing server, wherein the hierarchical data includes information identifying the main elements and logs collected or generated in association with the main elements.

[0006] Furthermore, in another aspect of this disclosure, an information processing method is provided in which a processor clusters hierarchical data based on a plurality of inference models distributed from an information processing server, performs learning using the corresponding inference model for each cluster, and transmits intermediate results generated for each cluster in the learning to the information processing server, wherein the hierarchical data includes information identifying a principal element and logs collected or generated in association with the principal element.

[0007] Furthermore, according to another aspect of this disclosure, the computer comprises a learning unit that clusters hierarchical data based on a plurality of inference models distributed from an information processing server and performs learning using the corresponding inference model for each cluster, and a communication unit that transmits intermediate results generated for each cluster in the learning by the learning unit to the information processing server, wherein the hierarchical data is provided on a non-temporary storage medium readable by a computer that stores a program for it to function as an information processing device, including information that identifies the main elements and logs collected or generated in association with the main elements.

[0008] Furthermore, according to another aspect of this disclosure, an information processing server is provided, comprising: a learning unit that generates a plurality of inference models corresponding to a plurality of clusters; and a communication unit that transmits information relating to the plurality of inference models generated by the learning unit to a plurality of information processing devices, wherein the communication unit receives from the plurality of information processing devices intermediate results generated by learning based on hierarchical data clustered based on the plurality of inference models and the inference models corresponding to each cluster, and the learning unit updates the plurality of inference models based on the plurality of intermediate results, and the hierarchical data includes information identifying the main elements and logs collected or generated in association with the main elements.

[0009] Furthermore, in another aspect of this disclosure, an information processing method is provided in which a processor generates a plurality of inference models corresponding to a plurality of clusters, transmits information relating to the generated plurality of inference models to a plurality of information processing devices, receives from the plurality of information processing devices intermediate results generated by learning based on hierarchical data clustered based on the plurality of inference models and the inference models corresponding to each cluster, and updates the plurality of inference models based on the plurality of intermediate results, wherein the hierarchical data includes information identifying a principal element and logs collected or generated in association with the principal element.

[0010] Furthermore, according to another aspect of this disclosure, the computer comprises a learning unit that generates a plurality of inference models corresponding to a plurality of clusters, and a communication unit that transmits information relating to the plurality of inference models generated by the learning unit to a plurality of information processing devices, wherein the communication unit receives from the plurality of information processing devices intermediate results generated by learning based on hierarchical data clustered based on the plurality of inference models and the inference models corresponding to each cluster, the learning unit updates the plurality of inference models based on the plurality of intermediate results, and the hierarchical data is provided on a non-temporary storage medium readable by a computer that stores a program for it to function as an information processing server, including information identifying a main element and logs collected or generated in association with the main element. [Brief explanation of the drawing]

[0011] [Figure 1] This is a block diagram showing an example configuration of System 1 according to one embodiment of the present disclosure. [Figure 2] This block diagram shows an example configuration of the information processing device 10 according to the same embodiment. [Figure 3] This block diagram shows an example configuration of the information processing server 20 according to the same embodiment. [Figure 4] This figure shows an example of a model corresponding to hierarchical data according to the same embodiment. [Figure 5]This is a schematic diagram illustrating the flow of processing performed by System 1 according to the same embodiment. [Figure 6] This flowchart shows the processing flow executed by System 1 according to the same embodiment. [Figure 7] This figure shows an example of hierarchical data clustering according to the same embodiment. [Figure 8] This diagram illustrates the operation of System 1 when the main element according to the same embodiment is an access point 40. [Figure 9] This diagram illustrates the operation of System 1 when the main element according to the same embodiment is the product category GC. [Figure 10] This diagram illustrates the operation of System 1 when the main element of the embodiment is a person and label y is an indicator related to the person's health status. [Figure 11] This figure illustrates the operation of System 1 when the main element according to the same embodiment is a person and the label y is the person's facial expression. [Figure 12] This is a block diagram showing an example of the hardware configuration of an information processing device 90 according to one embodiment of the present disclosure. [Modes for carrying out the invention]

[0012] Preferred embodiments of this disclosure will be described in detail below with reference to the attached drawings. In this specification and the drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant descriptions will be omitted.

[0013] The explanation will be given in the following order. 1. Embodiment 1.1. Overview 1.2. Example System Configuration 1.3. Example of Information Processing Device 10 Configuration 1.4. Example Configuration of Information Processing Server 20 1.5. Feature Details 1.6. Application Examples 2. Hardware Configuration Example 3. Summary

[0014] <1. Embodiments> <<1.1. Overview>> First, an overview of one embodiment of this disclosure will be described.

[0015] As mentioned above, in recent years, there has been development of models (inference models) that perform some kind of inference based on collected data.

[0016] Inference models can accurately perform various inferences based on unknown data. Therefore, the generation and utilization of inference models are actively pursued in various fields.

[0017] However, depending on the structure of the data used to generate the inference model and the inference performed by the model, it may be difficult to generate a highly accurate inference model.

[0018] Examples of the above-mentioned data include hierarchical data.

[0019] Hierarchical data may be defined, for example, as data that includes information identifying a primary element and logs collected or generated in association with that primary element.

[0020] As an example, consider a case where device 80 communicates with multiple other devices and performs some kind of inference based on logs collected from each communication partner.

[0021] In this case, the main element may be a device that communicates with device 80.

[0022] In this case, the generated logs may have characteristics that depend on the communication partner, the pair of device 80 and the communication partner, etc.

[0023] Therefore, if learning using logs is performed without distinguishing between communication partners, it becomes difficult to generate an inference model with high accuracy, and there is also a possibility of increased convergence time and overfitting.

[0024] On the other hand, there is also a method, for example, disclosed in Patent Document 1, which involves clustering the data and then performing learning on each data point belonging to each cluster.

[0025] Thus, the method of classifying data with different characteristic tendencies into multiple clusters and performing learning on data belonging to each cluster appears to be useful for hierarchical data as well.

[0026] However, with methods like the one disclosed in Patent Document 1, it is difficult to generate highly accurate inference models because the clustering results cannot be modified based on the inference results.

[0027] Furthermore, if, for example, a server collects data from multiple devices and performs learning based on that data, protecting privacy becomes an issue.

[0028] On the other hand, when performing learning using data collected from multiple devices, there is a method called Federated Learning that protects data privacy.

[0029] However, general associative learning aims to generate a single inference model from data received from multiple devices.

[0030] Therefore, when using data with various characteristics, such as hierarchical data, in general associative learning, there is a possibility of increased convergence time, overfitting, and other problems.

[0031] To mitigate the possibilities described above, one possible approach is to cluster the data according to the characteristics of the device from which it was collected, and then perform learning on each data point belonging to that cluster.

[0032] However, in this case, all data collected from a particular device would be classified into the same cluster.

[0033] Therefore, when performing inference using logs generated for each of the multiple communication partners mentioned above, it can be difficult to reduce the possibility of increased convergence time, overfitting, and other issues.

[0034] The technical concept of one embodiment of this disclosure was conceived with the above-mentioned points in mind, and aims to achieve both privacy protection and high inference accuracy.

[0035] To this end, an information processing device 10 according to one embodiment of the present disclosure includes a learning unit 120 that clusters hierarchical data based on a plurality of inference models distributed from an information processing server 20 and performs learning using the corresponding inference model for each cluster.

[0036] Furthermore, an information processing device 10 according to one embodiment of this disclosure includes a communication unit 130 that transmits intermediate results generated for each cluster in the learning process by the learning unit 120 to an information processing server 20.

[0037] In other words, an information processing device 10 according to one embodiment of the present disclosure clusters its own hierarchical data using multiple inference models distributed from an information processing server 20, and transmits the intermediate results generated by learning for each cluster to the information processing server 20.

[0038] Here, the above intermediate results may include information that contains values ​​calculated from features and labels included in the hierarchical data, and may also be information that makes it difficult to reconstruct the features and labels.

[0039] On the other hand, an information processing server 20 according to one embodiment of the present disclosure includes a learning unit 210 that generates multiple inference models corresponding to multiple clusters, and a communication unit that transmits information relating to the multiple inference models generated by the learning unit 210 to multiple information processing devices 10.

[0040] One of the features of the communication unit 220 is that it receives hierarchical data clustered based on multiple inference models and intermediate results generated by learning based on the above inference models corresponding to each cluster from multiple information processing devices 10.

[0041] Furthermore, one of the features of the learning unit 210 is that it updates multiple inference models based on the multiple intermediate results mentioned above.

[0042] In other words, an information processing server 20 according to one embodiment of the present disclosure updates the inference model corresponding to each cluster based on the intermediate results for each cluster received from a plurality of information processing devices 10, and distributes the information related to the update to each of the information processing devices 10.

[0043] Thus, in the system 1 according to one embodiment of this disclosure, the information processing device 10 may repeatedly perform clustering and transmission of intermediate results, and the server may repeatedly update the inference model and distribute information related to said update.

[0044] The processing described above ensures the convergence of the inference model and simultaneously improves clustering accuracy and inference precision.

[0045] Furthermore, the processing described above makes it possible to enhance privacy protection by using intermediate results that are difficult to restore to the original data.

[0046] The following provides a detailed explanation of a system configuration example that achieves the above.

[0047] <<1.2. Example System Configuration>> Figure 1 is a block diagram showing an example configuration of System 1 according to one embodiment of the present disclosure.

[0048] As shown in Figure 1, the system 1 according to this embodiment comprises a plurality of information processing devices 10 and an information processing server 20.

[0049] Each information processing device 10 and information processing server 20 are connected to each other via a network 30, enabling them to communicate with one another.

[0050] Although Figure 1 illustrates a case where System 1 includes information processing devices 10a and 10b, the number of information processing devices 10 in this embodiment is not limited to this example.

[0051] (Information processing device 10) The information processing device 10 according to this embodiment performs hierarchical data clustering using an inference model distributed from the information processing server 20.

[0052] Furthermore, the information processing device 10 according to this embodiment performs learning using data belonging to each cluster and an inference model corresponding to the cluster, and transmits the intermediate results to the information processing server 20.

[0053] The information processing device 10 according to this embodiment may be, for example, a personal computer, a smartphone, a tablet, a game console, a wearable device, or the like.

[0054] (Information processing server 20) In this embodiment, the information processing server 20 generates an inference model corresponding to the configured cluster and distributes it to multiple information processing devices 10.

[0055] Furthermore, the information processing server 20 according to this embodiment receives intermediate results from multiple information processing devices 10 corresponding to each cluster, and updates each inference model based on these intermediate results.

[0056] The information processing server 20 according to this embodiment distributes information related to the updating of each inference model to a plurality of information processing devices 10.

[0057] (Network 30) The network 30 in this embodiment mediates communication between the information processing device 10 and the information processing server 20.

[0058] <<1.3. Example Configuration of Information Processing Device 10>> Next, a configuration example of the information processing device 10 according to this embodiment will be described in detail.

[0059] Figure 2 is a block diagram showing an example configuration of the information processing device 10 according to this embodiment.

[0060] As shown in Figure 2, the information processing device 10 according to this embodiment may include a sensor unit 110, a learning unit 120, and a communication unit 130.

[0061] (Sensor unit 110) The sensor unit 110 according to this embodiment collects various types of sensor information.

[0062] The sensor information collected by the sensor unit 110 may be used as an element (feature) of hierarchical data.

[0063] For this purpose, the sensor unit 110 may be equipped with various sensors for collecting sensor information that can be used as an element of hierarchical data.

[0064] On the other hand, if the hierarchical data does not include sensor information, the information processing device 10 does not need to include the sensor unit 110.

[0065] (Learning Section 120) In this embodiment, the learning unit 120 clusters hierarchical data based on multiple inference models distributed from the information processing server 20.

[0066] Furthermore, the learning unit 120 according to this embodiment performs learning using an inference model corresponding to each cluster.

[0067] As described above, the hierarchical data according to this embodiment may include information that identifies the main elements, and logs that are collected or generated in association with the main elements.

[0068] Furthermore, the learning unit 120 may cluster hierarchical data relating to different principal elements based on multiple inference models.

[0069] The functions of the learning unit 120 according to this embodiment are realized by various processors.

[0070] Details of the functions of the learning unit 120 according to this embodiment will be described separately.

[0071] (Communications Section 130) The communication unit 130 according to this embodiment communicates with the information processing server 20 via the network 30.

[0072] The communication unit 130 receives, for example, an inference model, information related to updating the inference model, etc., from the information processing server 20.

[0073] Furthermore, the communication unit 130 transmits the intermediate results generated by the learning unit 120 to the information processing server 20.

[0074] In addition, the communication unit 130 according to this embodiment may communicate with other devices other than the information processing server 20.

[0075] In this case, the communication unit 130 may also generate and store logs related to communication with other devices.

[0076] Logs related to communication with other devices may be used as part of hierarchical data.

[0077] The above describes an example configuration of the information processing device 10 according to this embodiment. Note that the above configuration, as illustrated in Figure 2, is merely an example, and the configuration of the information processing device 10 according to this embodiment is not limited to this example.

[0078] The information processing device 10 according to this embodiment may further include, for example, an input unit for receiving information input from a user, a display unit for displaying various types of information, and so on.

[0079] Furthermore, if the information processing device 10 is equipped with an input unit, the input information may be used as part of hierarchical data.

[0080] The configuration of the information processing device 10 according to this embodiment can be flexibly modified according to specifications and operation.

[0081] <<1.4. Example Configuration of Information Processing Server 20>> Next, a configuration example of the information processing server 20 according to this embodiment will be described in detail.

[0082] Figure 3 is a block diagram showing an example configuration of the information processing server 20 according to this embodiment.

[0083] As shown in Figure 3, the information processing server 20 according to this embodiment may include a learning unit 210 and a communication unit 220.

[0084] (Learning Section 210) The learning unit 210 according to this embodiment generates multiple inference models, each corresponding to a multiple cluster.

[0085] Furthermore, the learning unit 210 according to this embodiment updates the plurality of inference models based on the plurality of intermediate results received by the communication unit 220.

[0086] The functions of the learning unit 210 according to this embodiment are realized by various processors.

[0087] Details of the functions of the learning unit 210 according to this embodiment will be described separately.

[0088] (Communications Section 220) The communication unit 220 according to this embodiment communicates with a plurality of information processing devices 10 via the network 30.

[0089] The communication unit 220 transmits, for example, the inference model generated by the learning unit 210 and information related to updates to the inference model to multiple information processing devices 10.

[0090] Furthermore, the communication unit 220 receives intermediate results from multiple information processing devices 10.

[0091] The above describes an example configuration of the information processing server 20 according to this embodiment. Note that the above configuration, as illustrated in Figure 3, is merely an example, and the configuration of the information processing server 20 according to this embodiment is not limited to this example.

[0092] The information processing server 20 according to this embodiment may further include, for example, an input unit for receiving information input from a user, a display unit for displaying various types of information, and so on.

[0093] The configuration of the information processing server 20 according to this embodiment can be flexibly modified according to specifications and operation.

[0094] <<1.5. Feature Details>> Next, the functions of the information processing device 10 and the information processing server 20 according to this embodiment will be described in detail.

[0095] As described above, the information processing device 10 according to this embodiment clusters its own hierarchical data using multiple inference models distributed from the information processing server 20, and transmits the intermediate results generated by learning for each cluster to the information processing server 20.

[0096] Furthermore, the information processing server 20 according to this embodiment updates the inference model corresponding to each cluster based on the intermediate results for each cluster received from the multiple information processing devices 10, and distributes the information related to the update to each of the information processing devices 10.

[0097] In order to achieve the above processing, it is necessary for the information processing device 10 and the information processing server 20 to share a model that corresponds to hierarchical data.

[0098] Figure 4 shows an example of a model corresponding to hierarchical data according to this embodiment.

[0099] The left side of Figure 4 shows an example of a graphical model corresponding to hierarchical data, and the right side of Figure 4 shows an example of a generative model corresponding to hierarchical data.

[0100] In each model shown in Figure 4, η, θ, and κ correspond to the set (global) of information processing devices 10, the information processing device 10, and the principal element, respectively.

[0101] However, the models shown in Figure 4 are merely examples, and the graphical and generative models according to this embodiment should be designed appropriately according to the characteristics of the hierarchical data, the characteristics of the inferred labels (dependent variables), etc.

[0102] Next, with reference to Figures 5 and 6, the processing flow executed by System 1 according to this embodiment will be described in detail.

[0103] Figure 5 is a schematic diagram illustrating the flow of processing performed by System 1 according to this embodiment.

[0104] Figure 6 is a flowchart showing the processing flow performed by System 1 according to this embodiment.

[0105] Although Figure 5 illustrates the processing performed by information processing devices 10a and 10b, as described above, the system 1 according to this embodiment may include three or more information processing devices 10.

[0106] Furthermore, Figure 5 illustrates a case where the information processing server 20 generates three inference models M1 to M3, each corresponding to one of the three clusters C1 to C3. However, the number of clusters and inference models according to this embodiment is not limited to this example.

[0107] The number of clusters and inference models according to this embodiment may be appropriately designed according to the characteristics of the hierarchical data, the characteristics of the labels (target variables) to be inferred, etc.

[0108] First, as shown in Figure 6, the information processing server 20 initializes the inference model (S100).

[0109] In the example shown in Figure 5, the information processing server 20 initializes the inference models M1 to M3.

[0110] Next, the information processing server 20 distributes information related to the inference model (S101).

[0111] In the example shown in Figure 5, the information processing server 20 may transmit all the information that constitutes the inference models M1 to M3.

[0112] Next, each of the information processing devices 10 clusters the hierarchical data using an inference model (S102).

[0113] In the example shown in Figure 5, the information processing device 10a uses the distributed inference models M1 to M3 to classify the hierarchical data D1 into one of the clusters C1 to C3, which correspond to each of the inference models M1 to M3.

[0114] Similarly, the information processing device 10b uses the distributed inference models M1 to M3 to classify the hierarchical data D2 into one of the clusters C1 to C3, which correspond to the inference models M1 to M3, respectively.

[0115] Next, each information processing unit performs learning for each cluster (S103).

[0116] In the example shown in Figure 5, the information processing device 10a performs learning for each cluster C1 to C3, and the intermediate results w corresponding to each cluster C1 to C3 are obtained. 11 ~w 31 Generates.

[0117] Similarly, the information processing device 10b performs learning for each cluster C1 to C3 and obtains intermediate results w corresponding to each cluster C1 to C3. 12 ~w 32 Generates.

[0118] Next, each of the information processing apparatuses 10 transmits the intermediate results to the information processing server 20 (S104).

[0119] In the case of the example shown in FIG. 5, the information processing apparatus 10a transmits the intermediate results w 11 ~w 31 to the information processing server.

[0120] Similarly, the information processing apparatus 10b transmits the intermediate results w 12 ~w 32 to the information processing server 20.

[0121] Next, the information processing server 20 aggregates the intermediate results received for each cluster and updates the inference model corresponding to each cluster (S105).

[0122] In the case of the example shown in FIG. 5, the information processing server 20 calculates w1 from the intermediate results w 11 and w 12 and updates the inference model M1.

[0123] Similarly, the information processing server 20 calculates w2 from the intermediate results w 21 and w 22 and updates the inference model M2.

[0124] Similarly, the information processing server 20 calculates w3 from the intermediate results w 31 and w 32 and updates the inference model M3.

[0125] Subsequently, the information processing server 20 determines whether each inference model has converged (S106).

[0126] If the information processing server 20 determines that each inference model has converged (S106: Yes), the system 1 ends the series of processes.

[0127] On the other hand, if the information processing server 20 determines that each inference model has not converged (S106: Yes), the information processing server 20 returns to step S101 and distributes the information related to the inference model.

[0128] In the example shown in Figure 5, the information processing server 20 may, for example, send w1, w2, and w3 to the information processing devices 10a and 10b.

[0129] If the information processing server 20 returns to step S101, the information processing device 10 and the information processing server 20 repeatedly execute the subsequent processes.

[0130] The processing flow of System 1 according to this embodiment has been described in detail with an example.

[0131] Next, the information transmitted and received between the information processing device 10 and the information processing server 20 according to this embodiment will be described in more detail.

[0132] Figure 7 shows an example of hierarchical data clustering according to this embodiment.

[0133] In the example shown in Figure 7, cluster C1 is hierarchically classified, having principal elements ID: ME1, ME2, and ME3.

[0134] Furthermore, cluster C2 is classified hierarchically, with principal elements ID:ME4 and ID:ME5.

[0135] Furthermore, cluster C2 is classified hierarchically, with principal elements ID: ME6, ME7, and ME8.

[0136] Here, the main element ID is an example of information used to identify a main element.

[0137] Furthermore, hierarchical data includes not only the main element ID but also logs collected or generated in association with the main element.

[0138] The above log may include, for example, features and labels (target variable).

[0139] In the example shown in Figure 7, the feature quantity is x n1 ~x n5 It has five elements, but this is merely an example, and the number of elements in the feature quantity according to this embodiment is not limited to this example.

[0140] The intermediate results according to this embodiment may be values ​​calculated from the features and labels included in the hierarchical data.

[0141] When the information processing device 10 performs hierarchical data clustering as shown in Figure 7, intermediate results w corresponding to clusters C1 to C3 are generated. 11 , w 21 , w 31 You can also set it as follows.

[0142] w 11 ={A1,b1} w 21 ={A2,b2} w 31 ={A3,b3}

[0143] Here, A in the above k and b k These may be values ​​calculated based on the features and labels belonging to cluster Ck, respectively.

[0144] Below, A k and b k An example of the calculation is shown below.

[0145] A1=A(x 11 ,x 12 ,x 13 ,x 14 ,x 15 y1, x 21 ,x 22 ,x 23 ,x 24 ,x 25 ,y2, x 31 ,x 32 ,x 33 ,x 34 ,x35 ,y3) b1=b(x 11 ,x 12 ,x 13 ,x 14 ,x 15 ,y1, x 21 ,x 22 ,x 23 ,x 24 ,x 25 ,y2, x 31 ,x 32 ,x 33 ,x 34 ,x 35 ,y3)

[0146] A2=A(x 41 ,x 42 ,x 43 ,x 44 ,x 45 ,y4, x 51 ,x 52 ,x 53 ,x 54 ,x 55 ,y5) b2=b(x 41 ,x 42 ,x 43 ,x 44 ,x 45 ,y4, x 51 ,x 52 ,x 53 ,x 54 ,x 55 ,y5)

[0147] A3=A(x 61 ,x 62 ,x 63 ,x 64 ,x 65 ,y6, x 71 ,x 72 ,x 73 ,x 74 ,x 75 ,y7, x 81 ,x 82 ,x 83 ,x 84 ,x 85 ,y8) b3 = b(x 61 , x 62 , x 63 , x 64 , x 65 , y6, x 71 , x 72 , x 73 , x 74 , x 75 , y7, x 81 , x 82 , x 83 , x 84 , x 85 , y8)

[0148] As described above, the information processing apparatus 10 according to the present embodiment may calculate an intermediate result without using information for specifying a main element.

[0149] Also, according to the above calculation, the larger the number of hierarchical data belonging to the cluster Ck, the more A k and b k it is difficult to restore the original feature amount x ij and the label y i from them.

[0150] From these facts, according to the generation of the intermediate result according to the present embodiment, it is possible to effectively enhance the privacy protection performance.

[0151] Subsequently, the information related to the update of the inference model transmitted from the information processing server 20 to the information processing apparatus 10 will be described in more detail.

[0152] Here, it is assumed that inference models M1 to M3 corresponding to clusters C1 to C3 are generated respectively.

[0153] In this case, the information processing server 20 may calculate the information w1 to w3 related to the update of each of the inference models M1 to M3, for example, as follows.

[0154] w1 = (w 11 , w 12 , w 13,…,w 1n ) w2=(w 21 ,w 22 ,w 23 ,…,w 2n ) w3=(w 31 ,w 32 ,w 33 ,…,w 3n )

[0155] According to the calculations above, the more information processing devices n that calculate intermediate results, the greater the w k From the original feature x ij and label y i This makes it difficult to recover the data, effectively enhancing privacy protection.

[0156] On the other hand, the information processing device 10 can perform more accurate inferences by receiving information w1 to w3 related to the updates of each of the inference models M1 to M3.

[0157] For example, when inferring the label y9 of feature x9 belonging to cluster C3, the information processing device 10 only needs to calculate f(w3,x9).

[0158] Thus, the learning unit 120 of the information processing device 10 according to this embodiment is capable of inferring labels based on feature quantities and inference models.

[0159] <<1.6. Application Examples>> Next, we will explain the application of System 1 according to this embodiment with specific examples.

[0160] For example, the main elements according to this embodiment may be various devices that communicate with the information processing device 10.

[0161] In this case, the information processing device 10 may use various logs collected in connection with communication with the device as hierarchical data.

[0162] In the following section, we will describe an example in which the main element of this embodiment is an access point that communicates with the information processing device 10.

[0163] Figure 8 is a diagram illustrating the operation of System 1 when the main element according to this embodiment is the access point 40.

[0164] Figure 8 illustrates the operation when the information processing device 10 is a smartphone.

[0165] In this example, the information processing device 10 stores the communication log L40 for each access point 40.

[0166] Communication log L40 is associated with information that identifies access point 40 and is used as hierarchical data.

[0167] In this case, each communication log L40 contains feature quantities x1 to x n This includes the label y.

[0168] Features x1~x n For example, the received signal strength, CCA (Clear Channel Assessment) busy time, etc., may be used.

[0169] Furthermore, the label y may be an index representing the communication quality related to access point 40.

[0170] Each of the information processing devices 10 uses multiple inference models distributed from the information processing server 20 to generate the above-mentioned feature quantities x1~x n Cluster the communication log L40, which includes label y.

[0171] For example, in the case shown in Figure 8, the information processing device 10a classifies the three records of communication log L40a corresponding to access point 40a and the two records of communication log L40b corresponding to access point 40b into cluster C1.

[0172] Furthermore, the information processing device 10a classifies the two records of communication log L40c corresponding to the access point 40c into cluster C2.

[0173] Similarly, the information processing device 10b classifies the two records of communication log L40a corresponding to access point 40a into cluster C1, and the two records of communication log L40c corresponding to access point 40c into cluster C2.

[0174] Thus, the information processing device 10 according to this embodiment may perform clustering so that hierarchical data relating to the same main element is classified into the same cluster.

[0175] According to this, features can be learned according to a predetermined type of access point 40, a predetermined type of information processing device 10 and a predetermined type of access point 40 pair, and more accurate inference can be achieved.

[0176] Furthermore, the information processing device 10 may simultaneously perform clustering so that hierarchical data relating to different main elements are classified into the same cluster.

[0177] This allows for a reduction in the number of clusters, thereby improving learning efficiency.

[0178] Each of the information processing devices 10 according to this embodiment transmits the intermediate results calculated as described above to the information processing server 20 after performing the clustering described above, and receives information related to updating the inference model.

[0179] Subsequently, each of the information processing devices 10 according to this embodiment infers a label y using the collected feature quantity x and the inference model.

[0180] For example, the information processing device 10 can use an inference model to infer the communication quality when using a certain access point 40 from the received radio wave strength from that access point 40.

[0181] In this case, the information processing device 10 may also perform control such as connecting to an access point 40 whose inferred communication quality satisfies predetermined conditions, or connecting to the access point 40 with the highest inferred communication quality among multiple access points 40.

[0182] The above describes the operation when the main element of this embodiment is a device that communicates with the information processing device 10.

[0183] Next, we will describe an example where the main element of this embodiment is a product category.

[0184] The main element in this embodiment does not necessarily have to be a device.

[0185] Figure 9 is a diagram illustrating the operation of System 1 when the main element of this embodiment is the product category GC.

[0186] Figure 9 illustrates the operation when the information processing device 10 is a game device and the product is a game.

[0187] In this example, the information processing device 10 stores the purchase log Lgc together for each game category GC.

[0188] The purchase log Lgc is associated with information that identifies the game category GC and is used as hierarchical data.

[0189] In this case, each purchase log Lgc has features x1 to x n This includes the label y.

[0190] Features x1~x n For example, game manufacturers and sales rankings may be used.

[0191] Furthermore, label y may be an indicator related to the purchase of games belonging to category GC (for example, whether or not it was purchased, or whether or not it was pre-ordered).

[0192] Each of the information processing devices 10 uses multiple inference models distributed from the information processing server 20 to generate the above-mentioned feature quantities x1~x n Cluster the purchase log Lgc, which includes label y.

[0193] For example, in the case shown in Figure 9, the information processing device 10a classifies the purchase log Lgc1, which consists of 3 records corresponding to game category GC1, and the purchase log Lgc2, which consists of 2 records corresponding to game category GC2, into cluster C1.

[0194] Furthermore, the information processing device 10a classifies the two purchase log records Lgc3 corresponding to the game category GC3 into cluster C2.

[0195] Similarly, the information processing device 10a classifies the two purchase log records Lgc1 corresponding to game category GC1 into cluster C1, and the two purchase log records Lgc3 corresponding to game category GC3 into cluster C3.

[0196] Each of the information processing devices 10 according to this embodiment transmits the intermediate results calculated as described above to the information processing server 20 after performing the clustering described above, and receives information related to updating the inference model.

[0197] Subsequently, each of the information processing devices 10 according to this embodiment infers a label y using the collected feature quantity x and the inference model.

[0198] For example, the information processing device 10 can use an inference model to infer the likelihood that a user will purchase a certain game.

[0199] In this case, the information processing device 10 may also take measures such as explicitly presenting games that exceed a threshold in terms of the likelihood of purchase to the user, or placing them in an easily visible location in the online shop.

[0200] Next, we will describe an example where the main element of this embodiment is a person.

[0201] In this case, label y may be an indicator representing, for example, the person's physical or mental state.

[0202] Figure 10 is a diagram illustrating the operation of System 1 when the main element of this embodiment is a person and label y is an indicator related to the person's health status.

[0203] Figure 10 illustrates the operation of the information processing device 10 when it is installed in a medical institution.

[0204] In this example, the information processing device 10 stores a collection of medical examination logs Lpe for each person P.

[0205] Features x1~x included in the consultation log Lpe n Examples include the results of various tests such as blood pressure and heart rate, as well as symptoms.

[0206] Furthermore, label y may be a diagnosis made by a physician.

[0207] Each of the information processing devices 10 uses multiple inference models distributed from the information processing server 20 to generate the above-mentioned feature quantities x1~x n Cluster the consultation log Lpe, which includes label y.

[0208] For example, in the case shown in Figure 10, the information processing device 10a classifies the 3 records of medical examination log Lpe1 corresponding to person P1 and the 2 records of medical examination log Lpe2 corresponding to person P2 into cluster C1.

[0209] Furthermore, the information processing device 10a classifies the two records of medical examination logs Lpe3 corresponding to person P3 into cluster C2.

[0210] Similarly, the information processing device 10a classifies the two records of medical examination log Lpe4 corresponding to person P4 into cluster C1, and the two records of medical examination log Lpe5 corresponding to person P5 into cluster C3.

[0211] Each of the information processing devices 10 according to this embodiment transmits the intermediate results calculated as described above to the information processing server 20 after performing the clustering described above, and receives information related to updating the inference model.

[0212] Subsequently, each of the information processing devices 10 according to this embodiment infers a label y using the collected feature quantity x and the inference model.

[0213] For example, the information processing device 10 can use an inference model to infer the health status of a person based on new test results relating to that person.

[0214] This would make it possible to make a provisional assessment of a person's health condition without an actual medical diagnosis by a doctor.

[0215] Next, we will discuss the case where label y is an indicator representing a person's emotion.

[0216] Indicators representing the emotions of the person mentioned above include, for example, the user's facial expressions.

[0217] Figure 11 is a diagram illustrating the operation of System 1 when the main element of this embodiment is a person and the label y is the person's facial expression.

[0218] Figure 11 illustrates the operation of the information processing device 10 when it is a robot that communicates with the user.

[0219] In this example, the information processing device 10 stores a collection of shooting logs Lpp for each person P.

[0220] Features x1~x included in the shooting log Lpp nExamples thereof include, for example, the captured image, the position of each part on the face, the size of each part, and the like.

[0221] Also, the label y may be various inferred expressions.

[0222] Each of the information processing apparatuses 10 clusters the captured log Lpp including the feature amounts x1 to x and the label y as described above using a plurality of inference models distributed from the information processing server 20. n and the label y.

[0223] For example, in the case of an example shown in FIG. 11, the information processing apparatus 10a classifies the captured log Lpp1 for three records corresponding to the person P1 and the captured log Lpp2 for two records corresponding to the person P2 into the cluster C1.

[0224] Also, the information processing apparatus 10a classifies the captured log Lpp3 for three records corresponding to the person P3 into the cluster C2.

[0225] Similarly, the information processing apparatus 10a classifies the captured log Lpp4 for three records corresponding to the person P4 into the cluster C1 and classifies the captured log Lpp5 for three records corresponding to the person P5 into the cluster C3.

[0226] After performing the clustering as described above, each of the information processing apparatuses 10 according to the present embodiment transmits the intermediate result calculated as described above to the information processing server 20 and receives information related to the update of the inference model.

[0227] Thereafter, each of the information processing apparatuses 10 according to the present embodiment infers the label y using the collected feature amount x and the inference model.

[0228] For example, the information processing apparatus 10 can infer the expression of a person from an image of the person as a subject using the inference model.

[0229] In addition, each of the information processing apparatuses 10 according to the present embodiment may perform control such as changing the behavior toward the user according to the inferred expression.

[0230] <2. Hardware Configuration Example> Next, a hardware configuration example common to the information processing apparatus 10 and the information processing server 20 according to an embodiment of the present disclosure will be described.

[0231] FIG. 12 is a block diagram showing a hardware configuration example of an information processing apparatus 90 according to an embodiment of the present disclosure.

[0232] The information processing apparatus 90 may be an apparatus having a hardware configuration equivalent to that of the information processing apparatus 10 and the information processing server 20.

[0233] As shown in FIG. 19, the information processing apparatus 90 includes, for example, a processor 871, a ROM 872, a RAM 873, a host bus 874, a bridge 875, an external bus 876, an interface 877, an input device 878, an output device 879, a storage 880, a drive 881, a connection port 882, and a communication device 883. Note that the hardware configuration shown here is an example, and some of the components may be omitted. Further, other components not shown here may be further included.

[0234] (Processor 871) The processor 871 functions as, for example, an arithmetic processing unit or a control unit, and controls the overall operation or a part of the operation of each component based on various programs recorded in the ROM 872, the RAM 873, the storage 880, or the removable storage medium 901.

[0235] (ROM 872, RAM 873) The ROM 872 is a means for storing programs read by the processor 871, data used for arithmetic operations, and the like. In the RAM 873, for example, programs read by the processor 871 and various parameters that change as appropriate when executing the programs are stored temporarily or permanently.

[0236] (Host bus 874, bridge 875, external bus 876, interface 877) The processor 871, ROM 872, and RAM 873 are interconnected, for example, via a host bus 874 capable of high-speed data transmission. On the other hand, the host bus 874 is connected to an external bus 876, which has a relatively low data transmission speed, via a bridge 875. The external bus 876 is also connected to various components via an interface 877.

[0237] (Input device 878) The input device 878 may include, for example, a mouse, keyboard, touch panel, buttons, switches, and levers. Furthermore, a remote controller (hereinafter referred to as a remote control) capable of transmitting control signals using infrared or other radio waves may also be used as the input device 878. Additionally, the input device 878 may include audio input devices such as microphones.

[0238] (Output device 879) The output device 879 is, for example, a display device such as a CRT (Cathode Ray Tube), LCD, or organic EL; an audio output device such as a speaker or headphones; a printer, mobile phone, or facsimile, or any other device capable of visually or audibly notifying the user of acquired information. Furthermore, the output device 879 according to this disclosure includes various vibration devices capable of outputting tactile stimuli.

[0239] (Storage 880) Storage 880 is a device for storing various types of data. Examples of storage 880 include magnetic storage devices such as hard disk drives (HDDs), semiconductor storage devices, optical storage devices, or magneto-optical storage devices.

[0240] (Drive 881) The drive 881 is a device that reads information recorded on a removable storage medium 901, such as a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory, or writes information to the removable storage medium 901.

[0241] (Removable storage medium 901) The removable storage medium 901 is, for example, DVD media, Blu-ray® media, HD DVD media, various semiconductor storage media, etc. Of course, the removable storage medium 901 may also be, for example, an IC card equipped with a contactless IC chip, or an electronic device, etc.

[0242] (Connection port 882) Connection port 882 is a port for connecting external devices 902, such as a USB (Universal Serial Bus) port, IEEE1394 port, SCSI (Small Computer System Interface), RS-232C port, or optical audio terminal.

[0243] (External connection device 902) External connected devices 902 include, for example, a printer, a portable music player, a digital camera, a digital video camera, or an IC recorder.

[0244] (Communication device 883) The communication device 883 is a communication device for connecting to a network, and is, for example, a communication card for wired or wireless LAN, Bluetooth®, or WUSB (Wireless USB), a router for optical communication, a router for ADSL (Asymmetric Digital Subscriber Line), or a modem for various types of communication.

[0245] <3. Summary> As described above, the information processing apparatus 10 according to an embodiment of the present disclosure clusters hierarchical data based on a plurality of inference models distributed from the information processing server 20, and includes a learning unit 120 that performs learning using the corresponding inference model for each cluster.

[0246] Also, the information processing apparatus 10 according to an embodiment of the present disclosure includes a communication unit 130 that transmits the intermediate results generated for each cluster in the above learning by the learning unit 120 to the information processing server 20.

[0247] According to the above configuration, it is possible to achieve both privacy protection and high inference accuracy.

[0248] As described above, the preferred embodiments of the present disclosure have been described in detail with reference to the accompanying drawings, but the technical scope of the present disclosure is not limited to such examples. It is obvious that those having ordinary knowledge in the technical field of the present disclosure can conceive of various modification examples or correction examples within the scope of the technical idea described in the claims, and it is naturally understood that these also belong to the technical scope of the present disclosure.

[0249] Also, each step related to the processing described in this specification does not necessarily need to be processed in time series in the order described in the flowchart or sequence diagram. For example, each step related to the processing of each device may be processed in an order different from the described order or may be processed in parallel.

[0250] Furthermore, the series of processes performed by each device described herein may be implemented using software, hardware, or a combination of software and hardware. The programs constituting the software are pre-stored in a non-transitory computer-readable medium located inside or outside each device and readable by a computer. Each program is then loaded into RAM, for example, when executed by a computer, and executed by various processors. The storage medium may be, for example, a magnetic disk, an optical disk, a magneto-optical disk, or flash memory. Alternatively, the computer programs may be distributed without using a storage medium, for example, via a network.

[0251] Furthermore, the effects described herein are merely descriptive or illustrative and not limiting. In other words, the technology relating to this disclosure may produce other effects that are obvious to those skilled in the art from the description herein, in addition to or instead of the effects described herein.

[0252] Furthermore, the following configurations also fall within the technical scope of this disclosure. (1) A learning unit that clusters hierarchical data based on multiple inference models distributed from an information processing server, and performs learning using the corresponding inference model for each cluster, A communication unit that transmits the intermediate results generated for each cluster in the learning process by the learning unit to the information processing server, Equipped with, The hierarchical data includes information that identifies the main element, and logs collected or generated in association with the main element. Information processing device. (2) The learning unit clusters the hierarchical data relating to different principal elements based on a plurality of inference models. The information processing device described in (1) above. (3) The learning unit performs clustering so that the hierarchical data relating to the same principal element are classified into the same cluster. The information processing device described in (1) or (2) above. (4) The learning unit performs clustering so that the hierarchical data relating to different principal elements are classified into the same cluster. An information processing device as described in any of (1) to (3) above. (5) The log includes features and labels, An information processing device as described in any of (1) to (4) above. (6) The aforementioned intermediate results include values ​​calculated from the features and labels, The information processing device described in (5) above. (7) The learning unit infers the label based on the features and the inference model. The information processing device described in (5) or (6) above. (8) The communication unit receives information relating to the inference model that has been updated based on the intermediate results received by the information processing server from multiple devices, and passes it on to the learning unit. An information processing device as described in any of (1) to (7) above. (9) The main element includes a device that communicates with the communication unit. An information processing device as described in any of (5) to (7) above. (10) The aforementioned main element includes an access point that communicates with the communication unit, The information processing device described in (9) above. (11) The label includes an indicator representing the communication quality related to the access point. The information processing device described in (10) above. (12) The aforementioned main elements include the product category, An information processing device as described in any of (5) to (7) above. (13) The aforementioned label includes an indicator related to the purchase of products belonging to the aforementioned category. The information processing device described in (12) above. (14) The aforementioned main elements include, An information processing device as described in any of (5) to (7) above. (15) The aforementioned label includes an indicator representing the physical or mental state of a person. The information processing device described in (14) above. (16) The aforementioned label includes an indicator that represents a person's emotions. The information processing device described in (14) above. (17) The processor, The process involves clustering hierarchical data based on multiple inference models distributed from an information processing server, and performing training using the corresponding inference model for each cluster. The intermediate results generated for each cluster during the learning process are transmitted to the information processing server. Includes, The hierarchical data includes information that identifies the main element, and logs collected or generated in association with the main element. Information processing methods. (18) Computers, A learning unit that clusters hierarchical data based on multiple inference models distributed from an information processing server, and performs learning using the corresponding inference model for each cluster, A communication unit that transmits the intermediate results generated for each cluster in the learning process by the learning unit to the information processing server, Equipped with, The hierarchical data includes information that identifies the main element, and logs collected or generated in association with the main element. Information processing device, A non-temporary storage medium readable by a computer that has stored a program to function as such. (19) A learning unit that generates multiple inference models corresponding to multiple clusters, A communication unit that transmits information relating to multiple inference models generated by the learning unit to multiple information processing devices, Equipped with, The communication unit receives from multiple information processing devices intermediate results generated by learning based on hierarchical data clustered based on multiple inference models and the corresponding inference model for each cluster, The learning unit updates the multiple inference models based on the multiple intermediate results, The hierarchical data includes information that identifies the main element, and logs collected or generated in association with the main element. Information processing server. (20) The processor, This involves generating multiple inference models, each corresponding to a different cluster, Transmitting information related to multiple generated inference models to multiple information processing devices, Intermediate results generated by learning based on hierarchical data clustered based on multiple inference models and the corresponding inference model for each cluster are received from multiple information processing devices. Updating multiple inference models based on multiple intermediate results, Includes, The hierarchical data includes information that identifies the main element, and logs collected or generated in association with the main element. Information processing methods. (twenty one) Computers, A learning unit that generates multiple inference models corresponding to multiple clusters, A communication unit that transmits information relating to multiple inference models generated by the learning unit to multiple information processing devices, Equipped with, The communication unit receives from multiple information processing devices intermediate results generated by learning based on hierarchical data clustered based on multiple inference models and the corresponding inference model for each cluster, The learning unit updates the multiple inference models based on the multiple intermediate results, The hierarchical data includes information that identifies the main element, and logs collected or generated in association with the main element. Information processing server, A non-temporary storage medium readable by a computer that has stored a program to function as such. [Explanation of Symbols]

[0253] 1 System 10 Information Processing Devices 110 Sensor section 120 Learning Department 130 Communications Department 20 Information Processing Server 210 Learning Department 220 Communications Department 40 access points

Claims

1. To generate multiple inference models corresponding to multiple clusters, This involves clustering hierarchical data obtained from multiple information processing devices, each of which is transmitted from the respective devices, based on multiple inference models, and updating the multiple inference models based on the intermediate results generated by learning each cluster. Distributing multiple inference models to multiple information processing devices, including, An information processing method executed by a server system.

2. Hierarchical data includes sensor data as one element. The information processing method according to claim 1.

3. Sensor data is obtained by a sensor provided by the information processing device. The information processing method according to claim 2.

4. Hierarchical data includes information that identifies the main element, and logs collected or generated in association with the main element. The information processing method according to claim 1.

5. Hierarchical data includes logs collected in connection with communication, The information processing method according to claim 1.

6. The hierarchical data includes information that identifies the access point with which the information processing device communicates, and a log associated with the access point. The information processing method according to claim 5.

7. The log includes at least one of the following: a feature quantity relating to the signal strength of the communication or a label representing the communication quality. The information processing method according to claim 5.

8. Distributing information related to the updating of an inference model to multiple information processing devices, Further including, The information processing method according to claim 1.

9. Intermediate results obtained using an updated inference model, which is updated using information related to the update of the inference model distributed from the server system, are received from multiple information processing devices. Further including, The information processing method according to claim 1.

10. The main element includes a device that communicates with an information processing device, The information processing method according to claim 4.

11. A learning unit that generates multiple inference models corresponding to multiple clusters, A communication unit that distributes multiple inference models to multiple information processing devices, Equipped with, The learning unit updates the multiple inference models based on the intermediate results generated by the learning of each cluster, which are obtained by clustering hierarchical data transmitted from multiple information processing devices. Server system.

12. Hierarchical data includes sensor data as one element. The server system according to claim 11.

13. Sensor data is obtained by a sensor provided by the information processing device. The server system according to claim 12.

14. Hierarchical data includes information that identifies a main element, and logs collected or generated in association with the main element. The server system according to claim 11.

15. Hierarchical data includes logs collected in connection with communication, The server system according to claim 11.

16. The hierarchical data includes information that identifies the access point with which the information processing device communicates, and a log associated with the access point. The server system according to claim 15.

17. The log includes at least one of the following: a feature quantity relating to the signal strength of the communication or a label representing the communication quality. The server system according to claim 15.

18. The communication unit receives from a plurality of information processing devices intermediate results obtained using an updated inference model using information related to the update of the inference model distributed from the server system. The server system according to claim 11.

19. The main element includes a device that communicates with an information processing device, The server system according to claim 14.

Images