Information processing device, program, and information processing method

Abstract

An information processing device (100) is provided with: an attention mechanism unit (113) that uses an attention mechanism learning model, which is a learning model of an attention mechanism, to weight and add a plurality of input data as a time series or a plurality of variables calculated from the plurality of input data by a plurality of weight values, thereby calculating a context variable; a determination unit (114) that estimates one determination from among a plurality of determinations on the basis of the reliability of the plurality of determinations calculated on the basis of the context variable and the latest one input data included in the plurality of input data or the latest one variable included in the plurality of variables, on the basis of the reliability of the plurality of determinations calculated on the basis of the context variable and the latest one input data included in the plurality of input data or on the basis of the latest one variable included in the plurality of variables; and a data extraction unit (116) that extracts one or more pieces of input data from the plurality of pieces of input data, said one or more pieces of input data being an estimation factor for one determination, by referring to the plurality of weight values.

Description

Technical Field

[0001] The present disclosure relates to an information processing device, a program, and an information processing method. Background Art

[0002] As a technology for improving the estimation accuracy based on the learning model, there is an attention mechanism. For example, the anomaly detection device described in Patent Document 1 has an anomaly detection unit for detecting anomalies in time series data. The anomaly detection unit includes: an encoding unit that uses multiple LSTM units to encode the time series data; an attention layer that calculates the weight of attention for the output from the encoding unit; a context generation unit that applies the weight to the output from the encoding unit to generate a context vector; and a decoding unit that uses multiple LSTM units to reconstruct the time series data based on the context vector. As a result, it is possible to achieve improved accuracy and efficient learning for anomaly detection.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: International Publication No. 2021 / 100179 Summary of the invention

[0006] Problems to be solved by the invention

[0007] However, in a learning model using deep reinforcement learning, internal processing becomes a black box, and therefore the internal processing cannot be seen. Therefore, the user cannot easily understand how the judgment in the learning model is performed.

[0008] Therefore, an object of one or more embodiments of the present disclosure is to easily grasp data that serves as a basis for judgment in a learning model that uses an attention mechanism.

[0009] Means for solving problems

[0010] An information processing device of one embodiment of the present invention is characterized in that it comprises: an attention mechanism unit, which uses a learning model of the attention mechanism, i.e., an attention mechanism learning model, to weight and add multiple input data as a time series or multiple variables calculated based on the multiple input data by multiple weight values, thereby calculating a context variable; a judgment unit, which estimates one judgment from the multiple judgments based on the reliabilities of the multiple judgments, wherein the reliability of the multiple judgments is calculated based on the context variable and the latest input data included in the multiple input data or the latest variable included in the multiple variables; and a data extraction unit, which refers to the multiple weight values ​​to extract one or more input data from the multiple input data that become estimation factors of the one judgment.

[0011] A program of one embodiment of the present invention is characterized in that it enables a computer to function as the following parts: an attention mechanism unit, which uses a learning model of the attention mechanism, i.e., an attention mechanism learning model, to weight and add multiple input data as a time series or multiple variables calculated based on the multiple input data by multiple weight values, thereby calculating a context variable; a judgment unit, which estimates one judgment from the multiple judgments based on the reliabilities of the multiple judgments, wherein the reliability of the multiple judgments is calculated based on the context variable and the latest one of the input data included in the multiple input data or the latest one of the variables included in the multiple variables; and a data extraction unit, which refers to the multiple weight values ​​to extract one or more input data from the multiple input data that become estimation factors of the one judgment.

[0012] An information processing method of one embodiment of the present invention is characterized in that a learning model of an attention mechanism, i.e., an attention mechanism learning model, is used to weight and add multiple input data as a time series or multiple variables calculated based on the multiple input data by multiple weight values, thereby calculating a context variable, and estimating one judgment from the multiple judgments based on the reliabilities of the multiple judgments, wherein the reliabilities of the multiple judgments are calculated based on the context variables and the latest input data included in the multiple input data or the latest variable included in the multiple variables, and referring to the multiple weight values, one or more input data that become estimation factors of the one judgment are extracted from the multiple input data.

[0013] Effects of the Invention

[0014] According to one or more aspects of the present disclosure, it is possible to easily grasp data that serves as a basis for judgment in a learning model using an attention mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] [ Figure 1 ] is a block diagram schematically showing the structure of the information processing device of embodiment 1.

[0016] [ Figure 2 ](A) and (B) are block diagrams showing hardware configuration examples.

[0017] [ Figure 3 ] is a schematic diagram used to illustrate the processing in the information processing device of embodiment 1.

[0018] [ Figure 4 ] is a block diagram schematically showing the structure of the information processing device of embodiment 2.

[0019] [ Figure 5] is a schematic diagram used to illustrate the processing in the information processing device of embodiment 2.

[0020] [ Figure 6 ] is a block diagram schematically showing the structure of an information processing device according to embodiment 3.

[0021] [ Figure 7 ] is a schematic diagram used to illustrate the processing in the information processing device of embodiment 3. DETAILED DESCRIPTION

[0022] Implementation method 1.

[0023] Figure 1 This is a block diagram schematically showing the configuration of information processing device 100 according to Embodiment 1.

[0024] The information processing device 100 includes a storage unit 101 , a communication unit 102 , an input unit 103 , a display unit 104 , and a control unit 110 .

[0025] The storage unit 101 stores programs and data necessary for processing in the information processing device 100 .

[0026] For example, the storage unit 101 stores at least an attention mechanism learning model which is a learning model used in the attention mechanism executed by the control unit 110. In the first embodiment, as described later, the storage unit 101 also stores an extraction learning model and a judgment learning model.

[0027] In addition, the storage unit 101 also stores judgment input data information indicating input data judged to be important (in other words, a factor that becomes a judgment result) based on the estimation result of the attention mechanism.

[0028] The communication unit 102 communicates with other devices. For example, the communication unit 102 communicates with other devices via a network such as the Internet.

[0029] The input unit 103 receives input from a user of the information processing device 100 .

[0030] The display unit 104 displays information to a user of the information processing device 100. For example, the display unit 104 displays various screen images.

[0031] The control unit 110 controls the processing in the information processing device 100. For example, the control unit 110 obtains input data and calculates variables required for making judgments based on the input data, namely, state variables. In addition, the control unit 110 performs weighted addition of the state variables through the attention mechanism, thereby calculating context state variables, and estimates a certain judgment based on the context state variables. Then, the control unit 110 refers to the weights based on the attention mechanism, thereby extracting the input data that becomes the factor of the estimated judgment, namely, the judgment result, and stores the judgment input data information representing the input data in the storage unit 101. Here, it can be judged that the extracted input data has a greater impact on the estimation.

[0032] In addition, below, state variables are also simply referred to as variables, and context state variables are also simply referred to as context variables.

[0033] The control unit 110 includes a data acquisition unit 111 , a variable extraction unit 112 , an attention mechanism unit 113 , a determination unit 114 , an attention time information extraction unit 115 , and a data extraction unit 116 .

[0034] The data acquisition unit 111 acquires input data. The data acquisition unit 111 may acquire input data, for example, via the communication unit 102. In addition, when the input data is stored in the storage unit 101, the data acquisition unit 111 may acquire the input data from the storage unit 101. It is assumed that the input data acquired here is time series data. The acquired input data is provided to the variable extraction unit 112 and the data extraction unit 116.

[0035] The variable extraction unit 112 extracts state variables, which are variables that can be used for determination, from the input data acquired by the data acquisition unit 111 .

[0036] Here, the variable extraction unit 112 extracts the state variables using an extraction learning model which is a learning model for extracting the state variables from the input data. In addition, it is assumed that the state variables extracted by the variable extraction unit 112 are in a time series.

[0037] The attention mechanism unit 113 performs a weighted sum based on a well-known attention mechanism on the state variables extracted by the variable extraction unit 112, thereby calculating the context state variable. For example, the attention mechanism unit 113 estimates a plurality of weight values ​​for the state variables extracted by the variable extraction unit 112 using the attention mechanism learning model stored in the storage unit 101, performs weighting by the plurality of weight values, and adds the weighted state variables, thereby calculating the context state variable as the estimation result.

[0038] The judgment unit 114 estimates a judgment from the multiple judgments based on the reliability of the multiple judgments, and the reliability of the multiple judgments is calculated based on the context state variable estimated by the attention mechanism unit 113 and the latest state variable included in the multiple state variables.

[0039] Here, the judgment unit 114 performs the estimation using a judgment learning model which is a learning model for estimating a judgment based on context variables.

[0040] The attention time information extraction unit 115 generates attention time information and provides the attention time information to the data extraction unit 116. The attention time information represents the time of input data corresponding to multiple weight values ​​estimated by the attention mechanism unit 113 and the state variables respectively weighted by the multiple weight values.

[0041] The data extraction unit 116 refers to the weight value represented by the attention time information to extract one or more input data that are factors of the judgment result from the input data. In other words, the data extraction unit 116 extracts the input data that is considered to have a greater impact on the judgment result and generates judgment input data information representing the input data. Then, the data extraction unit 116 stores the judgment input data information in the storage unit 101.

[0042] Specifically, the data extraction unit 116 can judge the input data corresponding to the weight value when the weight value represented by the attention time information exceeds the first threshold value as a predetermined threshold value, and the two input data corresponding to the two weight values ​​when the change size of the two weight values ​​corresponding to the continuous time represented by the attention time information exceeds the second threshold value as a predetermined threshold value, as input data that become factors of the judgment result. In addition, the change size can be a difference or a ratio.

[0043] For example, Figure 2 As shown in (A), part or all of the control unit 110 described above can be composed of a memory 10 and a processor 11 such as a CPU (Central Processing Unit) that executes a program stored in the memory 10. In other words, the information processing device 100 can be implemented by a so-called computer. Such a program can be provided through a network, and can also be provided by recording it in a recording medium. That is, such a program is provided as a program product, for example.

[0044] In addition, for example, Figure 2As shown in (B), part or all of the control unit 110 can also be composed of a processing circuit 12 such as a single circuit, a composite circuit, a processor operated by a program, a parallel processor operated by a program, an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

[0045] As described above, the control unit 110 can be implemented by a processing circuit network.

[0046] The storage unit 101 can be realized by a storage device such as a HDD (Hard Disk Drive) or an SSD (Solid State Drive).

[0047] The communication unit 102 can be realized by a communication interface such as a NIC (Network Interface Card).

[0048] The input unit 103 can be realized by an input interface such as a keyboard or a mouse.

[0049] The display unit 104 can be realized by a display.

[0050] Figure 3 This is a schematic diagram for explaining the processing in the information processing device 100 according to the first embodiment.

[0051] First, the data acquisition unit 111 acquires input data X t-n , X t-n+1 , X t-1 , X t (S10). Here, suppose the input data X t-n , X t-n+1 , X t-1 , X t is a sensor value as an observation value, and is data of a time series of tn, t-n+1, ​​t-1, t (t and n are positive integers). For example, image data can be used as input data.

[0052] The data acquisition unit 111 converts the acquired input data X t-n , X t-n+1 , X t-1 , X t Provided to the variable extraction unit 112 and the data extraction unit 116.

[0053] The variable extraction unit 112 extracts the input data X t-n , X t-n+1 , X t-1, X t Extract the state variable S which is helpful for the judgment of the judgment unit 114 t-n , S t-n+1 , S t-1 , S t (S11).

[0054] Here, the variable extraction unit 112 uses the neural network model stored in the storage unit 101, that is, the extraction learning model, to extract the input data X t-n , X t-n+1 , X t-1 , X t Extract the state variable S t-n , S t-n+1 , S t-1 , S t .

[0055] The variable extraction unit 112 extracts the state variable S t-n , S t-n+1 , S t-1 , S t Provided to the attention mechanism unit 113.

[0056] In addition, here, the variable extraction unit 112 uses the extraction learning model, but the first embodiment is not limited to such an example, as long as a certain function is used to extract the state variable S t-n , S t-n+1 , S t-1 , S t That's it.

[0057] The attention mechanism unit 113 is for the state variable S t-n , S t-n+1 , S t-1 , S t , the weighted sum is calculated by estimating the weight values ​​using the learning model, thereby calculating the context state variable (S12).

[0058] The attention mechanism unit 113 provides the calculated context state variable to the judgment unit 114.

[0059] The determination unit 114 makes a determination based on the context state variable and the latest state variable St ( S13 ).

[0060] Here, the judgment unit 114 uses the judgment learning model which is a neural network model stored in the storage unit 101 to estimate the judgment based on the context state variables and the latest state variables.

[0061] The attention time information extraction unit 115 extracts the weight value estimated in the attention mechanism unit 113 in step S12 and the time of the corresponding input data, and generates attention time information indicating the extracted weight value and time (S14). The generated attention time information is provided to the data extraction unit 116.

[0062] The data extraction unit 116 refers to the attention time information to extract input data that is a factor of the determination result in the determination unit 114 from among the input data, and generates determination input data information representing the input data ( S15 ).

[0063] Then, the storage unit 101 stores the determination input data information generated by the data extraction unit 116 ( S16 ).

[0064] As described above, according to Embodiment 1, it is possible to easily grasp the data that serves as a basis for judgment in a learning model using an attention mechanism.

[0065] Implementation method 2.

[0066] Figure 4 This is a block diagram schematically showing the structure of the information processing device 200 according to the second embodiment.

[0067] The information processing device 200 includes a storage unit 201 , a communication unit 102 , an input unit 103 , a display unit 104 , and a control unit 210 .

[0068] The communication unit 102 , the input unit 103 , and the display unit 104 of the information processing device 200 according to the second embodiment are the same as the communication unit 102 , the input unit 103 , and the display unit 104 of the information processing device 100 according to the first embodiment.

[0069] The storage unit 201 stores programs and data necessary for processing in the information processing device 200 .

[0070] The storage unit 201 in the second embodiment stores the same data as in the first embodiment, and also stores input data with meaning generated by the control unit 210 described later.

[0071] The control unit 210 controls processing in the information processing device 200 .

[0072] The control unit 210 in the second embodiment extracts the meaning of the input data and performs processing to interpret the basis for the determination.

[0073] The control unit 210 has a data acquisition unit 111, a variable extraction unit 112, an attention mechanism unit 113, a judgment unit 114, an attention time information extraction unit 115, a data extraction unit 216 and a data meaning acquisition unit 217.

[0074] The data acquisition unit 111, variable extraction unit 112, attention mechanism unit 113, judgment unit 114 and attention time information extraction unit 115 of the control unit 210 in implementation mode 2 are the same as the data acquisition unit 111, variable extraction unit 112, attention mechanism unit 113, judgment unit 114 and attention time information extraction unit 115 of the control unit 110 in implementation mode 1.

[0075] However, the data acquisition unit 111 in the second embodiment supplies the acquired input data to the variable extraction unit 112 and the data meaning acquisition unit 217 .

[0076] The data meaning acquisition unit 217 acquires the meaning of the input data from the data acquisition unit 111. Here, the data meaning acquisition unit 217 receives an input of the meaning of the input data from the user via the input unit 103, thereby acquiring the meaning of the input data. For example, when the input data is image data, the meaning of the input data is identification information (e.g., the name of the object) for identifying an object such as a person or an object contained in the image data.

[0077] Then, the data meaning acquisition unit 217 stores the input data with meaning obtained by adding the meaning of the input data to the input data in the storage unit 201 .

[0078] The data extraction unit 216 interprets the basis of the judgment result, that is, the basis of the judgment result, based on the meaning of the input data that is a factor of the judgment result by the judgment unit 114 .

[0079] For example, the data extraction unit 216 refers to the attention time information to extract the meaningful input data stored in the storage unit 201, which is considered to have a greater impact on the judgment result. Then, the data extraction unit 216 interprets the judgment basis of the judgment result according to the meaning of the extracted meaningful input data. For the interpretation of the judgment basis, it is assumed that the interpretation method is predetermined according to the judgment and weight value performed by the judgment unit 114. For example, in the case where the weight value represented by the attention time information rises sharply, it can be interpreted that the objects appearing before and after the sharp rise are the judgment basis. In addition, in the case where the weight value represented by the attention time information exceeds a certain threshold, it can be interpreted that the physical quantity such as the distance, position, size, etc. of the predetermined object in the time is the judgment basis. The predetermined object here can be determined according to the judgment performed by the judgment unit 114. For example, in the case where the judgment performed by the judgment unit 114 is the operation of the car, the predetermined object is an oncoming car, a wall or a pedestrian.

[0080] Then, the data extraction unit 216 stores the judgment basis information indicating the content interpreted based on the extracted meaningful input data, that is, the judgment basis, in the storage unit 201. The judgment basis information may also include the meaningful input data used in interpreting the judgment basis.

[0081] Figure 5 This is a schematic diagram for explaining the processing in the information processing device 200 according to the second embodiment.

[0082] about Figure 5 The processing of S10 to S14 is Figure 3 The processes of S10 to S14 shown are the same.

[0083] In the second embodiment, the data meaning acquisition unit 217 extracts the meaning of the input data from the data acquisition unit 111 and generates input data with meaning by adding the meaning of the input data to the input data ( S27 ). The generated input data with meaning is stored in the storage unit 201 .

[0084] The data extraction unit 216 refers to the attention time information to extract the meaningful input data that is a factor of the judgment result from the meaningful input data stored in the storage unit 201, and interprets the judgment basis of the judgment result based on the meaning of the extracted meaningful input data (S28). Then, the data extraction unit 216 generates judgment basis information indicating the content and the interpretation based on the extracted meaningful input data.

[0085] Then, the storage unit 201 stores the determination basis information (S29).

[0086] As described above, according to the second embodiment, since the judgment basis information indicating the content interpreted as the judgment basis is generated, the judgment basis can be presented so that the user can easily understand it.

[0087] Implementation method 3.

[0088] Figure 6 This is a block diagram schematically showing the structure of the information processing device 300 according to the third embodiment.

[0089] The information processing device 300 includes a storage unit 301 , a communication unit 102 , an input unit 103 , a display unit 104 , and a control unit 310 .

[0090] The communication unit 102 , the input unit 103 , and the display unit 104 of the information processing device 300 according to the third embodiment are the same as the communication unit 102 , the input unit 103 , and the display unit 104 of the information processing device 100 according to the first embodiment.

[0091] The storage unit 301 stores programs and data necessary for processing in the information processing device 300 .

[0092] The storage unit 301 in the third embodiment stores the same data as in the first embodiment, and also stores input data with meaning and a determination rule generated by a control unit 310 described later.

[0093] The control unit 310 controls processing in the information processing device 300 .

[0094] Similar to the second embodiment, the control unit 310 in the third embodiment extracts the meaning of the input data and performs processing to interpret the basis for the determination.

[0095] Furthermore, the control unit 310 in the third embodiment generates a judgment rule indicating the judgment basis and a judgment result inferred from the meaningful input data corresponding to the judgment basis, and stores the judgment rule in the storage unit 301 .

[0096] The control unit 310 has a data acquisition unit 111, a variable extraction unit 112, an attention mechanism unit 113, a judgment unit 114, an attention time information extraction unit 115, a data extraction unit 216, a data meaning acquisition unit 217 and a judgment rule generation unit 318.

[0097] The data acquisition unit 111, variable extraction unit 112, attention mechanism unit 113 and judgment unit 114 of the control unit 310 in implementation mode 3 are the same as the data acquisition unit 111, variable extraction unit 112, attention mechanism unit 113 and judgment unit 114 of the control unit 110 in implementation mode 1.

[0098] Furthermore, the data extraction unit 216 and the data meaning acquisition unit 217 of the control unit 310 in the third embodiment are the same as the data extraction unit 216 and the data meaning acquisition unit 217 of the control unit 210 in the second embodiment.

[0099] The judgment rule generation unit 318 generates a judgment rule that associates the judgment basis, which is the content interpreted from the meaningful input data by the data extraction unit 216, with the estimated judgment result. In addition, the judgment rule may also include the meaningful input data used in interpreting the judgment basis.

[0100] Then, the determination rule generating unit 318 stores the determination rule in the storage unit 301 .

[0101] Figure 7 This is a schematic diagram for explaining the processing in the information processing device 300 according to the third embodiment.

[0102] about Figure 7 The processing of S10 to S14 is Figure 3 The processes of S10 to S14 shown are the same.

[0103] In the third embodiment, similarly to the second embodiment, the data meaning acquisition unit 217 extracts the meaning of the input data from the data acquisition unit 111 and generates input data with meaning by adding the meaning of the input data to the input data (S27). The generated input data with meaning is stored in the storage unit 301.

[0104] The data extraction unit 216 refers to the attention time information to extract the meaningful input data stored in the storage unit 301 that serves as a factor for the judgment result, and interprets the judgment basis of the judgment result based on the meaning of the extracted meaningful input data (S28).

[0105] The judgment rule generation unit 318 generates a judgment rule indicating the judgment basis interpreted in step S28 and the judgment result inferred by the judgment unit 114. In addition, the judgment rule may include the meaningful input data extracted in step S27.

[0106] Then, the storage unit 301 stores the determination rule ( S31 ).

[0107] As described above, according to Embodiment 3, since a judgment rule representing a judgment basis and a judgment result derived from the judgment basis is generated, by accumulating the judgment rule, for example, a rule-based AI (Artificial Intelligence) model such as a well-known decision tree can be generated. In addition, since the judgment rule is in a form that is easy for the user to understand, the user can easily understand the content of the inference performed by the information processing device 300.

[0108] In the above-described embodiments 1 to 3, the variable extractor 112 extracts state variables from input data, but the embodiments 1 to 3 are not limited to such examples. For example, when the input data is suitable for processing, the processing in the variable extractor 112 may not be performed.

[0109] In this case, the attention mechanism unit 113 uses a learning model of the attention mechanism, i.e., an attention mechanism learning model, to weight and add a plurality of input data as a time series by a plurality of weight values, thereby calculating a context variable. In addition, the judgment unit 114 estimates a judgment from a plurality of judgments based on the reliability of the plurality of judgments, the reliability of the plurality of judgments being calculated based on the context variable and the latest input data included in the plurality of input data.

[0110] Description of symbols

[0111] 100, 200, 300: information processing device; 101, 201, 301: storage unit; 102: communication unit; 103: input unit; 104: display unit; 110, 210, 310: control unit; 111: data acquisition unit; 112: variable extraction unit; 113: attention mechanism unit; 114: judgment unit; 115: attention time information extraction unit; 116, 216: data extraction unit; 217: data meaning acquisition unit; 318: judgment rule generation unit.

Claims

1. An information processing device, characterized in that: The information processing device has: An attention mechanism unit, which uses a learning model of an attention mechanism, i.e., an attention mechanism learning model, to weight and add a plurality of input data as a time series or a plurality of variables calculated based on the plurality of input data by a plurality of weight values, thereby calculating a context variable; a judgment unit that estimates a judgment from the plurality of judgments based on reliabilities of the plurality of judgments, wherein the reliabilities of the plurality of judgments are calculated based on the context variable and a latest input data included in the plurality of input data or a latest variable included in the plurality of variables; as well as A data extraction unit refers to the plurality of weight values ​​to extract one or more input data to be estimation factors of the one judgment from the plurality of input data.

2. The information processing device according to claim 1, characterized in that When one weight value included in the plurality of weight values ​​exceeds a first threshold value that is a predetermined threshold value, the data extraction unit extracts input data corresponding to the one weight value.

3. The information processing device according to claim 1 or 2, characterized in that: When a change in two weight values ​​corresponding to consecutive times in the time series included in the plurality of weight values ​​exceeds a second threshold value which is a predetermined threshold value, the data extraction unit extracts two input data corresponding to the two weight values.

4. The information processing device according to any one of claims 1 to 3, characterized in that: The information processing device further includes a data meaning acquisition unit, the data meaning acquisition unit acquiring the meaning of the one or more input data, The data extraction unit estimates a basis for the one judgment based on the meaning interpretation of the one or more input data.

5. The information processing device according to claim 4, characterized in that The information processing device further includes a judgment rule generating unit that generates a judgment rule that associates the judgment basis with the one judgment.

6. The information processing device according to claim 5, characterized in that: The information processing device further includes a storage unit that stores the determination rule.

7. A program, characterized in that The program enables the computer to function as: An attention mechanism unit, which uses a learning model of an attention mechanism, i.e., an attention mechanism learning model, to weight and add a plurality of input data as a time series or a plurality of variables calculated based on the plurality of input data by a plurality of weight values, thereby calculating a context variable; a judgment unit that estimates a judgment from the plurality of judgments based on reliabilities of the plurality of judgments, wherein the reliabilities of the plurality of judgments are calculated based on the context variable and a latest input data included in the plurality of input data or a latest variable included in the plurality of variables; as well as A data extraction unit refers to the plurality of weight values ​​to extract one or more input data to be estimation factors of the one judgment from the plurality of input data.

8. An information processing method, characterized in that: A learning model using an attention mechanism, i.e., an attention mechanism learning model, weights and adds a plurality of input data as a time series or a plurality of variables calculated based on the plurality of input data by a plurality of weight values, thereby calculating a context variable, estimating a judgment from the plurality of judgments according to the reliabilities of the plurality of judgments, wherein the reliabilities of the plurality of judgments are calculated according to the context variable and the latest one of the input data included in the plurality of input data or the latest one of the variables included in the plurality of variables, The plurality of weight values ​​are referred to, thereby extracting one or more input data to be estimation factors of the one judgment from the plurality of input data.

Images