Information processing equipment, program and information processing procedures

Abstract

An information processing facility (100) comprises an attention mechanism unit (113) that utilizes an attention mechanism learning model, which is a learning model of an attention mechanism, to compute a context variable by weighting a plurality of input data elements organized as a time series, or a plurality of variables computed from the input data elements, with a plurality of weight values, and adding the weighted input data elements or the weighted variables; a decision unit (114) that infers a single decision from a plurality of decisions based on confidence levels of the plurality of decisions computed from the context variable and a most recent input data element contained in the plurality of input data elements, or a most recent variable contained in the plurality of variables;and a data extraction unit (116) that extracts at least one input data element, which is a factor in the inference of the individual decision, from the multitude of input data elements by reference to the multitude of weighting values.

Description

TECHNICAL FIELD

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

[0002] An attention mechanism is a technique for improving the accuracy of inference made by a learning model. For example, the anomaly detection device described in PTL 1 includes an anomaly detection unit that detects anomalies in time series data. The anomaly detection unit includes an encoding unit that encodes the time series data using a plurality of LSTM cells; an attention layer that calculates attention weight for an output from the encoding unit; a context generation unit that applies the weight to an output from the encoding unit to generate a context vector; and a decoding unit that reconstructs the time series data using the LSTM cells based on the context vector to achieve improved accuracy and efficient learning for anomaly detection. PRIOR ART PATENT REFERENCES

[0003] Patent Literature 1: International Publication No. 2021 / 100179 SUMMARY OF THE INVENTION PROBLEM TO BE SOLVED BY THE INVENTION

[0004] However, the internal processing of a learning model using deep reinforcement learning is a black box and therefore invisible. Therefore, a user cannot easily understand how a decision is made in a learning model.

[0005] Accordingly, it is an object of at least one aspect of the disclosure to easily locate data that serves as the basis for a decision made by a learning model using an attention mechanism. MEANS TO SOLVED THE TASK

[0006] An information processing device according to one aspect of the disclosure comprises: an attention mechanism unit configured to use an attention mechanism learning model to calculate a context variable by weighting a plurality of input data items organized as a time series or a plurality of variables calculated from the input data items with a plurality of weight values, and adding the weighted input data items or the weighted variables, wherein the attention mechanism learning model is a learning model of an attention mechanism;a decision unit configured to infer a single decision from a plurality of decisions based on confidence levels of the plurality of decisions calculated from the context variable and a most recent input data item included in the plurality of input data items or a last variable included in the plurality of variables; a data extraction unit configured to extract at least one input data item from the plurality of input data items by referring to the plurality of weight values, wherein the at least one input data item is a factor in the inference of the single decision;

[0007] A program according to one aspect of the disclosure causes a computer to operate as: an attention mechanism unit configured to utilize an attention mechanism learning model to calculate a context variable by weighting a plurality of input data items organized as a time series or a plurality of variables calculated from the input data items with a plurality of weight values, and adding the weighted input data items or the weighted variables, wherein the attention mechanism learning model is a learning model of an attention mechanism;a decision unit configured to infer a single decision from a plurality of decisions based on confidence levels of the plurality of decisions calculated from the context variable and a most recent input data item included in the plurality of input data items or a most recent variable included in the plurality of variables; a data extraction unit configured to extract at least one input data item from the plurality of input data items by referring to the plurality of weight values, wherein the at least one input data item is a factor in the inference of the single decision;

[0008] An information processing method according to one aspect of the disclosure includes: using an attention mechanism learning model to calculate a context variable by weighting a plurality of input data items organized as a time series or a plurality of variables calculated from the input data items with a plurality of weight values, and summing the weighted input data items or the weighted variables, wherein the attention mechanism learning model is a learning model of an attention mechanism; inferring a single decision from a plurality of decisions based on confidence levels of the plurality of decisions calculated from the context variable and a most recent input data item included in the plurality of input data items or a most recent variable included in the plurality of variables;and extracting at least one input data item from the plurality of input data items by referring to the plurality of weighting values, wherein the at least one input data item is a factor in the inference of the single decision.; EFFECTS OF THE INVENTION

[0009] Accordingly, at least one aspect of the disclosure may easily localize data that serves as the basis for a decision made by a learning model using an attention mechanism. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a block diagram schematically illustrating a configuration of an information processing device according to a first embodiment. Fig. 2A and Fig. 2B are block diagrams showing hardware configuration examples. Fig. 3 is a schematic diagram for explaining processing executed in the information processing device according to the first embodiment. Fig. 4 is a block diagram schematically illustrating a configuration of an information processing device according to a second embodiment. Fig. 5 is a schematic diagram for explaining processing executed in the information processing device according to the second embodiment. Fig. 6 is a block diagram schematically illustrating a configuration of an information processing device according to a third embodiment. Fig. 7 is a schematic diagram for explaining processing executed in the information processing device according to the third embodiment. MODE FOR CARRYING OUT THE INVENTION FIRST EMBODIMENT

[0010] Fig. 1 is a block diagram schematically illustrating a configuration of an information processing device 100 according to a first embodiment.

[0011] The information processing device 100 comprises a storage unit 101, a communication unit 102, an input unit 103, a display unit 104 and a control unit 110.

[0012] The storage unit 101 stores programs and data required for the processing performed in the information processing device 100.

[0013] For example, the storage unit 101 stores at least one attention mechanism learning model, which is a learning model used in an attention mechanism executed by the control unit 110. In the first embodiment, the storage unit 101 also stores an extractive learning model and a decision learning model, as described later.

[0014] The storage unit 101 also stores information about decision input data indicating input data determined to be important by a result of inference made by the attention mechanism, in other words, input data that is a factor in the decision result.

[0015] The communication unit 102 communicates with other devices. The communication unit 102 communicates with other devices, for example, via a network such as the Internet.

[0016] The input unit 103 accepts an input from a user of the information processing device 100.

[0017] 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.

[0018] The control unit 110 controls processing executed in the information processing device 100. The control unit 110 acquires, for example, input data and calculates state variables required for making a decision from the input data. The control unit 110 calculates a context state variable by performing a weighted sum of the state variables through an attention mechanism and infers a specific decision from the context state variable. The control unit 110 then refers to the weights added by the attention mechanism to extract the input data that represents a factor for the decision result, i.e., the inferred decision, and stores the decision input data information indicating the input data in the storage unit 101.Here it can be determined that the extracted input data had a significant impact on the inference.

[0019] In the following, a state variable is also simply referred to as a variable, and a context state variable is also simply referred to as a context variable.

[0020] The control unit 110 includes a data acquisition unit 111, a variable extraction unit 112, an attention mechanism unit 113, a decision unit 114, an attention timing information extraction unit 115, and a data extraction unit 116.

[0021] The data acquisition unit 111 acquires input data. The data acquisition unit 111 can acquire input data, for example, via the communication unit 102. If input data is stored in the storage unit 101, the data acquisition unit 111 can acquire input data from the storage unit 101. The input data acquired here is assumed to be time series data. The acquired input data is provided to the variable extraction unit 112 and the data extraction unit 116.

[0022] The variable extraction unit 112 extracts state variables, which are variables that can be used to make a decision, from the input data acquired by the data acquisition unit 111.

[0023] Here, the variable extraction unit 112 extracts state variables using an extractive learning model, which is a learning model for extracting state variables from input data. It is assumed that the state variables extracted by the variable extraction unit 112 are organized in the form of a time series.

[0024] The attention mechanism unit 113 calculates a context state variable by performing weight summation using a known attention mechanism on the state variables extracted by the variable extraction unit 112. For example, the attention mechanism unit 113 infers multiple weight values ​​by applying an attention mechanism learning model stored in the storage unit 101 to the state variables extracted by the variable extraction unit 112, adding weights using the weight values, and summing the weighted state variables to calculate a context state variable as an inference result.

[0025] The decision unit 114 infers a single decision from multiple decisions based on the confidence levels of the multiple decisions calculated from the context state variable inferred by the attention mechanism unit 113 and the most recent state variable included in the multiple state variables.

[0026] Here, the decision unit 114 performs inference by using a decision learning model, which is a learning model for inferring a single decision from a context variable.

[0027] The attention timing information extraction unit 115 generates attention timing information indicating a plurality of weight values ​​inferred by the attention mechanism unit 113 and the timings of the input data corresponding to the state variables weighted by the respective weight values, and outputs the attention timing information to the data extraction unit 116.

[0028] The data extraction unit 116 refers to the weight values ​​indicated by the attention timing information to extract one or more input data items that represent a factor for the decision result from the input data. In other words, the data extraction unit 116 extracts input data that is believed to have a significant influence on the decision result and generates decision input data information indicating this input data. The data extraction unit 116 then stores the decision input data information in the storage unit 101.

[0029] Specifically, the data extraction unit 116 may determine that the input data that is a factor for the decision result is formed from two sets of input data: one set is formed from input data items corresponding to weight values ​​indicated in the attention timing information that exceed a predetermined threshold or a first threshold, and the other set is formed from pairs of input data items corresponding to pairs of weight values ​​indicated in the attention timing information that correspond to consecutive points in time and whose magnitude of change exceeds a predetermined threshold or a second threshold. The magnitude of the change may be a difference or a percentage.

[0030] Part or all of the control unit 110 described above may be implemented, for example, by a working memory 10 and a processor 11, such as a central processing unit (CPU), which executes the programs stored in the working memory 10, as shown in Fig. 2A. In other words, the information processing device 100 can be implemented by a well-known computer. Such programs can be provided via a network or can be recorded and provided on a recording medium. That is, such programs can be provided, for example, as a program product.

[0031] A part or the whole of the control unit 110 may also be implemented, for example, by a single circuit, a composite circuit, a program-operated processor, a program-operated parallel processor, a processing circuit 12 such as an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA), as in Fig. 2B shown, be implemented.

[0032] As described above, the control unit 110 may be implemented by a processing circuit.

[0033] The storage unit 101 may be implemented by a memory such as a hard disk drive (HDD) or a solid state drive (SSD).

[0034] The communication unit 102 may be implemented by a communication interface, such as a network interface card (NIC).

[0035] The input unit 103 may be implemented by an input interface such as a keyboard or a mouse.

[0036] The display unit 104 may be implemented by a display.

[0037] Fig. 3 is a schematic diagram for explaining processing executed in the information processing device 100 according to the first embodiment.

[0038] First, the data acquisition unit 111 acquires the input data X t-n , X t-n+1 , X t-1 , Xt (step S10). For the input data X t-n , X t-n+1 , X t-1 , Xt are sensor values, which are observation values, and the time series data tn, t-n+1, ​​t-1, t (where t and n are positive integers). For example, image data can be used as input data.

[0039] The data acquisition unit 111 outputs the acquired input data X t-n , X t-n+1 , Xt-1 , Xt to the variable extraction unit 112 and the data extraction unit 116.

[0040] The variable extraction unit 112 extracts from the input data X t-n , X t-n+1 , X t-1 , Xt state variables S t-n , S t-n+1 , S t-1 and St, which are advantageous variables for the decision unit 114 to make a decision (step S11).

[0041] The variable extraction unit 112 uses the extractive learning model, which is a neural network model stored in the storage unit 101, to extract the state variables S t-n , S t-n+1 , S t-1 and St from the input data X t-n , X t-n+1 , X t-1 , to extract Xt.

[0042] The variable extraction unit 112 outputs the extracted state variables S t-n , S t-n+1 , S t-1and St to the attention mechanism unit 113.

[0043] Here, the variable extraction unit 112 uses the extractive learning model; however, the first embodiment is not limited to such an example as long as the state variables S t-n , S t-n+1 , S t-1 and St can be extracted using a function.

[0044] The attention mechanism unit 113 uses a learning model to generate weighted values ​​for the state variables S t-n , S t-n+1 , S t-1 and St, and calculates a weighted sum to calculate a context state variable (step S12).

[0045] The attention mechanism unit 113 passes the calculated context state variable to the decision unit 114.

[0046] The decision unit 114 makes a decision based on the context state variables and the most recent state variable St (step S13).

[0047] Here, the decision unit 114 uses the decision learning model, which is a neural network model stored in the storage unit 101, to infer a decision from the context state variable and the recent state variable.

[0048] In step S12, the attention timing information extraction unit 115 extracts the weight values ​​inferred by the attention mechanism unit 113 and the timings of the corresponding input data, and generates attention timing information indicating the extracted weight values ​​and timings (step S14). The generated attention timing information is passed to the data extraction unit 116.

[0049] The data extraction unit 116 refers to the attention timing information to extract from the input data input data which is a factor in the decision result of the decision unit 114, and generates decision input data information indicating the input data items (step S15).

[0050] The storage unit 101 then stores the decision input data information generated by the data extraction unit 116 (step S16).

[0051] Accordingly, according to the first embodiment, it is possible to easily locate data serving as the basis for a decision made by a learning model using an attention mechanism. SECOND EMBODIMENT

[0052] Fig. 4 is a block diagram schematically illustrating a configuration of an information processing device 200 according to the second embodiment.

[0053] 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.

[0054] 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 respectively 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.

[0055] The storage unit 201 stores programs and data required for the processing performed in the information processing device 200.

[0056] The storage unit 201 according to the second embodiment stores the same data as in the first embodiment and also stores semantic input data generated by a control unit 210 described below.

[0057] The control unit 210 controls the processing executed in the information processing device 200.

[0058] The control unit 210 according to the second embodiment extracts the meaning of the input data and performs a process of interpreting a decision basis.

[0059] The control unit 210 includes a data acquisition unit 111, a variable extraction unit 112, an attention mechanism unit 113, a decision unit 114, an attention timing information extraction unit 115, a data extraction unit 216, and a data meaning acquisition unit 217.

[0060] The data acquisition unit 111, the variable extraction unit 112, the attention mechanism unit 113, the decision unit 114, and the attention timing information extraction unit 115 of the control unit 210 according to the second embodiment are respectively the same as the data acquisition unit 111, the variable extraction unit 112, the attention mechanism unit 113, the decision unit 114, and the attention timing information extraction unit 115 of the control unit 110 according to the first embodiment.

[0061] The data acquisition unit 111 according to the second embodiment passes the acquired input data to the variable extraction unit 112 and the data meaning acquisition unit 217.

[0062] 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 acquires the meaning of the input data by accepting the input of the meaning of the input data from a user via the input unit 103. For example, when the input data is image data, the meaning of the input data is identification information (for example, the name of a target) for identifying a target such as a person or an object in the image data.

[0063] The data meaning acquisition unit 217 then stores semantic input data obtained by adding the meaning of the input data to the input data in the storage unit 201.

[0064] The data extraction unit 216 interprets a decision basis, which is the basis of a decision result by the decision unit 114, from the meaning content of the input data, which is a factor in the decision result.

[0065] In other words, the data extraction unit 216 refers to attention timing information to extract, from the semantic input data stored in the storage unit 201, semantic input data that is believed to have a significant influence on the decision result. The data extraction unit 216 then interprets the decision basis of the decision result from the meaning content of the extracted semantic input data. Regarding the interpretation of the decision basis, the interpretation method should be determined in advance in accordance with the decision to be made by the decision unit 114 and the weight values.For example, if the weight values ​​indicated by the attention timing information rapidly increase, it can be interpreted that the target that appeared immediately before or after the rapid increase is the decision basis. If a weight value indicated by the attention timing information exceeds a certain threshold, it can be interpreted that a physical quantity such as a distance, a position, or a size of a predetermined target at that time is the decision basis. Here, the predetermined target can be defined in accordance with the decision to be made by the decision unit 114. For example, if the decision to be made by the decision unit 114 is to use an automobile, the predetermined target is an oncoming vehicle, a wall, a pedestrian, or the like.

[0066] The data extraction unit 216 then stores, in the storage unit 201, decision basis information indicating the decision basis, which is the content interpreted from the extracted semantic input data. The decision basis information may include the semantic input data used to interpret the decision basis.

[0067] Fig. 5 is a schematic diagram for explaining processing executed in the information processing device 200 according to the second embodiment.

[0068] The processing of steps S10 to S14 in Fig. 5 is the same as the processing of steps S10 to S14 in Fig. 3.

[0069] 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 semantic input data by adding the meaning to the input data (step S27). The generated semantic input data is stored in the storage unit 201.

[0070] The data extraction unit 216 refers to attention timing information to extract semantic input data that is a factor in the decision result from the semantic input data stored in the storage unit 201, and interprets the decision basis of the decision result from the meaning content of the extracted semantic input data (step S28). The data extraction unit 216 then generates decision basis information indicating the content interpreted from the extracted semantic input data.

[0071] The storage unit 201 then stores the decision basis information (step S29).

[0072] As described above, according to the second embodiment, since decision basis information indicating the content interpreted as a decision is generated, the decision basis can be presented to a user in an easily understandable manner. THIRD EMBODIMENT

[0073] Fig. 6 is a block diagram schematically illustrating a configuration of an information processing device 300 according to the third embodiment.

[0074] 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.

[0075] 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 respectively 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.

[0076] The storage unit 301 stores programs and data required for the processing executed in the information processing device 300.

[0077] The storage unit 301 according to the third embodiment stores the same data as in the first embodiment, and also stores semantic input data and decision rules generated by a control unit 310 described below.

[0078] The control unit 310 controls the processing performed in the information processing device 300.

[0079] As in the second embodiment, the control unit 310 according to the third embodiment extracts the meaning of the input data and performs a process of interpreting a decision basis.

[0080] The control unit 310 according to the third embodiment generates a decision rule indicating a decision basis and a decision result inferred from semantic input data corresponding to the decision basis, and stores the decision rule in the storage unit 301.

[0081] The control unit 310 includes a data acquisition unit 111, a variable extraction unit 112, an attention mechanism unit 113, a decision unit 114, an attention timing information extraction unit 115, a data extraction unit 216, a data meaning acquisition unit 217, and a decision rule generation unit 318.

[0082] The data acquisition unit 111, the variable extraction unit 112, the attention mechanism unit 113, and the decision unit 114 of the control unit 310 according to the third embodiment are respectively the same as the data acquisition unit 111, the variable extraction unit 112, the attention mechanism unit 113, and the decision unit 114 of the control unit 110 according to the first embodiment.

[0083] The data extraction unit 216 and the data meaning acquisition unit 217 of the control unit 310 according to the third embodiment are respectively the same as the data extraction unit 216 and the data meaning acquisition unit 217 of the control unit 210 according to the second embodiment.

[0084] The decision rule generation unit 318 generates a decision rule that correlates a decision basis, which is content extracted from semantic input data by the data extraction unit 216, and an inferred decision result. The decision rule may contain semantic input data used to interpret the decision basis.

[0085] The decision rule generation unit 318 then stores the decision rule in the storage unit 301.

[0086] Fig. 7 is a schematic diagram for explaining processing executed in the information processing device 300 according to the third embodiment.

[0087] The processing of steps S10 to S14 in Fig. 7 is the same as the processing of steps S10 to S14 in Fig. 3.

[0088] As in the second embodiment, in the third embodiment, the data meaning acquisition unit 217 extracts the meaning of input data from the data acquisition unit 111 and generates semantic input data by adding the meaning of the input data to the input data (step S27). The generated semantic input data is stored in the storage unit 301.

[0089] The data extraction unit 216 refers to attention timing information to extract the semantic input data, which is a factor in the decision result, from the semantic input data stored in the storage unit 301, and interprets the decision basis of the decision result from the meaning content of the extracted semantic input data (step S28).

[0090] The decision rule generation unit 318 generates a decision rule that specifies the decision basis interpreted in step S28 and the decision result inferred by the decision unit 114. The decision rule may include the semantic input data extracted in step S27.

[0091] The storage unit 301 then stores the decision rule (step S31).

[0092] As described above, according to the third embodiment, a decision rule is generated that specifies a decision basis and a decision result inferred from the decision basis. By accumulating such decision rules, a rule-based artificial intelligence (AI) model, such as a well-known decision tree, can be generated. Since a decision rule is provided in a format easily understandable to a user, the user can easily understand the content of an inference performed by the information processing device 300.

[0093] In the first to third embodiments described above, state variables are extracted from the input data by the variable extraction unit 112, but the first to third embodiments are not limited to such an example. For example, if the input data is data suitable for processing, the processing by the variable extraction unit 112 cannot be performed.

[0094] In such a case, the attention mechanism unit 113 uses an attention mechanism learning model, which is a learning model of an attention mechanism, calculates a context variable by weighting a plurality of input data items organized as a time series with a plurality of weight values, and adds the input data items. The decision unit 114 infers a decision from a plurality of decisions based on the context variables and the confidence levels of the multiple decisions calculated from the most recent input data included in the plurality of input data items. LIST OF REFERENCE SYMBOLS

[0095] 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 Decision unit; 115 Attention timing information extraction unit; 116, 216 Data extraction unit; 217 Data meaning acquisition unit; 318 Decision rule generation unit.

Claims

[1] Information processing device, comprising: an attention mechanism unit configured to use an attention mechanism learning model to calculate a context variable by weighting a plurality of input data items organized as a time series or a plurality of variables calculated from the input data items with a plurality of weight values, and adding the weighted input data items or the weighted variables, wherein the attention mechanism learning model is a learning model of an attention mechanism; a decision unit configured to infer a single decision from a plurality of decisions based on confidence levels of the plurality of decisions calculated from the context variable and a most recent input data item included in the plurality of input data items or a most recent variable included in the plurality of variables; and a data extraction unit configured to extract at least one input data item from the plurality of input data items by referring to the plurality of weight values, wherein the at least one input data item is a factor in the inference of the individual decision. [2] The information processing device according to claim 1, wherein the data extraction unit is configured to extract an input data item corresponding to a weight value included in the plurality of weight values ​​when the one weight value exceeds a first threshold value, the first threshold value being a predetermined threshold value. [3] The information processing device according to claim 1 or 2, wherein the data extraction unit is configured to extract two input data items corresponding to two weight values ​​included in the plurality of weight values ​​when a magnitude of change between the two weight values ​​corresponding to two consecutive points in time in the time series exceeds a second threshold, the second threshold being a predetermined threshold. [4] Information processing device according to one of claims 1 to 3, further comprising: a data meaning acquisition unit configured to acquire a meaning of the at least one input data element, wherein the data extraction unit is configured to interpret a decision basis on which the one decision is inferred from the meaning of the at least one input data element. [5] Information processing device according to claim 4, further comprising: a decision rule generation unit configured to generate a decision rule that correlates the decision basis and a decision. [6] Information processing device according to claim 5, further comprising: a storage unit configured to store the decision rule. [7] Program that causes a computer to act as: an attention mechanism unit configured to use an attention mechanism learning model to calculate a context variable by weighting a plurality of input data items organized as a time series or a plurality of variables calculated from the input data items with a plurality of weight values, and adding the weighted input data items or the weighted variables, wherein the attention mechanism learning model is a learning model of an attention mechanism; a decision unit configured to infer a single decision from a plurality of decisions based on confidence levels of the plurality of decisions calculated from the context variable and a most recent input data item included in the plurality of input data items or a most recent variable included in the plurality of variables; and a data extraction unit configured to extract at least one input data item from the plurality of input data items by referring to the plurality of weight values, wherein the at least one input data item is a factor in the inference of the individual decision. [8] Information processing methods comprising: Using an attention mechanism learning model to compute a context variable by weighting a plurality of input data items organized as a time series or a plurality of variables computed from the input data items with a plurality of weight values, and summing the weighted input data items or the weighted variables, wherein the attention mechanism learning model is a learning model of an attention mechanism; Inferring a single decision from a plurality of decisions based on confidence levels of the plurality of decisions calculated from the context variable and a most recent input data item included in the plurality of input data items or a most recent variable included in the plurality of variables; and Extracting at least one input data item from the plurality of input data items by referring to the plurality of weight values, wherein the at least one input data item is a factor in the inference of the single decision.

Images