Abstract: An information processing apparatus (2000) acquires a feature vector (20) obtained based on signal data (14) of a detected value of a sensor (10) that senses gas to be measured. The information processing apparatus (2000) decomposes the feature vector (20) into a product of a coefficient vector and a feature matrix by using a non-negative matrix factorization (NMF). The detected value of the sensor (10) changes according to an attachment and a detachment of a molecule contained in a sensed gas. A value of each element of the feature vector (20) is equal to or greater than zero.

Abstract: A learning apparatus acquires learning data in which odor data of each object and a label representing the object in a label space expressing features of odors are associated with each other, and learns, based on odor data, a model for predicting a label of the odor data in the label space, by using the learning data. In a data processing apparatus for processing odor data, an acquisition unit acquires odor data from an outside. A prediction unit predicts a label of the acquired odor data in the label space by using the model in which a relationship between sets of odor data and labels in the label space expressing the features of the odors is learned.

Abstract: A reward function estimation unit 81 estimates a reward function by multiple importance sampling using samples of a decision-making history of a subject and of a decision-making history generated based on a sampling policy. A policy estimation unit 82 estimates a policy by reinforcement learning using the estimated reward function. The reward function estimation unit 81 sets the policy estimated by the policy estimation unit as a new sampling policy, and estimates the reward function by the multiple importance sampling using the samples of the decision-making history of the subject and of the decision-making history generated based on the sampling policy.

Abstract: The intention feature extraction device 80 includes an input unit 81, a learning unit 82, and a feature extraction unit 83. The input unit 81 receives input of a decision-making history of a subject. The learning unit 82 learns an objective function in which factors of an intended behavior of the subject are explanatory variables, based on the decision-making history. The feature extraction unit 83 extracts weights of the explanatory variables of the learned objective function as features which represent intention of the subject.

Abstract: To mitigate degradation in the accuracy of a prediction model by re-learning the prediction model with consideration given to the characteristics of a detection value of a sensor. This prediction model re-learning device comprises: a calculation unit that, on the basis of data related to smell detection by a sensor, calculates an index for determining whether or not to re-learn a prediction model for smell; and a re-learning unit that re-learns the prediction model in cases where the calculated index satisfies a predetermined condition.

Abstract: A model estimation system estimates a model of a system represented by an ordinary differential equation with all coefficients being non-zero, and with which input data and a state at each time can be obtained. When an order of the ordinary differential equation and input data and a state at multiple past times in the system are inputted, a model expression construction unit constructs an expression representing a model by using a first matrix that is a matrix according to the order and has only some elements as unknown elements and a second matrix that is a matrix according to the order and has only some one element as an unknown element. A model estimation unit uses input data and a state at multiple past times, to estimate the model by learning unknown elements of the first matrix and the unknown element of the second matrix.

Abstract: An expert model unit 81 generates predicted expert control actions based on an expert model which is a machine learning model trained using data collected when an expert operated a plant which is a control target or a plant of the same or similar characteristics. A transformer 82 constructs metrics or error measures involving the predicted expert control actions from the expert model unit 81 as an objective term. A combiner 83 collects different objective terms from the transformer 82 and a learner which outputs machine-learning models as objective terms and computes an optimal set of weights or combinations of the objective terms to construct an aggregated cost function for use in an optimizer.

Abstract: A vehicle control system for controlling driving of a vehicle reflecting an environment and a characteristic of a user, while suppressing increase in learning time, is provided. The vehicle control system includes classification means for classifying, by using one or more attributes selected from accumulation means for accumulating data including attributes relating to driving of a vehicle, driving properties included in the data, learning means for learning a model representing the driving property, for each of types that are a result of classification by the classification means, and control information determination means for determining, by using the model learned for the type associated with a value of the attribute at time of driving of a control target vehicle, control information for the driving.

Abstract: An information processing apparatus (20) includes a model generating unit (210) and a feature value computation unit (220). The model generating unit (210) generates an Auto-Regressive with eXogenous input (ARX) model of a smell sensor by use of input data controlling an input operation of gas including a smell component being a measurement target, and output data acquired by inputting the gas to the smell sensor, based on the input data. The feature value computation unit (220) computes a transfer function of the smell sensor relating to the smell component by subjecting the ARX model to Z-Transform, and further computes a first-order lag transfer function feature value of the smell sensor relating to the smell component by subjecting the transfer function to partial fraction decomposition.

Abstract: A selection unit (11) in a learning device (10) inputs a plurality of “learning candidate data units.” The plurality of learning candidate data units are respectively related to a plurality of subjects including a plurality of cancer patients and a plurality of non-cancer patients. Further, each learning candidate data unit at least includes a “urine odor data unit” and a “cancer label.” Then, from the plurality of input learning candidate data units, the selection unit (11) selects part of the plurality of learning candidate data units as a “learning target data set,” based on a “selection rule.” By using the learning target data set selected by the selection unit (11), a determination model formation unit (12) forms a “determination model” for determining which of urine of a cancer patient and urine of a non-cancer patient a determination target urine odor data unit is related to.

Abstract: A server apparatus 10 is communicably connected to a terminal apparatus 20 that collects sensor data from an odor sensor 40. The server apparatus 10 includes an analyzer holding unit 11 that holds a plurality of analyzers for analyzing specific odor analysis targets, based on sensor data, an analyzer management unit 12 that selects an analyzer, determines preprocessing to be performed on the sensor data, according to the selected analyzer, and causes the terminal apparatus 20 to execute the preprocessing, an analysis execution unit 13 that executes analysis processing of the designated odor analysis target, by applying the selected analyzer to the preprocessed sensor data from the terminal apparatus, and an analysis result transmission unit 14 that transmits information indicating a result of the analysis processing to the terminal apparatus 20.

Abstract: A processing apparatus (20) includes a prediction equation generation unit (210) and an output unit (250). The prediction equation generation unit (210) generates, through machine learning having a plurality of feature values based on outputs from a set of a plurality of kinds of sensors and correct answer data as inputs, a prediction equation that has the plurality of feature values as variables and is used for predicting an odor component. The output unit (250) outputs a plurality of weights as information indicating the prediction equation in association with the feature values, respectively.

Abstract: An information processing apparatus (2000) acquires a feature vector (20) obtained based on signal data (14) of a detected value of a sensor (10) that senses gas to be measured. The information processing apparatus (2000) decomposes the feature vector (20) into a product of a coefficient vector and a feature matrix by using a non-negative matrix factorization (NMF). The detected value of the sensor (10) changes according to an attachment and a detachment of a molecule contained in a sensed gas. A value of each element of the feature vector (20) is equal to or greater than zero.

Abstract: An information processing apparatus (20) includes a sensor output data acquisition unit (210), a prediction equation generation unit (220), and an operation setting unit (230). The sensor output data acquisition unit (210) acquires sensor output data for each sampling length of an odor sensor with respect to a target gas. The prediction equation generation unit (220) generates, by using the sensor output data for each sampling length, a prediction equation for making a prediction for an odor component of the target gas. The operation setting unit (230) determines, by using the prediction equation, a sampling length for operating the odor sensor.

Abstract: An information processing apparatus (20) includes a use environment information acquisition unit (210), a model selection unit (220), and a prediction unit (230). The use environment information acquisition unit (210) acquires use environment information indicating a use environment of a physical system having input-output. The model selection unit (220) selects, from a storage unit storing a plurality of prediction models of the physical system in association with section information indicating a section based on the use environment, a prediction model being associated with section information of a section matching the use environment indicated by the use environment information. The prediction unit (230) performs prediction based on output of the physical system by use of the selected prediction model.

Abstract: A terminal apparatus 20 includes a sensor data collection unit 21 that collects sensor data from an odor sensor 40 that outputs the sensor data in reaction to a plurality of types of odors, an analyzer acquisition unit 22 that, in the case where an analyzer capable of analyzing a designated odor analysis target is transmitted thereto from a server apparatus 10 that holds a plurality of analyzers for analyzing odor analysis targets by analyzing the sensor data, acquires the analyzer transmitted thereto, an analysis execution unit 23 that executes analysis processing of the designated odor analysis target, by applying the acquired analyzer to the collected sensor data, and an analysis result holding unit 24 that holds information indicating a result of the analysis processing.

Abstract: An information processing apparatus (2000) acquires time-series data (14) output by a sensor (10) and computes a plurality of feature constants ?i and a contribution value ?i representing contribution with respect to the time-series data (14) for each feature constant ?i. Thereafter, the information processing apparatus (2000) outputs information in which a set ? of the feature constants ?i and a set ? of the contribution values ?i are associated with each other as a feature value of a target gas. As the feature constant ?, a velocity constant ? or a time constant ? that is a reciprocal of the velocity constant can be adopted.

Abstract: An information processing apparatus (2000) acquires time-series data (14) output by a sensor (10) and computes a contribution value ?i representing contribution with respect to the time-series data (14) for each of a plurality of feature constants ?i. Thereafter, the information processing apparatus (2000) outputs a set ? of the contribution values ?i as a feature value of a target gas. As the feature constant ?, a velocity constant ? or a time constant ? that is a reciprocal of the velocity constant can be adopted.

Abstract: A learning model generation support apparatus 10 is an apparatus for supporting generation of a learning model to be utilized in odor detection using an odor sensor that reacts to a plurality of types of odors. The learning model generation support apparatus 10 includes a data acquisition unit 11 that acquires sensor data output by the odor sensor under specific measurement conditions and condition data specifying the measurement conditions, and inputs, as training data, the acquired sensor data and condition data to a machine learning engine 31 that generates the learning model, and a condition setting unit 12 that acquires a predictive accuracy output by the machine learning engine in response to input of the training data, and sets new measurement conditions for when the odor sensor newly outputs sensor data as training data, based on the acquired predictive accuracy.

Abstract: An input unit 81 inputs action data, in which a state of an environment and an action performed under the environment are associated with each other, a prediction model for predicting a state according to the action on the basis of the action data, and explanatory variables of objective functions for evaluating the state and the action together. A structure setting unit 82 sets a branch structure in which the objective functions are placed at lowermost nodes of a hierarchical mixtures of experts model. A learning unit 83 learns the objective functions including the explanatory variables and branching conditions at nodes of the hierarchical mixtures of experts model, on the basis of the states predicted with the prediction model applied to the action data divided in accordance with the branch structure.