A data processing apparatus including a waveform data acquisition unit which acquires waveform data of a consumption current and/or a voltage of a target device, a feature value extraction unit which extracts a waveform feature value from the waveform data, an environment data acquisition unit which acquires environment data indicating an environment of the target device at the time when the waveform data is acquired, an operation state data acquisition unit which acquires operation state data indicating an operation state of the target device at the time the waveform data is acquired, a distance calculation unit which calculates a distance between each of members including the waveform feature value, the environment data, and the operation state data, and each of a plurality of reference members, a grouping unit which groups the members, and a registration unit which registers a group satisfying a predetermined condition as training data.

Methods and apparatus related to generating a model for an object encountered by a robot in its environment, where the object is one that the robot is unable to recognize utilizing existing models associated with the robot. The model is generated based on vision sensor data that captures the object from multiple vantages and that is captured by a vision sensor associated with the robot, such as a vision sensor coupled to the robot. The model may be provided for use by the robot in detecting the object and/or for use in estimating the pose of the object.

A machine learning device performs reinforcement learning on a controller that performs multiple processes for controlling a machine tool in parallel at multiple operation units. The machine learning device comprises: behavior information output means that outputs behavior information containing allocation of arithmetic units that perform the multiple processes to the controller; state information acquisition means that acquires state information containing a machining condition as a condition for machining set at the machine tool, and determination information generated by monitoring the implementation of the multiple processes by the multiple operation units based on the allocation in the behavior information; reward calculation means that calculates the value of a reward to be given by the reinforcement learning based on the determination information in the state information; and value function update means that updates a behavior value function based on the reward value, the state information, and the behavior information.

Methods and apparatus related to generating a model for an object encountered by a robot in its environment, where the object is one that the robot is unable to recognize utilizing existing models associated with the robot. The model is generated based on vision sensor data that captures the object from multiple vantages and that is captured by a vision sensor associated with the robot, such as a vision sensor coupled to the robot. The model may be provided for use by the robot in detecting the object and/or for use in estimating the pose of the object.

Techniques described herein relate to generating a posteriori knowledge about where objects are typically located within environments to improve object location. In various implementations, output from vision sensor(s) of a robot may include visual frame(s) that capture at least a portion of an environment in which a robot operates/will operate. The visual frame(s) may be applied as input across a machine learning model to generate output that identifies potential location(s) of an object of interest. The robot's position/pose may be altered based on the output to relocate one or more of the vision sensors. One or more subsequent visual frames that capture at least a not-previously-captured portion of the environment may be applied as input across the machine learning model to generate subsequent output identifying the object of interest. The robot may perform task(s) that relate to the object of interest.

Techniques described herein relate to generating a posteriori knowledge about where objects are typically located within environments to improve object location. In various implementations, output from vision sensor(s) of a robot may include visual frame(s) that capture at least a portion of an environment in which a robot operates/will operate. The visual frame(s) may be applied as input across a machine learning model to generate output that identifies potential location(s) of an object of interest. The robot's position/pose may be altered based on the output to relocate one or more of the vision sensors. One or more subsequent visual frames that capture at least a not-previously-captured portion of the environment may be applied as input across the machine learning model to generate subsequent output identifying the object of interest. The robot may perform task(s) that relate to the object of interest.

To perform reinforcement learning that enables selecting action information for shortening a cycle time while also avoiding the occurrence of overheating. An action information learning device (300) includes: a state information acquisition means (310) for acquiring state information including an operation pattern of a spindle and a combination of parameters related to machining of a machine tool (100); an action information output means (320) for outputting action information including adjustment information for the operation pattern and the combination of parameters included in the state information; a reward calculation means (333) for acquiring judgment information which is information for temperature of the machine tool (100) and a machining time related to the machining of the machine tool (100), and calculating a value of a reward for reinforcement learning based on the judgment information thus acquired; and a value function update means (332) for updating a value function by performing the reinforcement learning based on the value of the reward, the state information and the action information.

A numerical controller has a machine learning device that performs machine learning of the adjustment of a setting value used in override control. The machine learning device acquires state data showing states of the numerical controller and a machine, sets reward conditions, calculates a reward based on the state data and the reward conditions, performs the machine learning of the adjustment of the setting value used in override control, and determines the adjustment of the setting value used in override control, based on a machine learning result and the state data.

Techniques described herein relate to generating a posteriori knowledge about where objects are typically located within environments to improve object location. In various implementations, output from vision sensor(s) of a robot may include visual frame(s) that capture at least a portion of an environment in which a robot operates/will operate. The visual frame(s) may be applied as input across a machine learning model to generate output that identifies potential location(s) of an object of interest. The robot's position/pose may be altered based on the output to relocate one or more of the vision sensors. One or more subsequent visual frames that capture at least a not-previously-captured portion of the environment may be applied as input across the machine learning model to generate subsequent output identifying the object of interest. The robot may perform task(s) that relate to the object of interest.

A data processing apparatus including a waveform data acquisition unit which acquires waveform data of a consumption current and/or a voltage of a target device, a feature value extraction unit which extracts a waveform feature value from the waveform data, an environment data acquisition unit which acquires environment data indicating an environment of the target device at the time when the waveform data is acquired, an operation state data acquisition unit which acquires operation state data indicating an operation state of the target device at the time the waveform data is acquired, a distance calculation unit which calculates a distance between each of members including the waveform feature value, the environment data, and the operation state data, and each of a plurality of reference members, a grouping unit which groups the members, and a registration unit which registers a group satisfying a predetermined condition as training data.