An information processing apparatus includes a controller. The controller is configured to obtain a measurement value of a sensor provided in mechanical equipment. The controller is configured to generate a first model by machine learning using the measurement value of the sensor measured in a first period of the mechanical equipment and store the first model in a storage portion. The controller is configured to generate a second model by machine learning using the measurement value of the sensor measured in a second period after a trigger event has occurred in the mechanical equipment and store the second model in the storage portion. The controller is configured to determine a state of the mechanical equipment by using the measurement value of the sensor measured in an evaluation period and the first model and the second model stored in the storage portion.

This learning device provides a learned model to an adjuster containing a learned model learned to output a predetermined compensation amount to a controller, in a control system including the controller outputting a command value obtained by compensating a target value based on a compensation amount and a control object controlled to process an object to be processed. The learning device includes: an evaluation part obtaining operation data including the target value, command value and control variable and evaluates the quality of the control variable; a learning part generating candidate compensation amounts based on the operation data, and learning, as teacher data, the generated candidate compensation amount and the specific parameter of the object, and generating a learned model; and a setting part providing the learned model to the adjuster if the evaluated quality is within an allowable.

Interactions between a training server and a plurality of environment controllers are used for updating the weights of a predictive model used by a neural network executed by the plurality of environment controllers. Each environment controller executes the neural network using a current version of the predictive model to generate outputs based on inputs, modifies the outputs, and generates metrics representative of the effectiveness of the modified outputs for controlling the environment. The training server collects the inputs, the corresponding modified outputs, and the corresponding metrics from the plurality of environment controllers. The collected inputs, modified outputs and metrics are used by the training server for updating the weights of the current predictive model through reinforcement learning. A new predictive model comprising the updated weights is transmitted to the environment controllers to be used in place of the current predictive model.

The step for performing machine learning includes acquiring shape data; acquiring geometric information for each of a plurality of machining faces; acquiring a tool path pattern selected for the machining faces from among a plurality of tool path patterns; and performing machine learning by using the geometric data for known workpieces and the tool path patterns wherein the input is the geometric information for the machining faces and the output is the tool path pattern for the machining faces. The step for generating a new tool path includes: acquiring shape data for the workpiece; acquiring geometric information for each of the plurality of machining faces of the workpiece to be machined; and generating a tool path pattern for each of the plurality of machining faces on the workpiece on the basis of the results of the machine learning using the geometric information of the workpiece to be machined.

An artificial intelligence (AI) device capable of being controlled according to a user's gaze includes a communication unit, a camera configured to capture an image of a user, and a processor configured to acquire user state information from the image of the user, acquire a gaze position of the user based on the acquired user state information, calculate a distance between the acquired gaze position and the camera, receive, from one or more external AI devices, one or more distances between gaze positions of the user respectively acquired by the external AI devices and cameras respectively provided in the external AI devices through the communication unit, and compare the calculated distance with the received one or more distances to select a controlled device.

A sensorless collision detection method of robotic arm based on motor current includes acquiring an output current of a robotic arm joint motor; building a neural network, and using a backpropagation algorithm to update the weights and the deviations of the neural network to obtain an estimated current value; judging whether collision occurs by comparing the collision detection threshold with the error value between the output current of the robotic arm joint motor and the estimated output current of the neural network. The detection method is easy to operate and has higher universality.

Methods, apparatus, systems, and articles of manufacture are disclosed that optimize layers of a machine learning model for a target hardware platform. An example apparatus includes a communication processor to obtain information specific to the target hardware platform (THP) on which to execute the machine learning model; a layer generation controller to generate layers of the machine learning model based on the information specific to the THP; and a deployment controller to, in response to the machine learning model satisfying a threshold error metric, deploy the machine learning model to the THP.

Described herein is a method and system for controlling an interior-mounted permanent magnet (IPM) alternating-current (AC) electrical machine utilizing a space vector pulse-width modulated (SVPWM) converter operably connected between an electrical power source and the IPM AC electrical machine comprising three neural networks (NNs), including a controller NN operably connected to the SVPWM converter, a parameter estimator NN, and a flux-weakening and MTPA NN.

A search range of machine learning is changed and relearned when the search range is not appropriate. A machine learning device (200) that searches for a first parameter of a component of a servo control device (100) that controls a servo motor (300) includes: a search solution detection unit (2024A) that acquires a set of evaluation function values used for machine learning device during machine learning or after machine learning, plots the set of evaluation function values in a search range of the first parameter or a second parameter used for searching for the first parameter, and detects whether a search solution is at an edge of the search range or is in a predetermined range from the edge; an evaluation function expression estimation unit (2024B) that estimates an evaluation function expression on the basis of the set of evaluation function values when the search solution is at the edge of the search range or is in the predetermined range; and a search range changing unit (2024C) that changes the search range to a new search range of the first parameter or the second parameter on the basis of the estimated evaluation function expression.

An apparatus is provided which includes a processing circuit and a plurality of sensors connected to a vehicle, where at least one of the plurality of sensors is positioned on an undercarriage of the vehicle. The plurality of sensors can detect variations in a road on which the vehicle is traveling. The plurality of sensors can also generate information corresponding to the variations of the road. The plurality of sensors can also transmit the information corresponding to the variations in the road to the processing circuit. The information collected by the plurality of sensors may then be used to augment a driving capability of the vehicle.