A method for detecting fallen cargo, the method may include receiving by a computerized system, sensed information related to driving sessions of multiple vehicles; applying a machine learning process on the sensed information to detect fallen cargo and to classify the fallen cargo to fallen cargo classes; estimating, from the sensed information, an impact of at least some of the fallen cargo classes on a behavior of at least some of the multiple vehicles; and determining, based on the impact, at least one suggested vehicle behavior as a response to a detection of at least some of the fallen cargo classes.

A computer-implemented method for the self-calibration of an inertial sensor. The method includes: establishing data that relate to the inertial sensor; subdividing the data into training data and test data; setting a first target accuracy value for a first artificial neural network that includes linear activation functions; training the first artificial neural network using the training data; inputting the test data into the trained first artificial neural network in order to obtain a first output value of the first artificial neural network;

establishing a first output accuracy value based on a comparison result between the first output value and the test data; storing weightings and the linear activation functions of the first artificial neural network in a memory unit of the inertial sensor if the first output accuracy value is greater than the first target accuracy value, or otherwise, training the first artificial neural network again using the training data.

A computer-implemented learning method for optimizing a control policy controlling a system is provided. The method includes receiving states of the system being operated for a specific task, initializing the control policy as a function approximator including neural networks, collecting state transition and reward data using a current control policy, estimating an advantage function and a state visitation frequency based on the current control policy, updating the current control policy using the second-order approximation of the objective function, a second-order approximation of the KL-divergence constraint on the permissible change in the policy using a quasi-newton trust region policy optimization, and determining an optimal control policy, for controlling the system, based on the average reward accumulated using the updated current control policy.

The present disclosure provides for customizable content moderation using neural networks with fine-grained and dynamic image classification ontology. A content moderation system of the present disclosure may provide a plurality of image categories in which a subset of of image categories may be designated as restricted categories. The restricted categories may be chosen by a content provider or an end-user. The content moderation system may utilize a neural network to classify image data (e.g., still images, video, etc.) into one or more of the plurality of image categories, and determine that an image is a restricted image upon classifying the image into one of the restricted categories. The restricted image may by flagged, rejected, removed, or otherwise filtered upon being classified as a restricted image.

Data indicative of a plurality of observations of an environment are received at a control system. Machine learning using deep reinforcement learning is applied to determine an action based on the observations. The deep reinforcement learning applies a convolutional neural network or a deep auto encoder to the observations and applies a training set to locate one or more regions having a higher reward. The action is applied to the environment. A reward token indicative of alignment between the action and a desired result is received. A policy parameter of the control system is updated based on the reward token. The updated policy parameter is applied to determine a subsequent action responsive to a subsequent observation.

A control system for controlling an energy storage system includes a controller including a plurality of layered nodes configured to form an artificial neural network trained to generate a forecasted transmission level load and confidence value for an entire jurisdiction of a utility distribution system. The controller includes at least one memory and at least one processor configured for: identifying a potential coincident peak for the utility distribution system based on the forecasted transmission level load and confidence value generated by the artificial neural network; and upon identifying a potential coincident peak, transmitting signals to cause the electrical infrastructure to consume energy stored at the energy storage system thereby reducing the energy drawn from the utility distribution system during the identified potential coincident peak.

An automated unloader system and method. A process performed by an automatic unloader system includes performing an automatic unloading operation of parcels from a container. The process includes monitoring the automatic unloading operation using a plurality of sensors. The process includes automatically detecting an exception based on current sensor data and a knowledge base storing past sensor data. The process includes automatically resolving the exception using the knowledgebase and using resolution knowledge generated from a previous manual resolution by an operator of a corresponding type of exception.

A system is described for configuring another system, e.g., a robotics system. The other system interacts with an environment according to a deterministic policy by repeatedly obtaining, from a sensor, sensor data indicative of a state of the environment, determining a current action, and providing, to an actuator, actuator data causing the actuator to effect the current action in the environment. To configure the other system, the system optimizes a loss function based on an accumulated reward distribution with respect to a set of parameters of the policy. The accumulated reward distribution includes an action probability of an action of a previous interaction log being performed according to the current set of parameters. The action probability is approximated using a probability distribution defined by an action selected by the deterministic policy according to the current set of parameters.

In order to control a technical system by means of control model a data container is received, in which data container a control model having a training structure and model type information are encoded over all the model types. One of multiple model-typespecific execution modules is selected for the technical system as a function of the model type information. Furthermore, operating data channels of the technical system are assigned input channels of the control model as a function of the model type information. Operating data of the technical system are acquired via a respective operating data channel and are transferred to the control model via an input channel assigned to this operating data channel. The control model is executed by means of the selected execution module, wherein control data are derived from the transferred operating data according to the training structure and are output to control the technical system.

Systems and methods for establishing relationships between building automation system components and controlling building automation system components. Data for a building automation system components may be received from the building automation system components and one or more models may be applied to the received data to determine types of the building automation system components and relationships between building automation system components. Once the types of building automation system components have been determined or identified, uniform names may be applied to the building automation system components. The received data may include, among other data, naming data and telemetry data from the building automation system components.