Abstract: A computer-implemented system and method for synthesizing a controller for an under actuated robotic manipulator includes a machine learning based model having a plurality of neural network modules. Each module is configured to approximate a function related to an underactuated controller for a robotic manipulator. Parameters of each function are learned during training of the model using a loss function that satisfies one or more conditions including structure preservation, integrability and equilibrium assignment.

Abstract: Techniques for dynamic profile data ingestion are described. The system may process profile data associated with one device, such as a mobile device, to associate it with another device, such as a vehicle. For example, when a connection is made between the first device and the second device, the profile data associated with the first device may be associated with the second device, in a manner using a remote system. The remote system may temporarily associate the profile data with the second device, enabling the system to interpret a command using the profile data.

Abstract: Systems, methods, and computer-readable media are disclosed for generating and training a deep convolutional generative model for multivariate time series modeling and utilizing the model to assess time series data indicative of a machine or machine component's operational state over a period of time to detect and localize potential operational anomalies.

Abstract: System and method for synthesizing a controller for a dynamical system includes a feeder neural network trained to estimate an ordinary differential equation (ODE) from time series training data (X) of a trajectory having embedded angular data and configured to learn dynamics of a physical system by encoding a generalization of a Hamiltonian representation of the dynamics using a constant external control term (u). A neural ODE solver receives the estimate of the ODE from the feeder neural network and synthesizes a controller to control the system to track a reference configuration.

Abstract: Techniques for dynamic profile data ingestion are described. The system may process profile data associated with one device, such as a mobile device, to associate it with another device, such as a vehicle. For example, when a connection is made between the first device and the second device, the profile data associated with the first device may be associated with the second device, in a manner using a remote system. The remote system may temporarily associate the profile data with the second device, enabling the system to interpret a command using the profile data.

Abstract: Techniques for dynamic contact ingestion are described. A system may interpret a voice command received from a first device based on contact data or other information associated with a second device connected to the first device. For example, when a data connection is made between the first device and the second device, the first device may receive the contact data and send the contact data to a remote system. The remote system may temporarily associate the contact data with the first device, enabling the remote system to interpret a voice command received from the first device using the contact data. The remote system may use the contact data to perform disambiguation, enabling the remote system to initiate outbound calls, announce inbound calls, and/or the like. When the second device is disconnected from the first device, the remote system may remove the association between the contact data and the first device.

Abstract: System and method for robust machine learning (ML) includes an attack detector comprising one or more deep neural networks trained using adversarial examples generated from a generative adversarial network (GAN), producing an alertness score based on a likelihood of an input being adversarial. A dynamic ensemble of individually robust ML models of various types and sizes and all being trained to perform an ML-based prediction is dynamically adapted by types and sizes of ML models to be deployed during the inference stage of operation. The adaptive ensemble is responsive to the alertness score received from the attack detector. A data protector module with interpretable neural network models is configured to prescreen training data for the ensemble to detect potential data poisoning or backdoor triggers in initial training data.

Abstract: A method for identifying underperforming agents in a multi-agent cooperative system includes receiving information relating to the performance of each agent in the multi-agent system, calculating an estimated extracted resource value of each agent based on the received information, comparing the estimated extracted resource value of each agent to a threshold value, calculating a performance index based on the comparison and identifying an agent as an under-performing agent based on the performance index.

Abstract: System and method are disclosed for approximating unknown safety constraints during reinforcement learning of an autonomous agent. A controller for directing the autonomous agent includes a reinforcement learning (RL) algorithm configured to define a policy for behavior of the autonomous agent, and a control barrier function (CBF) algorithm configured to calculate a corrected policy that relocates policy states to an edge of a safety region. Iterations of the RL algorithm safely learn an optimal policy where exploration remains within the safety region. CBF algorithm uses standard least squares to derive estimates of coefficients for linear constraints of the safe region. This overcomes inaccurate estimation of safety region constraints caused by one or more noisy observations of constraints received by sensors.

Abstract: System and method for performing natural language processing are disclosed. An encoder includes a multi-head attention block for nonlinear transformation of inputs and a feed-forward network for learning parameters that result in best function approximation. Output of the multi-head attention block and the feed-forward network are coupled in parallel to produce a summed output. An ODE solver performs continuous depth integration of the summed output for reduced number of parameters.

Abstract: A computer-implemented method of scheduling jobs for an industrial process includes receiving jobs to be executed on machines within a manufacturing facility. A job schedule is generated based on an optimization function that minimizes total energy cost for all the machines during a time horizon based on a summation of energy cost at each time step between a start time and an end time. The energy cost at each time step is a summation of (a) a first energy cost associated with each machine in sleeping mode during the time step, (b) a second energy cost associated with each machine in stand-by mode during the time step, and (c) a third energy cost associated with each machine in processing mode during the time step. The jobs are executed on the machines based on the job schedule.

Abstract: System and method for modeling motion and collision of rigid bodies in a dynamic system includes a collision detector that detects active contacts of the rigid bodies. A differentiable contact impulse solver applies constraints on contact forces related to a compression phase, applies coefficient of restitution on contact forces related to a restitution phase, solves for contact forces and velocity impulses associated with the active contacts in the compression phase and the restitution phase, and estimates trajectories of the rigid bodies while optimizing for maximum rate of energy dissipation.

Abstract: A computer-implemented method for performing machine condition monitoring for fault diagnosis includes collecting multivariate time series data from a plurality of sensors in a machine and partitioning the multivariate time series data into a plurality of segment clusters. Each segment cluster corresponds to one of a plurality of class labels related to machine condition monitoring. Next, the segment clusters are clustered into segment cluster prototypes. The segment clusters and the segment cluster prototypes are used to learn a discriminative model that predicts a class label. Then, as new multivariate time series data is collected from the sensors in the machine, the discriminative model may be used to predict a new class label corresponding to segments included in the new multivariate time series data. If the new class label indicates a potential fault in operation of the machine, a notification may be provided to one or more users.

Abstract: A computer-implemented method for assessing material microstructure of a machine component involves obtaining a raw image of a section of the component captured via a microscope. The method further includes pre-processing the raw image to generate a ternary image defined by pixel data including three levels of intensities. The method further includes identifying, from the ternary image, phase boundaries delineating at a phase in a primary constituent material of the component. The method further includes determining a volume associated with the phase based on the identified phase boundaries. The proposed method may be utilized, for example, as an automated tool for assessing material degradation and for quality control of gas turbine engine components.

Abstract: A computer-implemented method for training a deep neural network includes defining a loss function corresponding to the deep neural network, receiving a training dataset comprising training samples, and setting current parameter values to initial parameter values. An optimization method is performed which iteratively minimizes the loss function. During each iteration, a steepest direction of the loss function is calculated by determining the gradient of the loss function at the current parameter values. A batch of samples included in training samples is selected. A matrix-free CG solver is applied to obtain an inexact solution to a linear system defined by the steepest direction of the loss function and a stochastic Hessian matrix with respect to the batch of samples. A descent direction is determined, and the parameter values are updated based on the descent direction. Following the optimization method, the parameter values are stored in relationship to the deep neural network.

Abstract: Systems, techniques, and computer-program products are provided to generate synthetic time series using a generative adversarial network. In some embodiment a technique includes configuring a first neural network having a first function representative of an output of the first neural network, and configuring a second neural network having a second function representative of an output of the second neural network. In addition, such a technique includes generating a generative adversarial network by solving an optimization problem with respect to an objective function based at least on the first function and the second function. The generative adversarial network includes a discriminator neural network and a generator neural network. A synthetic time series can be generated using at least the generator neural network.

Abstract: Systems, techniques, and computer-program products are provided to generate manufacturing schedules that integrate maintenance strategies. A manufacturing schedule can be generated by solving an optimization problem subject to operational constraints that preserve consistency in the order of the operations to be performed during the manufacture of a product, and further subject to maintenance constraints that enforce a desired maintenance strategy. The optimization problem can be solved by minimizing a makespan of a product subject to the operational and maintenance constraints.

Abstract: Techniques for dynamic contact ingestion are described. A system may interpret a voice command received from a first device based on contact data or other information associated with a second device connected to the first device. For example, when a data connection is made between the first device and the second device, the first device may receive the contact data and send the contact data to a remote system. The remote system may temporarily associate the contact data with the first device, enabling the remote system to interpret a voice command received from the first device using the contact data. The remote system may use the contact data to perform disambiguation, enabling the remote system to initiate outbound calls, announce inbound calls, and/or the like. When the second device is disconnected from the first device, the remote system may remove the association between the contact data and the first device.

Abstract: A method for identifying underperforming agents in a multi-agent cooperative system includes receiving information relating to the performance of each agent in the multi-agent system, calculating an estimated extracted resource value of each agent based on the received information, comparing the estimated extracted resource value of each agent to a threshold value, calculating a performance index based on the comparison and identifying an agent as an under-performing agent based on the performance index.

Abstract: A system for verification of the output of a sensor includes an industrial system comprising a plurality of sensors, one of the plurality of sensors being a target sensor, a plurality of machine learning networks, each machine learning network connecting a plurality of driving sensors associated with the target sensor and trained using simulation data. a selected machine learning network from the plurality of machine learning networks having an output representative of the target sensor, the selected machine learning network being trained with real-time data from the industrial plant and a processor for comparing an output of the selected machine learning network to a real output of the target sensor. Based on the comparison, the real sensor output is provided as final output when the values match, and the estimated value is output when the values do not match and the sensor output is flagged as an error.