Abstract: An apparatus for monitoring a heartbeat of a fetus comprises a memory comprising instruction data representing a set of instructions, and a processor configured to communicate with the memory and to execute the set of instructions. The set of instructions, when executed by the processor, cause the processor to: i) obtain ultrasound image data from an ultrasound transducer, ii) determine a location of the heart of the fetus in the ultrasound image data, iii) send a message to the ultrasound transducer to instruct the ultrasound transducer to switch to a Doppler mode, wherein the determined location of the heart of the fetus is used to set the target location of the Doppler mode, and iv) monitor the heartbeat of the fetus using ultrasound data obtained in the Doppler mode.

Abstract: A variety of improvements to docking systems for portable computing devices are disclosed. For example, improved techniques for maintaining a data connection between a base portion of the docking system and a case portion of the docking system are disclosed. As an example, the docking system can include improved magnetics that help maintain the data connection between the base portion and the case portion, even during rotational movements of the case portion relative to the base portion. Further still, examples are described where the detection of a docking action between the case portion and the base portion can trigger any of a number of responsive actions.

Abstract: An embodiment of the present invention provides an RF ion guide having four elongated electrodes arranged in parallel around the axial centerline. Each electrode is generally L-shaped in cross section, having first and second inner surfaces directed toward the interior of the ion guide. The first and second surfaces extend along axis that are transverse and preferably approximately perpendicular to one another. RF voltages of equal amplitude but opposite phases are applied to opposed pairs of electrodes, in the manner known in the art, to generate an RF field to radially confine ions and focus them to the centerline. Because the resultant RF field more closely approximates a quadrupolar field, relative to the field generated within a flatapole, better performance may be achieved in terms of improved transmission efficiencies and/or less mass discrimination.

Abstract: According to an aspect of the present invention, a computer-implemented method is provided for reinforcement learning. The method includes reading, by a processor device, an action manifold which is described as a n-polytope, at least one physical action limit, and at least one safety constraint. The method further includes updating, by the processor device, the action manifold based on the at least one physical action limit and the at least one safety constraint. The method also includes performing, by the processor device, the reinforcement learning by selecting a constrained action from among a set of constrained actions in the action manifold.

Abstract: A computer-implemented method, computer program product, and computer processing system are provided for Hierarchical Reinforcement Learning (HRL) with a target task. The method includes obtaining, by a processor device, a sequence of tasks based on hierarchical relations between the tasks, the tasks constituting the target task. The method further includes learning, by a processor device, a sequence of constraints corresponding to the sequence of tasks by repeating, for each of the tasks in the sequence, reinforcement learning and supervised learning with a set of good samples and a set of bad samples and by applying an obtained constraint for a current task to a next task.

Abstract: This disclosure is directed to product merchandising systems that and designed to prevent brute force attempts to steal a product on display. The merchandising systems include security features that enhances the strength of the connection between a puck assembly and a base assembly and between the base assembly and a display surface. The merchandising systems are suited for withstanding brute force pulling attacks on the puck assembly and the base assembly.

Abstract: A computer-implemented method, computer program product, and computer processing system are provided for obtaining a plurality of bad demonstrations. The method includes reading, by a processor device, a protagonist environment. The method further includes training, by the processor device, a plurality of antagonist agents to fail a task by reinforcement learning using the protagonist environment. The method also includes collecting, by the processor device, the plurality of bad demonstrations by playing the trained antagonist agents on the protagonist environment.

Abstract: A method is provided for reinforcement learning. The method includes obtaining, by a processor device, a first set and a second set of state-action tuples. Each of the state-action tuples in the first set represents a respective good demonstration. Each of the state-action tuples in the second set represents a respective bad demonstration. The method further includes training, by the processor device using supervised learning with the first set and the second set, a neural network which takes as input a state to provide an output. The output is parameterized to obtain each of a plurality of real-valued constraint functions used for evaluation of each of a plurality of action constraints. The method also includes training, by the processor device, a policy using reinforcement learning by restricting actions predicted by the policy according to each of the plurality of action constraints with each of the plurality of real-valued constraint functions.

Abstract: A computer-implemented method, computer program product, and system are provided for deep reinforcement learning to control a subject device. The method includes training, by a processor, a neural network to receive state information of a target of the subject device as an input and provide action information for the target as an output. The method further includes inputting, by the processor, current state information of the target into the neural network to obtain current action information for the target. The method also includes correcting, by the processor, the current action information minimally to obtain corrected action information that meets a set of constraints. The method additionally includes performing an action by the subject device based on the corrected action information for the target to obtain a reward from the target.

Abstract: An ultrasound system comprises a probe including an array of CMUT (capacitive micromachined ultrasound transducer) cells. Each cell comprises a substrate carrying a first electrode. The substrate is spatially separated from a flexible membrane including a second electrode. The flexible membrane comprises a mass element in a central region. The system also comprises a voltage supply adapted to, in a transmission mode provide, the respective electrodes with a bias voltage driving the CMUT cells into a collapsed state and a stimulus voltage having a set frequency for resonating the flexible membrane of the CMUT cells in said collapsed state The mass element of the CMUT cells forces the central region of the flexible membrane to remain in the collapsed state during said resonating. A pulse transmission method for such a system is also disclosed.

Abstract: This disclosure is directed to product merchandising systems that and designed to prevent brute force attempts to steal a product on display. The merchandising systems include security features that enhances the strength of the connection between a puck assembly and a base assembly and between the base assembly and a display surface. The merchandising systems are suited for withstanding brute force pulling attacks on the puck assembly and the base assembly.

Abstract: A mid-plane board including a first connector configured to receive a first signal from a first circuit board is provided. The mid-plane board includes a second connector configured to provide the first signal to a second circuit board. The first circuit board forms a first plane and the second circuit board forms a second plane, and the first plane and the second plane are substantially parallel. The mid-plane board also includes a cutout configured to allow a coplanar connector to bridge the mid-plane board and provide a second signal from the first circuit board to the second circuit board. The second signal is a high-end signal and the first signal is a low-end signal, and the mid-plane board is disposed on a plane substantially orthogonal to the first circuit board and the second circuit board.

Abstract: A variety of improvements to docking systems for portable computing devices are disclosed, For example, improved techniques for maintaining a data connection between a base portion of the docking system and a case portion of the docking system are disclosed. As an example, the docking system can include improved magnetics that help maintain the data connection between the base portion and the case portion, even during rotational movements of the case portion relative to the base portion. Further still, examples are described where the detection of a docking action between the case portion and the base portion can trigger any of a number of responsive actions.

Abstract: The present invention is directed at a relatively high modulus spunbond nonwoven material that is suitable for use in relatively high deep draw molding applications.

Abstract: According to an aspect of the present invention, a computer-implemented method is provided for reinforcement learning. The method includes reading, by a processor device, an action manifold which is described as a n-polytope, at least one physical action limit, and at least one safety constraint. The method further includes updating, by the processor device, the action manifold based on the at least one physical action limit and the at least one safety constraint. The method also includes performing, by the processor device, the reinforcement learning by selecting a constrained action from among a set of constrained actions in the action manifold.

Abstract: A computer-implemented method, computer program product, and computer processing system are provided for obtaining a plurality of bad demonstrations. The method includes reading, by a processor device, a protagonist environment. The method further includes training, by the processor device, a plurality of antagonist agents to fail a task by reinforcement learning using the protagonist environment. The method also includes collecting, by the processor device, the plurality of bad demonstrations by playing the trained antagonist agents on the protagonist environment.

Abstract: A computer-implemented method, computer program product, and computer processing system are provided for Hierarchical Reinforcement Learning (HRL) with a target task. The method includes obtaining, by a processor device, a sequence of tasks based on hierarchical relations between the tasks, the tasks constituting the target task. The method further includes learning, by a processor device, a sequence of constraints corresponding to the sequence of tasks by repeating, for each of the tasks in the sequence, reinforcement learning and supervised learning with a set of good samples and a set of bad samples and by applying an obtained constraint for a current task to a next task.

Abstract: A method is provided for reinforcement learning. The method includes obtaining, by a processor device, a first set and a second set of state-action tuples. Each of the state-action tuples in the first set represents a respective good demonstration. Each of the state-action tuples in the second set represents a respective bad demonstration. The method further includes training, by the processor device using supervised learning with the first set and the second set, a neural network which takes as input a state to provide an output. The output is parameterized to obtain each of a plurality of real-valued constraint functions used for evaluation of each of a plurality of action constraints. The method also includes training, by the processor device, a policy using reinforcement learning by restricting actions predicted by the policy according to each of the plurality of action constraints with each of the plurality of real-valued constraint functions.

Abstract: A computer-implemented method, computer program product, and system are provided for deep reinforcement learning to control a subject device. The method includes training, by a processor, a neural network to receive state information of a target of the subject device as an input and provide action information for the target as an output. The method further includes inputting, by the processor, current state information of the target into the neural network to obtain current action information for the target. The method also includes correcting, by the processor, the current action information minimally to obtain corrected action information that meets a set of constraints. The method additionally includes performing an action by the subject device based on the corrected action information for the target to obtain a reward from the target.

Abstract: A tablet case includes a frame, a connector, a swivel, and a hand strap. The frame is configured to retain a tablet computing device and has a front and a rear. The connector is coupled to the frame and protrudes from the rear of the frame. The connector is configured to releasably attach the frame to a docking station. The swivel is rotatable about the connector, and the hand strap is attached to the swivel. The hand strap is moveable between a handheld position and a dock ready position. The hand strap is resiliently biased to the dock ready position. The handheld position is a position in which the hand strap is spaced from the frame such that a user's hand can be inserted, and the dock ready position is a position in which the hand strap does not interfere with docking of the tablet case to a docking station.