Abstract: A method is provided for processing on an acceleration component a deep neural network. The method includes configuring the acceleration component to perform forward propagation and backpropagation stages of the deep neural network. The acceleration component includes an acceleration component die and a memory stack disposed in an integrated circuit package. The memory stack has a memory bandwidth greater than about 50 GB/sec and a power efficiency of greater than about 20 MB/sec/mW.

Abstract: A modular overhead lighting system includes a raceway element and one or more interchangeable lighting modules capable of variable positioning relative to the raceway element. Multiple raceway conductors are mounted in the raceway element, with a portion of each raceway conductor being exposed along a lower surface thereof to facilitate continuously variable positioning of lighting modules. An electrical connector for establishing electrical connections between a raceway element and a lighting module includes multiple ridges incorporating electrical contacts configured to contact raceway conductors, and includes terminals to removably receive electrical wires permitting electrical communication between a lighting module and the raceway element.

Abstract: A bedding block for mounting a barreled action for a rifle to a stock can include an elongated body configured to receive the barreled action for the rifle in a semicircular channel defined in top surface of the elongated body. The semicircular channel can have a first radius, the barreled action can have a second radius, and the first radius can be less than the second radius.

Abstract: Described technologies extend the information available from an execution trace of a program by providing heuristically-derived values for memory contents when the trace does not include data expressly showing the value of a memory cell at a particular execution time. Various heuristics are described. The heuristics may use information about the memory cell at other times to produce the derived value. Some heuristics use other trace data, such as whether the memory cell is in a stack, whether there are gaps in the trace, or whether garbage collection or compilation occurred near the time in question. Grounds for the derived value are reported along with the derived value. A time-travel debugger or other program analysis tool can then present the derived values to users, or make other use of the derived values and grounds to assist debugging and other efforts to improve the functioning of a computing system.

Abstract: An apparatus for cleaning a surface, such as the surface of a rotating disk, the apparatus comprising an articulating arm associated via a linkage to a rotating member driven by a motor. Rotation of the rotating member causes linkage to move the articulating arm in an oscillating pattern. A nozzle associated with the distal end of the articulating arm can convey pressurized air or liquid from a source to a surface in an oscillating pattern based on the construction of the arm, linkage, the rotating member and the speed of rotation.

Abstract: In aspects of variable relinearization in homomorphic encryption, a computing device stores homomorphic encrypted data as a dataset, and implements an encryption application that can perform a multiplication operation on a ciphertext in the homomorphic encrypted data, where the multiplication operation contributes to increase a noise component in the ciphertext. The encryption application can determine a relinearization amount by which to relinearize the ciphertext after the multiplication operation, where the determination is effective to optimize a noise increase in the ciphertext based at least in part on projected subsequent multiplication operations on the ciphertext. The encryption application can then relinearize the ciphertext utilizing the determined relinearization amount that optimizes the noise increase in the ciphertext for optimal relinearization performance.

Abstract: A wrappable textile sleeve for protecting an elongate member and method of construction thereof is provided. The sleeve includes a plurality of warp filaments and at least one weft filament woven with one another to form a woven wall. The woven wall has opposite sides extending lengthwise between opposite ends. The opposite sides are wrappable about a central longitudinal axis into overlapping relation with one another to bound the elongate member within a cavity of the sleeve. At least one warp and/or weft filament is provided as a bulky, fluffy multifilament, wherein the bulky, fluffy multifilament is woven to form at least one float extending over 2 or more weft and/or warp yarns, wherein the at least one float provides loops sufficient to readily fix themselves to hooks of a standard hook and loop fastener.

Abstract: Presented herein are embodiments of state-of-the-art speech recognition systems developed using end-to-end deep learning. In embodiments, the model architecture is significantly simpler than traditional speech systems, which rely on laboriously engineered processing pipelines; these traditional systems also tend to perform poorly when used in noisy environments. In contrast, embodiments of the system do not need hand-designed components to model background noise, reverberation, or speaker variation, but instead directly learn a function that is robust to such effects. A phoneme dictionary, nor even the concept of a “phoneme,” is needed. Embodiments include a well-optimized recurrent neural network (RNN) training system that can use multiple GPUs, as well as a set of novel data synthesis techniques that allows for a large amount of varied data for training to be efficiently obtained.

Abstract: A control valve has a body having an inlet and outlet, a valve seat between the inlet and outlet, a valve plug, and a cage adjacent the valve seat to provide guidance for the valve plug. The valve plug is movable between a closed position, where the valve plug sealingly engages the valve seat, and an open position, where the valve plug is spaced away from the valve seat. The cage has a solid, unitary circumferential wall having inner and outer surfaces and a plurality of passages formed through the wall. Each passage extends between the inner and outer surfaces and follows an arcuate, non-linear path from the inner surface to the outer surface.

Abstract: Machine learning techniques are described for generating recommendations using decision trees. A decision tree is generated based on training data that comprises multiple training instances, each of which comprises a feature value for each of multiple features and a label of a target variable. The multiple features correspond to attributes of multiple content delivery campaigns. Later, feature values of a content delivery campaign are received. The decision tree is traversed using the feature values to generate output. Based on the output, one or more recommendations are identified and the one or more recommendations are presented on a computing device.

Abstract: A system and method for determining a time of an event participant passing a detection point having a participant identifier number by a timing system coupled to an RFID tag reader in proximity to at a monitored point and receiving RFID tag read times and coupled to a portable device with a transceiver and a location detector for selectively receiving a participant number read, the timing system receiving the geographic location of the handheld device, the handheld read participant number and a handheld time, determining that the received RFID tag read times timing system are not usable times and determining a time of passing of the participant by the detection point by determining a delta time between the proximity detector and the handheld device and an adjusted time of participant passing based on the determined delta time and handheld time.

Abstract: Disclosed herein are compositions, probes, devices, and processes useful for detecting specific reactions and binding interactions with biological molecules. In certain embodiments, methods of binding one or more biomolecules to a solid support are disclosed. Methods of generating site-specific sequences for one or more biomolecules from a solid support are also disclosed. Biological complexes generated by these methods are also disclosed.

Abstract: Logic may create new images by adding more than one synthetic image on a template. Logic may provide a template with a background for a new image. Logic may provide a set of models, each model to comprise a generative adversarial network (GAN), the GANs pretrained independently to generate different synthetic images. Logic may select two or more models from the set of models. Logic may generate, by the two or more models, two or more of the different synthetic images. Logic may combine the two or more of the different synthetic images with the template to create the new image. And logic may train a set of GANs independently, to generate one of two or more different synthetic images on a blank image background, the different synthetic images to comprise a subset of the multiple features of a new image.

Abstract: Present systems and methods are directed to an interactive gaming system that includes a first game environment that includes user input devices that receive first user inputs indicative of team/individual performance of a first set of players, and a second game environment that includes user input devices that receive second user inputs indicative of team/individual performance of a second set of players. The interactive gaming system includes a controller that receives the first and second user inputs, accesses dynamic user profiles that correspond the first and the second set of players, updates the dynamic user profiles based on the first and second user inputs, selects a third set of players from the first and second set of players to qualify for a third game environment, and instructs the third game environment based on a subset of the dynamic user profiles that correspond to the third set of players.

Abstract: A customizable adhesive toy playscape is constructed of a combination of printed adhesive playscape tape and other accessories, such as printed stickers, upstanding signs, toy vehicles, and the tape roll core that can be used by children (or adults) for creating imaginary playscape tape worlds for play, education, or other uses. The playscape tape can be part of a single layer track construction that includes machine-readable codes for controlling movement of a mobile agent traveling thereover.

Abstract: A process is presented for the production of butadienes. The process includes the separation of oxygenates from the product stream from an oxidative dehydrogenation reactor. The process includes quenching the product stream and solvent and oxygenates from the product stream. The oxygenates are stripped from the solvent with an inert gas to reduce the energy consumption of the process, and the solvent is recycled and reused in the process.

Abstract: The present invention provides compounds and compositions which are useful in methods for modulating pain and fevers in a subject. These compounds are of the structure of formula (I), wherein Rx-R4, X, Y, and Z are defined herein, or a pharmaceutically acceptable salt, solvate, or prodrug thereof to the subject. In one embodiment, the compound is 5-hydroxy-7-methyloxindole.

Abstract: A hybrid-fabric apparatus comprises a black box memory configured to store a plurality of behavior metrics and an anomaly agent coupled to the black box. The anomaly agent determines a baseline vector corresponding to nominal behavior of the fabric, wherein the baseline vector comprises at least two different behavior metrics that are correlated with each other. The anomaly agent disaggregates anomaly detection criteria into a plurality of anomaly criterion to be distributed among network nodes in the fabric.

Abstract: The principles and embodiments of the invention relate to systems and computer-implemented methods for administering financial product accounts that provide reward benefits. An attribution rate on an investment in a financial account can be generated as benefit rewards received from the use of credit cards in a credit card account. The method can comprise issuing credit cards to customers, and receiving invested funds into corresponding financial accounts for each customer, wherein the financial accounts are configured by the computer-implemented method as a saving vehicle selected by the customers for investment. The computer-implemented administration system can comprise a credit or debit card provided to a person having an account, a first computer system having communication hardware for receiving credit or debit card data over a network, and a non-transitory computer readable media having an account database configured to record monetary values in the account.

Abstract: Embodiments of the present invention relate to integrated modular LED display devices. In one embodiment, a modular LED display devices comprises a plastic housing with an outer surface exposed to an external environment. The modular LED display device is configured to display images using an array of pixels attached to a front side of a printed circuit board attached to the plastic housing. The modular LED display device includes a circuit for controlling a plurality of LEDs, the circuit being attached to the opposite second side of the printed circuit board. The first side of the printed circuit board is sealed to be waterproof by an overlying compound. The modular LED display device further includes a power supply including a power converter for converting alternating current (AC) power to direct current (DC) power. The modular LED display device is configured to be exposed to the external environment without additional enclosures.