Abstract: A method includes receiving a training example for a listen-attend-spell (LAS) decoder of a two-pass streaming neural network model and determining whether the training example corresponds to a supervised audio-text pair or an unpaired text sequence. When the training example corresponds to an unpaired text sequence, the method also includes determining a cross entropy loss based on a log probability associated with a context vector of the training example. The method also includes updating the LAS decoder and the context vector based on the determined cross entropy loss.

Abstract: A system and method are provided for inter-ceding video in which encoder and decoder memory requirements associated with storage of motion information related to collocated coding units is reduced. In some embodiments motion information related to only a single collocated coding unit may be stored at the encoder and decoder. In operation, if the encoder determines that motion information for a current coding unit should replace the currently stored motion information for the currently stored motion information for the collocated coding unit, then the encoder can replace the motion information at the encoder and transmit an indicator with the current coding unit to signal to the decoder that the motion information currently stored should be updated or replaced with the motion information associated with the current coding unit.

Abstract: Systems, methods and computer-readable media are provided for determining optimum placement, within a grow space, of plants of different varieties. Sets of reference environmental setpoints each correspond to a desired predicted performance of a respective plant variety. Environmental condition metrics at different positions within the grow space are determined. An allocation of the plants is determined among different volumetric regions within the grow space based on predicted performance of the plant varieties as a function of position within the grow space.

Abstract: A method executed by a computing entity includes obtaining a transaction token representing a first pending transaction associated with a transaction item and establishing a set of required public benefit asset information for subsequent acquisition of a set of public benefit credit assets based on the first pending transaction. The method further includes identifying, using a securely passing process, a secure first token on an object distributed ledger representing the set of public benefit credit assets based on the set of required public benefit asset information. The method further includes establishing, in accordance with the securely passing process, control over the secure first token representing the set of public benefit credit assets to produce an updated secure first token, where only a device possessing control over the secure first token may modify the secure first token.

Abstract: A method for video decoding includes determining, based on prediction information, whether the current block uses a flexible partition type. Reconstructed samples of the reference block are stored in a cache memory, which stores reconstructed samples of coding regions of a first size comprising a coding tree unit (CTU) of the current block and reconstructed samples of at least one coding region of the first size comprising a CTU adjacent to the CTU of the current block. The cache memory stores either (i) only reconstructed samples of the coding regions of the CTU of the current block or (ii) reconstructed samples of coding regions of a second size of the CTU of the current block and reconstructed samples of at least one coding region of the second size of the CTU adjacent to the CTU of the current block.

Abstract: An intelligent lock, overlock, and lock system can be electrically controlled to open and close one or more locks. The lock, lock system, and method may function on a frequency selected to avoid cross talk, which permits numerous locks to function simultaneously on the same central system with little interruption or delay. A lock may be used as an overlock for an existing lock or be a primary lock and may include a pivoting engagement structure that engages a locking mechanism to provide added strength.

Abstract: Systems and methods for enhancing a headset user’s own voice include at least two outside microphones, an inside microphone, audio input components operable to receive and process the microphone signals, a voice activity detector operable to detect speech presence and absence in the received and/or processed signals, and a cross-over module configured to generate an enhanced voice signal. The audio processing components includes a low frequency branch comprising low pass filter banks, a low frequency spatial filter, a low frequency spectral filter and an equalizer, and a high frequency branch comprising highpass filter banks, a high frequency spatial filter, and a high frequency spectral filter.

Abstract: An electrical contact, a 3D LED and a method of manufacturing 3D LEDs are described. The electrical contact includes a printed circuit board (PCB) bridge and a PCB connector. The PCB bridge comprises at least one electrical contact. The at least one electrical contact is configured for electrical coupling with at least one interposer of a string of LEDs. The PCB connector is electrically coupled to the PCB bridge. The PCB connector comprises an electrical contact configured for electrical coupling with at least one external electrical wire.

Abstract: An inductively heatable susceptor for use with an inductively heated aerosol-generating device or system includes an open-porous inductively heatable ceramic material configured to hold an aerosol-forming liquid and configured to heat the aerosol-forming liquid under the influence of an alternating electromagnetic field. A cartridge for use with an aerosol-generating device includes an aerosol-forming liquid and a susceptor. An aerosol-generating device includes a susceptor.

Abstract: An electronic display system includes a display controller configured to receive a first hinge angle between a first display and a second display pivotally coupled to the first display via a hinge. A first portion of the rendered visual content is displayed on the first display and a second portion of the rendered visual content is displayed on the second display, the first portion and the second portion having rendered sizes based at least in part on the first hinge angle.

Abstract: Embodiments disclosed include a data center comprising a first reconfigurable utility module comprising interfaces for facilitating electrical power supply, data connectivity, and cooling media supply, and further comprising a plurality of interfaces through which components may be connected to said first reconfigurable utility module. Additionally, a plurality of server modules are connected to the reconfigurable utility module through the interfaces, each of said server modules comprising one or more racks of servers, each of said server modules receiving said power supply, said data connectivity, and said cooling media through an interface through which it is connected to said reconfigurable utility module.

Abstract: A method includes obtaining a graph that includes a plurality of nodes each representing a computational job and connected to one or more other nodes via edges. Each edge represents a dependency between two connected nodes. For each computing device of a distributed computing system, the method includes obtaining resource characteristics. The method includes slicing, based on the dependencies of connected nodes, the graph into a plurality of sub-graphs each including one or more nodes. For each respective sub-graph, the method includes determining a computational cost of the respective sub-graph and distributing, based on the computational cost and the resource characteristics, the respective sub-graph to the respective one of the computing devices. The method includes receiving a respective result from the respective one of the computing devices based on execution of the computational jobs. The method includes determining a total result based on each respective result.

Abstract: Examples of the present disclosure describe systems and methods for probabilistic neural network architecture generation. In an example, an underlying distribution over neural network architectures based on various parameters is sampled using probabilistic modeling. Training data is evaluated in order to iteratively update the underlying distribution, thereby generating a probability distribution over the neural network architectures. The distribution is iteratively trained until the parameters associated with the neural network architecture converge. Once it is determined that the parameters have converged, the resulting probability distribution may be used to generate a resulting neural network architecture. As a result, intermediate architectures need not be fully trained, which dramatically reduces memory usage and/or processing time.

Abstract: A control module adapted to be attached to a power adapter is described.

Abstract: Described is a system for conveying wet sand. The system includes containers with a slip coating for storing wet sand. Each container includes an outlet formed in a bottom portion to facilitate removal of the wet sand from the container to a hopper. Each hopper has end walls, side walls, and an interior wall extending at an angle from the top to the bottom of the hopper to reduce an area of the outlet. A flow gate over the outlet controls discharge of the wet sand from the hopper onto a conveyor belt of a conveyor system. The hopper also includes a vibration system coupled to its exterior surface for shaking the hopper to assist in releasing sand. A discharge assembly is located at the end of the conveyor system for directing wet sand from the conveyor belt.

Abstract: An in situ generation of a biocide, such as monochlorodimethylhydantoin (“MCDMH”) and/or dichlorodimethylhydantoin (“DCDMH”), suitable as a preservative for paints is provided along with a dried surface coating having a biocidal surface comprising MCDMH and/or DCDMH. Also, the present invention provides for a dried coating have a surface comprising an inert compound, such as dimethylhydantoin (“DMH”), whose biocidal capabilities may be activated by applying a chlorine treatment to the dried surface. Further, this invention provides for liquids that can be used to create the above-described coatings.

Abstract: A solids dissociation apparatus (SDA) may include a housing. SDA may also include at least one insert that is operably engaged with the housing and adapted to receive a continuous fluid stream. SDA may also include a transducer that is operably engaged with the housing and disposed about the at least one insert. The transducer is configured to create cavitation inside of the housing, via sonic waves, to eviscerate contaminants in the continuous fluid stream. SDA may also include at least one pair of electrodes that is positioned inside of the at least one insert. The at least one pair of electrodes is configured to provide electrolysis of the continuous fluid stream flowing through the at least one insert to produce at least one continuous stream of oxygen fuel and at least one continuous stream of hydrogen fuel from the continuous fluid stream flowing through the at least one insert.

Abstract: The present invention discloses access devices and methods to create an access channel for introduction of one or more working devices into an anatomical region. The access channel is created using a visualization modality that is later removed before inserting one or more working devices through the access channel. This allows the methods and devices of the present invention to be used even in small sized natural or surgically created insertion tracts leading to the anatomical region. The access channel can be made of a device such as a sheath, a guidewire, and an elongate device comprising a lumen. Examples of visualization modalities are endoscopes and body insertable ultrasound imaging devices. The working devices can be used to perform a variety of diagnostic, therapeutic, or preventive procedures. Endometrial ablation devices and procedures have been used as an example to describe various aspects of the present invention.

Abstract: Methods and systems for well tool wear classification system are provided. A wear classifier tool is configured to classify wear of a scanned well tool using a machine learning engine. Computer-readable memory stores a training dataset and a trained ML model. The training data set includes scanned image data and associated labels representative of classification types of failure. The trained ML model has a neural network. The wear classifier tool can output data identifying a failure mode of the scanned well tool based on classification of input by the machine learning engine. A database is configured to stored historical data on scanner type, patterns of scanner cutting elements, sensor type, and age and usage conditions. A scanning system includes a camera and a three-dimensional (3D) scanner configured to scan a drill bit.

Abstract: A pressure-energized bidirectional seal having a main body having a first end and a second end and a pair of legs extending from the second end of the main body, the pair of legs having an inner surface and an outer surface, the inner surface forming a hollow interior opening in a first direction. The bidirectional seal includes at least one rib extending from the outer surfaces and configured to sealingly engage a mating surface of a mating body the at least one rib forming a sealing zone opening in a second direction opposite of the first direction.