Abstract: Disclosed examples access data from a database, the data stored across multiple registers of the database; determine (a) a maximum rank for each of the multiple registers and (b) a maximum rank count for each of the multiple registers; determine a frequency distribution based on the maximum ranks and the maximum rank counts; and generate a report including at least one of the frequency distribution, the maximum ranks, or the maximum rank counts.

Abstract: Embodiments of the present disclosure include a system for optimizing an artificial neural network by configuring a model, based on a plurality of training parameters, to execute a training process, monitoring a plurality of statistics produced upon execution of the training process, and adjusting one or more of the training parameters, based on one or more of the statistics, to maintain at least one of the statistics within a predetermined range. In some embodiments, artificial intelligence (AI) processors may execute a training process on a model, the training process having an associated set of training parameters. Execution of the training process may produce a plurality of statistics. Control processor(s) coupled to the AI processor(s) may receive the statistics, and in accordance therewith, adjust one or more of the training parameters to maintain at least one of the statistics within a predetermined range during execution of the training process.

Abstract: A computer-implemented method includes receiving a sequence of acoustic frames as input to an automatic speech recognition (ASR) model. Here, the ASR model includes a causal encoder and a decoder. The method also includes generating, by the causal encoder, a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The method also includes generating, by the decoder, a first probability distribution over possible speech recognition hypotheses. Here, the causal encoder includes a stack of causal encoder layers each including a Recurrent Neural Network (RNN) Attention-Performer module that applies linear attention.

Abstract: A medication packet bundle includes one or more beacons. Each beacon may transmit a signal representing an event related to patient compliance with a medication regimen, such as removal of a packet of medication from a bundle or opening of a packet. The transmitted beacon signals can be monitored and processed to track patient compliance and adherence.

Abstract: Described herein are complete lifecycle management processes for IoT/M2M devices. In an example, devices are commissioned and de-commissioned in a given system without requiring a user/human administrator. A delegated life-cycle management process is described, wherein devices rely upon a delegatee, which may have more computing and battery resources than the devices, to perform complete or partial lifecycle management operations on behalf of the devices. The delegatee may be a trusted entity that may belong to the same domain as the devices. Further, a Trust Enabling Infrastructure (TEI) is described herein, which may belong to a different trusted domain than the given device and its delegatee.

Abstract: A wireless device may receive one or more radio resource control (RRC) messages comprising configuration parameters for one or more unlicensed cells. The wireless device may selectively monitor a downlink physical control channel (PDCCH) on one of the unlicensed cells, depending on whether there is an uplink transmission or downlink transmission on another unlicensed cell.

Abstract: A building electrical system provides electrical mounting plates which are embedded inside of building walls. The electrical mounting plates provide power to low voltage electrical fixtures such as motorized window coverings and allow the fixtures to be mounted and operated without external wiring.

Abstract: Parallel processing concepts such as wavefront parallel processing, are realized with a reduced end-to-end delay by giving up the usual slice concept according to which slices are either coded/decoded completely independent from areas of the picture outside of the respective slice, or at least independent from areas outside the respective slice as far as the entropy coding is concerned, namely in favor of slices of different modes, namely ones called dependent slices which allow for interdependencies across slice boundaries, and others which do not, called normal slices, for example. Combined with the aspect or not, WPP processing concept is made more efficiently by using the slices' start syntax portions to locate WPP entry points.

Abstract: The present disclosure relates to systems and methods implemented on a memory controller for detecting and mitigating memory attacks (e.g., row hammer attacks). For example, a memory controller may track activations of row addresses within a memory hardware (e.g., a DRAM device) and determine whether a pattern of activations is indicative of a row hammer attack. This is determined using a counting mode for corresponding memory sub-banks. Where a likely row hammer attack is detected, the memory controller may activate a sampling mode (rather than the counting mode) for a particular sub-bank to identify which of the row addresses should be refreshed on the memory hardware. The implementations described herein provide a low computational cost alternative to heavy-handed detection mechanisms that require access to significant computing resources to accurately detect and mitigate row hammer attacks.

Abstract: Methods and systems are provided for fetching graphics data for rendering a scene presented on a display device. The method includes receiving gaze information for eyes of a user while the user is interacting with the scene. The method includes tracking gestures of the user while the user is interacting with the scene. The method includes identifying a content item in the scene as being a potential focus of interactivity by the user. The method includes processing the gaze information and the gestures of the user to generate a prediction of interaction with the content item by the user. The method includes processing a pre-fetching operation to access and load the graphics data into system memory in anticipation of the user interacting with the content item. In this way, when a prediction of interaction with a content item in the scene is determined based on the game actions (e.g., eye gaze, gestures, etc.

Abstract: A radio communication route enables communication from an originating ground station to a destination ground station via one of multiple randomly orbiting, rotating satellites with no active attitude control. The ground stations and satellites include directional antennas for receiving radio signals from and transmitting radio signals in multiple directions. The satellites store an address of a destination ground station from which an initial information signal is transmitted and antenna information identifying the satellite antenna on which the initial information signal was received. Plural satellite antennas transmit linking information identifying the satellite to the originating ground station.

Abstract: A snap-in crank mechanism has a crank housing mounted out of a column, and a crank having a crank rod with two ends; the crank housing has a crank inner housing sleeved outside the column, and a crank outer housing detachably sleeved outside the crank inner housing; the crank inner housing has a left inner housing, a right inner housing and a lower inner housing, the crank outer housing is slidable on the periphery of a sleeve formed by the left inner housing and the right inner housing.

Abstract: The present disclosure relates to systems and methods for classifying data sets using associated functions from neural networks. In one example, a system for classifying data sets by corresponding functions includes at least one processor and at least one non-transitory memory storing instructions that, when executed by the at least one processor cause the system to perform operations including: obtaining a neural network associated with a data set, the neural network being trained to generate synthetic data sets related to the data set; selecting a test set of inputs to the neural network; obtaining a corresponding set of outputs by applying the neural network to the test set of inputs; estimating one or more functions describing the test set of inputs and the corresponding set of outputs; and indexing the estimated one or more functions to the data.