Abstract: A data storage device and method for auto-peeling of surveillance video content to increase archival storage is provided. In one embodiment, a data storage device is provided comprising a memory and a controller. The controller is configured to determine that available storage space in the memory is less than a threshold; in response to determining that the available storage space in the memory is less than the threshold: read a video file from the memory; and re-encode the video file to decrease a size of the video file, wherein re-encoding the video file increases available storage space in the memory without deleting the video file. Other embodiments are provided.

Abstract: A system configured to perform microphone occlusion event detection. When a device detects a microphone occlusion event, the device will modify audio processing performed prior to speech processing, such as by disabling spatial processing and only processing audio data from a single microphone. The device detects the microphone occlusion event by determining inter-level difference (ILD) values between two microphone signals and using the ILD values as input features to a classifier. For example, when a far-end reference signal is inactive, the classifier may process a first ILD value within a high frequency band. However, when the far-end reference signal is active, the classifier may process the first ILD value and a second ILD value within a low frequency band.

Abstract: Images of a hand may be used to identify users. Quality, detail, and so forth of these images may vary. An image is processed to determine a first spatial mask. A first neural network comprising many layers uses the first spatial mask at a first layer and a second spatial mask at a second layer to process images and produce an embedding vector representative of features in the image. The first spatial mask provides information about particular portions of the input image, and is determined by processing the image with an algorithm such as an orientation certainty level (OCL) algorithm. The second spatial mask is determined using unsupervised training and represents weights of particular portions of the input image as represented at the second layer. The use of the masks allows the first neural network to learn to use or disregard particular portions of the image, improving overall accuracy.

Abstract: An online meeting is established, via a meeting server, where the online meeting facilitates exchange of content between a plurality of separate meeting endpoints engaged in the online meeting. Video content is received at the meeting server from one or more of the meeting endpoints, and presentation content is received at the meeting server from a meeting endpoint associated with a presenter attendee that selects sharing of the presentation content within the online meeting. At the meeting server, an overlay window in the presentation content is identified, video content received at the meeting server from at least one meeting endpoint is inserted within the overlay window in the presentation content to form presentation content integrated with video content, and meeting endpoints engaged in the online meeting are provided for display the presentation content integrated with video content.

Abstract: Apparatuses and methods can be related to supplementing AI processing in memory. An accelerator and/or a host can perform AI processing. Some of the operations comprising the AI processing can be performed by a memory device instead of by an accelerator and/or a host. The memory device can perform AI processing in conjunction with the host and/or accelerator.

Abstract: Embodiments of the invention include a plasma system. The plasma system includes a plasma chamber; an RF driver configured to drive bursts into the plasma chamber with an RF frequency; a nanosecond pulser configured to drive pulses into the plasma chamber with a pulse repetition frequency, the pulse repetition frequency being less than the RF frequency; a high pass filter disposed between the RF driver and the plasma chamber; and a low pass filter disposed between the nanosecond pulser and the plasma chamber.

Abstract: A method and system for activity classification. A pressure sensor receives input data resulting from physical activity of a subject performing an activity. The input data includes pressure data from at least one pressure sensor, and may include other data acquired through other types of sensors. A deep learning neural network is applied to the input data for identifying the activity. The neural network is trained with reference to training data from a training database. The training data may include empirical data from a database of previous data of corresponding activities, synthesized data prepared from the empirical data or simulated data. The training data may include data from physical activity of the subject being monitored by the system. Different aspects of the neural network may be trained with reference to the training data, and some aspects may be locked or opened depending on the application and the circumstances.

Abstract: A pipe fitting having a first body and a second body that together at least partially define a fluid flow passage. The first body defines a first portion of the fluid flow passage that extends from a first end of the fluid flow passage to a first internal opening. The second body defines a second portion of the fluid flow passage that extends from a second internal opening to a second end of the fluid flow passage. The first body has a first interface surface that surrounds the first internal opening, the first interface surface having a plurality of anti-rotation grooves. The second body has a second interface surface that surrounds the second internal opening and engages with the first interface surface. The first internal opening is in fluid communication with the second internal opening. The second interface surface has a plurality of anti-rotation fingers that are each received by and engage with a corresponding one of the anti-rotation grooves.

Abstract: In one embodiment, an IoT server includes: processing circuitry, an I/O module operative to communicate with at least an IoT device and a vendor network server, and an onboarding application and operative to at least: receive an onboarding request from the IoT device via the I/O module, send a confirmation request to the vendor network server via the I/O module, where the confirmation request indicates a request to confirm an identity of the IoT device according to a connection to a network device authenticated by the vendor network server, receive a confirmation response from the vendor network server via the I/O module, where the confirmation response indicates whether the IoT device is connected to the network device, and if the confirmation response is a positive confirmation response that indicates that the IoT device is connected to the network device, onboard the IoT device for participation in an IoT-based system.

Abstract: A computer system includes physical memory devices of different types that store randomly-accessible data in a main memory of the computer system. In one approach, an operating system allocates memory from a namespace for use by an application. The namespace is a logical reference to physical memory devices in which physical addresses are defined. The namespace is bound to a memory type. In response to binding the namespace to the memory type, the operating system adjusts a page table to map a logical memory address in the namespace to a memory device of the memory type.

Abstract: An exercise machine resistance adjustment system for providing an exerciser the ability to immediately adjust the resistance force by changing their body position on the movable carriage or end platforms. The exercise machine resistance adjustment system generally includes a frame, a carriage movably positioned on the frame, a spring connected to the carriage to apply a biasing force to the carriage, one or more left projections within the carriage adapted for a left hand of an exerciser to grasp, and one or more right projections within the carriage adapted for a right hand of an exerciser to grasp. The exerciser is able to efficiently adjust the amount of resistance force applied to the carriage by repositioning their hands (or feet) from a first set of projections at a first distance to a second set of projections at a second distance from the first end of the exercise machine.

Abstract: Techniques are described for automatically generating a consistent configuration state version 2 for a network device with no or minimal help from a user and/or from a provider of the network device when updating from a configuration state version 1 to the configuration state version 2. The techniques and architecture also provide for migration from configuration state version 1 to configuration state version 2 when at least some of a configuration state are located in text files that are applied to the network device at start-up of the network device.

Abstract: Disclosed are systems and method for training an ensemble of machine learning models with a focus on feature engineering. For example, the training of the models encourages each machine learning model of the ensemble to rely on a different set of input features from the training data samples used to train the machine learning models of the ensemble. However, instead of telling each model explicitly which features to learn, in accordance with the disclosed implementations, ML models of the ensemble may be trained sequentially, with each new model trained to disregard input features learned by previously trained ML models of the ensemble and learn based on other features included in the training data samples.

Abstract: Techniques for custom time series analysis with large-scale datasets are described. A time series data analysis service provides an interactive development environment that is configured to accept user input to configure stages of a time series analysis data pipeline. The stages include one or more of a collection stage to place events from a dataset into groupings of ones of the events, a summary stage to generate a set of summary statistics based on the groupings of events, a fill and filter stage to add or remove summary statistics to or from the set of summary statistics, and/or an analytics stage to apply analytical functions based at least in part on the set of summary statistics. The stages can be executed at least partially in a distributed manner by a cluster of computing instances executing an analytics engine.

Abstract: A system for detecting salt ion concentration, comprising a device further comprising a sensor having a carbon printed electrode on a flexible substrate with adhesive on one side (backside) and the circuit electronics to generate pulse signal stimuli and measure salt concentrations to determine hydration level of a person or a living being, wherein the device is a wearable device. The electrode can be made into different shapes changing the area as necessary, since it is a carbon printed electrode and is flexible.

Abstract: Networked sleep mode management is provided by measuring network conditions for a first Access Point (AP) serving a plurality of Client Devices (CDs) configured to operate in one of a sleep mode and an active mode; in response to detecting, based on the measured network conditions, an amount of network usage devoted to transitioning members of the plurality of CDs from the sleep mode to the active mode satisfies a threshold: identifying a first subset of CDs from the plurality of CDs that are deprioritized for access to the sleep mode; receiving a sleep request from a given CD that is a member of the first subset of CDs; and denying the sleep request to force the given CD to maintain the active mode for at least a predefined amount of time.

Abstract: A method and system are provided for determining a relation between stimulation settings for a brain stimulation probe and a corresponding V-field. The brain stimulation probe comprises multiple stimulation electrodes. The V-field is an electrical field in brain tissue surrounding the stimulation electrodes.

Abstract: Disclosed are methods and apparatuses for image data encoding/decoding. A method for decoding a 360-degree image includes the steps of: receiving a bitstream obtained by encoding a 360-degree image; generating a prediction image by making reference to syntax information obtained from the received bitstream; adding the generated prediction image to a residual image obtained by dequantizing and inverse-transforming the bitstream, so as to obtain a decoded image; and reconstructing the decoded image into a 360-degree image according to a projection format. Therefore, the performance of image data compression can be improved.

Abstract: In one embodiment, a method in a multi-tenant wireless network comprises determining a first user private network (UPN) for a first device of a first user. The first UPN provides discovery, by the first device, of other devices on the wireless network to a first subset of other devices on the wireless network. The method further comprises determining a second UPN for the first device of the first user. The second UPN provides discovery, by the first device, of other devices on the wireless network to a second subset of other devices on the wireless network. The method further comprises providing discovery of the first subset and second subset of other devices on the wireless network to the first device of the first user. Discovery of the second subset is provided dynamically based on a current location of the first device.

Abstract: Operation of a head-mounted wearable device is responsive to whether a state of the device is being worn (donned) or not worn (doffed) on the head. The device may operate in a first mode while donned and a second mode while doffed. A capacitive sensor device having an electrode positioned in a bridge of the device provides output data. The output data is processed to determine a baseline value. Later, acquired output data is compared to the baseline value to determine don/doff data indicative of whether the device is donned or doffed. Data from other sensors, may also be used in conjunction with the output data to improve accuracy of the don/doff data. For example, if the output data is above the baseline value and accelerometer data is indicative of motion greater than a motion threshold, the don/doff data may be designated as donned.