Abstract: In various examples, the present disclosure relates to using temporal filters for automated real-time classification. The technology described herein improves the performance of a multiclass classifier that may be used to classify a temporal sequence of input signals—such as input signals representative of video frames. A performance improvement may be achieved, at least in part, by applying a temporal filter to an output of the multiclass classifier. For example, the temporal filter may leverage classifications associated with preceding input signals to improve the final classification given to a subsequent signal. In some embodiments, the temporal filter may also use data from a confusion matrix to correct for the probable occurrence of certain types of classification errors. The temporal filter may be a linear filter, a nonlinear filter, an adaptive filter, and/or a statistical filter.

Abstract: The present disclosure is directed to a method and system for increasing virtual machine (VM) density on a server system through adaptive rendering by dynamically shifting video rendering tasks to a client computing device. In one embodiment, a processor in a server manages virtual machines in the server by controlling a number of VMs and an amount of system resources allocated to the VMs. The number of VMs and the amount of resources allocated to the VMs are controlled by shifting video rendering from at least one of the VMs to a client device, and increasing the number of the VMs in the server after the shifting.

Abstract: Memory, used by a computer to store data, is generally prone to faults, including permanent faults (i.e. relating to a lifetime of the memory hardware), and also transient faults (i.e. relating to some external cause) which are otherwise known as soft errors. Since soft errors can change the state of the data in the memory and thus cause errors in applications reading and processing the data, there is a desire to characterize the degree of vulnerability of the memory to soft errors. In particular, once the vulnerability for a particular memory to soft errors has been characterized, cost/reliability trade-offs can be determined, or soft error detection mechanisms (e.g. parity) may be selectively employed for the memory. In some cases, memory faults can be diagnosed by redundant execution and a diagnostic coverage may be determined.

Abstract: The present invention provides methods, systems, computer program products that use deep learning with neural networks to denoise ATAC-seq datasets. The methods, systems, and programs provide for increased efficiency, accuracy, and speed in identifying genomic sites of chromatin accessibility in a wide range of tissue and cell types.

Abstract: A simultaneous bi-directional (SBD) transceiver includes a main transmit driver, a replica transmit driver, and a series-series-bridged (SSB) tri-impedance network. A pre-driver stage includes parallel delay paths for the main transmit driver and the replica transmit driver, enabling the delay for signals received by the main transmit driver and the replica transmit driver to be independently configured.

Abstract: An end-to-end system for data generation, map creation using the generated data, and localization to the created map is disclosed. Mapstreams—or streams of sensor data, perception outputs from deep neural networks (DNNs), and/or relative trajectory data—corresponding to any number of drives by any number of vehicles may be generated and uploaded to the cloud. The mapstreams may be used to generate map data—and ultimately a fused high definition (HD) map—that represents data generated over a plurality of drives. When localizing to the fused HD map, individual localization results may be generated based on comparisons of real-time data from a sensor modality to map data corresponding to the same sensor modality. This process may be repeated for any number of sensor modalities and the results may be fused together to determine a final fused localization result.

Abstract: Training deep neural networks requires a large amount of labeled training data. Conventionally, labeled training data is generated by gathering real images that are manually labelled which is very time-consuming. Instead of manually labelling a training dataset, domain randomization technique is used generate training data that is automatically labeled. The generated training data may be used to train neural networks for object detection and segmentation (labelling) tasks. In an embodiment, the generated training data includes synthetic input images generated by rendering three-dimensional (3D) objects of interest in a 3D scene. In an embodiment, the generated training data includes synthetic input images generated by rendering 3D objects of interest on a 2D background image. The 3D objects of interest are objects that a neural network is trained to detect and/or label.

Abstract: In various examples, potentially highlight-worthy video clips are identified from a gameplay session that a gamer might then selectively share or store for later viewing. The video clips may be identified in an unsupervised manner based on analyzing game data for durations of predicted interest. A classification model may be trained in an unsupervised manner to classify those video clips without requiring manual labeling of game-specific image or audio data. The gamer can select the video clips as highlights (e.g., to share on social media, store in a highlight reel, etc.). The classification model may be updated and improved based on new video clips, such as by creating new video-clip classes.

Abstract: A graph neural network to predict net parasitics and device parameters by transforming circuit schematics into heterogeneous graphs and performing predictions on the graphs. The system may achieve an improved prediction rate and reduce simulation errors.

Abstract: An apparatus includes at least one heat pipe that is adapted to be thermally coupled to an integrated circuit and has an evaporator portion and a first condenser portion, wherein the first condenser portion extends away from the evaporator portion; a first plurality of cooling fins that is attached to the first condenser portion; a first movable support that is thermally coupled to the first condenser portion and is configured to move a second plurality of cooling fins relative to the first plurality of cooling fins; and the second plurality of cooling fins, which is attached to the first movable support.

Abstract: One embodiment sets forth a technique for generating a tree structure within a computer memory for storing sparse data. The technique includes dividing a matrix into a first plurality of equally sized regions. The technique also includes dividing at least one region in the first plurality of regions into a second plurality of regions, where the second plurality of regions includes a first region and one or more second regions that have a substantially equal number of nonzero matrix values and are formed within the first region. The technique further includes creating the tree structure within the computer memory by generating a first plurality of nodes representing the first plurality of regions, generating a second plurality of nodes representing the second plurality of regions, and grouping, under a first node representing the first region, one or more second nodes representing the one or more second regions.

Abstract: A network device including a first data structure storing a set of buffer profile types. Each buffer profile type is associated with one or more configuration parameters. The network device further includes a second data structure storing a set of peer device identifiers, wherein each peer device identifier of the set of peer device identifiers is associated with a buffer profile type. The network device includes a buffer management application to receive first data associated with a first peer network device coupled via a first link to an interface component of the network device, determine the first data matches a first peer device identifier stored in the second data structure, and assign a first buffer profile type to the interface component of the network device, wherein the first buffer profile type is associated with the first peer device identifier in the second data structure.

Abstract: The various embodiments of the present disclosure are directed towards methods for tone mapping High-Dynamic-Range (HDR) image data, as well as controlling the brightness of the image encoded by HDR the image data and/or the tone-mapped image data. HDR image is captured. A tone mapping function for the HDR image data is generated. To generate the tone mapping function, control points are dynamically determined based on an analysis of the HDR image data. The tone mapping function is fit to the control points. The tone mapping function is a non-linear function, and is described by a curve in a plane. The shape of the curve is constrained by a line generated from a portion of the control points. The tone mapping function is applied to the HDR image data. A color-compression is applied to the tone mapped image data to generate Standard Dynamic Range or Low Dynamic Range image data.

Abstract: In at least one embodiment, a system determines a set of possible grasp poses that allow a robot to successfully grasp an object by generating a set of potential grasp poses, and then evaluating the performance of each potential grasp pose. In at least one embodiment, the system performs a refinement operation on the grasp poses, and based on an evaluation of the poses, creates an improved set of possible grasps for the object.

Abstract: In various examples, a deep neural network (DNN) is trained—using image data alone—to accurately predict distances to objects, obstacles, and/or a detected free-space boundary. The DNN may be trained with ground truth data that is generated using sensor data representative of motion of an ego-vehicle and/or sensor data from any number of depth predicting sensors—such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. The DNN may be trained using two or more loss functions each corresponding to a particular portion of the environment that depth is predicted for, such that—in deployment—more accurate depth estimates for objects, obstacles, and/or the detected free-space boundary are computed by the DNN. In some embodiments, a sampling algorithm may be used to sample depth values corresponding to an input resolution of the DNN from a predicted depth map of the DNN at an output resolution of the DNN.

Abstract: In various examples, an adaptive eye tracking machine learning model engine (“adaptive-model engine”) for an eye tracking system is described. The adaptive-model engine may include an eye tracking or gaze tracking development pipeline (“adaptive-model training pipeline”) that supports collecting data, training, optimizing, and deploying an adaptive eye tracking model that is a customized eye tracking model based on a set of features of an identified deployment environment. The adaptive-model engine supports ensembling the adaptive eye tracking model that may be trained on gaze vector estimation in surround environments and ensemble based on a plurality of eye tracking variant models and a plurality of facial landmark neural network metrics.

Abstract: Systems and methods for generating real-time synthetic crowd responses for events, to augment the experience of event participants, remote viewers, and the like. Various sensors monitor the event in question, and various event properties are derived from their output using an event state model. These event properties, along with various event parameters such as score, time remaining, etc., are then input to a machine learning model that determines a real-time synthetic audience reaction tailored to the immediate state of the event. Reaction parameters are used to generate a corresponding crowd or audience audio signal, which may be broadcast to event participants, viewers, spectators, or anyone who may be interested. This instantaneous, realistic crowd reaction more closely simulates the experience of events with full on-site audiences, enhancing the viewing experience of both event participants and those watching.

Abstract: Apparatuses, systems, and techniques to cool computer processors. In at least one embodiment, a system comprises one or more processors and a heatsink connected by a flexible heat conduit to the one or more processors, and a position of the heatsink is adjustable.

Abstract: In various examples, a VPU and associated components may be optimized to improve VPU performance and throughput. For example, the VPU may include a min/max collector, automatic store predication functionality, a SIMD data path organization that allows for inter-lane sharing, a transposed load/store with stride parameter functionality, a load with permute and zero insertion functionality, hardware, logic, and memory layout functionality to allow for two point and two by two point lookups, and per memory bank load caching capabilities. In addition, decoupled accelerators may be used to offload VPU processing tasks to increase throughput and performance, and a hardware sequencer may be included in a DMA system to reduce programming complexity of the VPU and the DMA system. The DMA and VPU may execute a VPU configuration mode that allows the VPU and DMA to operate without a processing controller for performing dynamic region based data movement operations.

Abstract: A three-dimensional (3D) object reconstruction neural network system learns to predict a 3D shape representation of an object from a video that includes the object. The 3D reconstruction technique may be used for content creation, such as generation of 3D characters for games, movies, and 3D printing. When 3D characters are generated from video, the content may also include motion of the character, as predicted based on the video. The 3D object construction technique exploits temporal consistency to reconstruct a dynamic 3D representation of the object from an unlabeled video. Specifically, an object in a video has a consistent shape and consistent texture across multiple frames. Texture, base shape, and part correspondence invariance constraints may be applied to fine-tune the neural network system. The reconstruction technique generalizes well—particularly for non-rigid objects.