Abstract: Generative adversarial models have several benefits; however, due to mode collapse, these generators face a quality-diversity trade-off (i.e., the generator models sacrifice generation diversity for increased generation quality). Presented herein are embodiments that improve the performance of adversarial content generation by decelerating mode collapse. In one or more embodiments, a cooperative training paradigm is employed where a second model is cooperatively trained with the generator and helps efficiently shape the data distribution of the generator against mode collapse. Moreover, embodiments of a meta learning mechanism may be used, where the cooperative update to the generator serves as a high-level meta task and which helps ensures the generator parameters after the adversarial update stay resistant against mode collapse. In experiments, tested employments demonstrated efficient slowdown of mode collapse for the adversarial text generators.

Abstract: Presented are systems and methods to automatically find efficient parallelization strategies for deep neural networks (DNNs). A computation graph comprising an efficiently ordered sequence of vertices aids in computing the best parallelizing strategy in a relatively short time. Effectiveness of the parallelization strategies is evaluated on various DNNs, and the performance of the strategies proposed by various embodiments is compared against data parallelism, expert-designed strategies, and other state-of-the-art approaches. Experimental results demonstrate that the proposed strategies outperform a baseline data parallelism strategy and achieve better performance than expert-designed strategies and state-of-the-art approaches.

Abstract: A dynamically adjustable cooling system is discussed and a system configuration design method is disclosed for cooling airflow management and dispatching. The system includes a first and second IT rooms that include electronic racks with IT components. The system also includes cooling units configured to provide cooling air to the IT components. The system also includes a cooling air supply channel to distribute the cooling air to the IT components. The system also includes multiple louvers installed at different locations such as on supply ports of the cooling units, within the cooling air supply channel, and on entrance ports to the IT rooms which are used for configuring the airflow within the supply channel. Each louver is capable of being opened and closed independently to create multiple different air flow combinations for cooling the IT components.

Abstract: In one embodiment, an image sensor captures an image and passes the image to an image processing system that detects one or more vehicle objects and one or more human objects, using a single detection function having a single artificial intelligence (AI) model. Each of the detected objects correspond to a portion of the captured image. For each vehicle object, the corresponding image portion is processed to determine a plurality of properties of the vehicle, including, e.g., type (truck, van, sedan, motorcycle, etc.), color, and license number. For each human object, the corresponding portion of the image is processed to determine a plurality of properties of the human, e.g., gender, approximate age, skin tone, eye color, hair color, and facial landmarks for facial recognition. The captured image is annotated to describe the objects and their properties, and annotated image is transmitted to a service for use by an application.

Abstract: In one embodiment, an emergency vehicle detection system can be provided in the ADV travelling on a road to detect the presence of an emergency vehicle in a surrounding environment of the ADV using both audio data and visual data. The emergency vehicle detection system can use a trained neutral network to independently generate a detection result from the audio data, and use another trained network to independently generate another detection result from the visual data. The emergency vehicle detection system can fuse the two detection results to determine the position and moving direction of the emergency vehicle. The ADV can take appropriate actions in response to the position and moving direction of the emergency vehicle.

Abstract: The present disclosure provides methods and techniques for evaluating and improving algorithms for autonomous driving planning and control (PNC), using one or more metrics (e.g., similarity scores) computed based on expert demonstrations. For example, the one or more metrics allow for improving PNC based on human, as opposed to or in addition to optimizing certain oversimplified properties, such as the least distance or time, as an objective. When driving in certain scenarios, such as taking a turn, people may drive in a distributed probability pattern instead of in a uniform line (e.g., different speeds and different curvatures at the same corner). As such, there can be more than one “correct” control trajectory for an autonomous vehicle to perform in the same turn. Safety, comfort, speeds, and other criteria may lead to different preferences and judgment as to how well the controlled trajectory has been computed.

Abstract: A driving environment is perceived based on sensor data obtained from a plurality of sensors mounted on the ADV, including detecting a traffic light, where the plurality of sensors includes at least one image sensor. A first sensor setting is applied to the at least one image sensor to capture a first frame, and a second sensor setting is applied to the at least one image sensor to capture a second frame. A color of the traffic light is determined based on sensor data of the at least one image sensor in the first frame. The ADV is controlled to drive autonomously according to the color of the traffic light determined based on sensor data of the at least one image sensor in the first frame and a driving environment perceived based on sensor data of the at least one image sensor in the second frame.

Abstract: Methods, systems, and devices for electronic rack level testing and/or verification are disclosed. The disclosed systems may provide for automated testing of electronic racks that may include fluid recirculation systems. To provide for automated testing, a test unit of a system may automatically and without manual intervention circulate coolant to an electronic rack under test. The testing unit may also pressurize the electronic rack. The testing unit may further depressurize the electronic rack to provide for efficient delivering, deployment commissioning, and reduce the likelihood of coolant circulation issues occurring when initially deployed. A testing server of the system may also provide for load testing and thermal cycling of the electronic rack in an automated manner.

Abstract: A liquid cooling distribution system can be installed to an information technology (IT) rack to deliver and distribute fluid to IT equipment. The liquid cooling system can include a fluid manifold and a container that is arranged to capture a fluid that leaks from the fluid manifold. A first fluid sensor can be arranged to detect the fluid at a first position in the container. A controller can be configured to reduce a flow of the fluid into the fluid manifold and pump out fluid from the fluid manifold, in response to the fluid in the container being detected at the first position. Further remedial measure can be taken based on various detected leak scenarios.

Abstract: In one embodiment, a system determines activation parameters for an autonomous driving vehicle (ADV), where the activation parameters include historical usages of a primary brake system or a secondary brake system. In response to determining that a brake is to be applied, the system determines whether to activate a primary or a secondary brake system based on the activation parameters. The system sends an activation flag to activate the primary or the secondary brake system based on the determining whether to activate the primary or the secondary brake system. The system sends a brake command to the primary and the secondary brake system to activate either the primary or the secondary brake system according to the activation flag.

Abstract: Various embodiments for tensor decomposition to compress neural network models are presented. In one or more embodiments, one or more neural network layers from a neural network model are compressed using tensor ring (TR) decomposition. In one or more embodiments, a TR-decomposed neural network utilizes fewer resources and can be more readily deployed on resource-constraint devices. With a smaller model size, during inference time, the compressed model runs faster than its original model and consumes less power. Furthermore, the TR-compressed model can achieve better performance in terms of stability and model accuracy, compared to other tensor decomposition methods.

Abstract: Presented herein are systems, methods, and datasets for automatically and precisely generating highlight or summary videos of content. In one or more embodiments, the inputs comprise a text (e.g., an article) of the key event(s) (e.g., a goal, a player action, etc.) in an activity (e.g., a game, a concert, etc.) and a video or videos of the activity. In one or more embodiments, the output is a short video of an event or events in the text, in which the video may include commentary about the highlighted events and/or other audio (e.g., music), which may also be automatically synthesized.

Abstract: Different from prior works that model the internal distribution of patches within an image implicitly with a top-down latent variable model (e.g., generator), embodiments explicitly represent the statistical distribution within a single image by using an energy-based generative framework, where a pyramid of energy functions, each parameterized by a bottom-up deep neural network, are used to capture the distributions of patches at different resolutions. Also, embodiments of a coarse-to-fine sequential training and sampling strategy are presented to train the model efficiently. Besides learning to generate random samples from white noise, embodiments can learn in parallel with a self-supervised task (e.g., recover an input image from its corrupted version), which can further improve the descriptive power of the learned model. Embodiments does not require an auxiliary model (e.g.

Abstract: A cooling unit includes a channel frame assembled to a server chassis to form a region enclosing an electronic chip that is disposed on a server board contained within the server chassis, an inlet port coupled to a first side of the channel frame to receive a coolant fluid, an outlet port coupled to a second side of the channel frame for coolant fluid in either a liquid phase or a vapor phase to exit the channel frame, and an internal structure disposed on an inner surface of the channel frame between the inlet port and the outlet port. The internal structure guides the coolant fluid flow and/or distribution along a surface of the electronic chip, where the electronic chip transfers heat to the coolant fluid to cause a portion of the coolant fluid to change from a liquid to a vapor phase.

Abstract: According to one embodiment, a sensing system for an immersion cooling system that includes several flow sensors that are coupled to several lines that are coupled together and are arranged to sense flow rates of coolant flowing through their respective lines. The sensing system also includes a controller that is communicatively coupled to the flow sensors and is configured to receive sensor data from the sensors. The first line couples to a first pump that moves liquid coolant into an information technology (IT) enclosure, a second line couples to a second pump that moves coolant drawn from the enclosure into the second line, and the third line couples to a third pump that moves coolant drawn from a coolant source into the third line, the coolant drawn from the first lime is a combination of coolant from at least one of the second and third lines.

Abstract: The present disclosure provides apparatus, systems, methods, and techniques for cooling using a passive internal return loop with two-phase fluids. The cooling techniques may be applied to an advanced server rack having high power density servers. The server rack may be either packaged with one or more cooling modules or employ a fully sealed immersion configuration. For example, coolant in the vapor state may be separated from coolant in the liquid state in the two-phase coolant, and the coolant in the liquid state may be recirculated to cool the server (or other sub cooling systems) by gravity (i.e., pump-free).

Abstract: A multi-layer seal formed with a first layer for visual detection of leak by incorporating coloring agent that is water soluble, a second layer for electrical detection of a leak by a sensor, the second layer incorporating disbursable agent which disburses into the water upon contacting water and causing a change in chemical, and/or electrical property of the water, and a third layer made of water swelling material which swells upon contact with water. The three layers may be arranged in an orientation wherein a leak would first reach the first layer, then reach the second layer, and lastly reach the swelling layer. The sealing layer arrangement may be correlated to the fluid flowing direction. The layers may form independent parts placed side-by-side or may be integrated into a single seal loop.

Abstract: A method of determining a vehicle travel trajectory, an electronic device, a storage medium and a vehicle, which relate to a field of an artificial intelligence technology, in particular to a field of autonomous driving and intelligent transportation. A specific implementation solution includes: determining an initial path information for a vehicle; optimizing the initial path information to generate a target optimized path information; determining an optimized mapping relationship for velocity according to the target optimized path information and a first energy consumption constraint parameter; and determining an optimized trajectory for the vehicle according to the target optimized path information and the optimized mapping relationship for velocity.

Abstract: Presented herein are embodiments of a two-stage methodology that integrates data-driven imitation learning and model-based trajectory optimization to generate optimal trajectories for autonomous excavators. In one or more embodiments, a deep neural network using demonstration data to mimic the operation patterns of human experts under various terrain states, including their geometry shape and material type. A stochastic trajectory optimization methodology is used to improve the trajectory generated by the neural network to ensure kinematics feasibility, improve smoothness, satisfy hard constraints, and achieve desired excavation volumes. Embodiments were tested on a Franka robot arm equipped with a bucket end-effector. Embodiments were also evaluated on different material types, such as sand and rigid blocks. Experimental results showed that embodiments of the two-stage methodology that comprises combining expert knowledge and model optimization increased the excavation weights by up to 24.

Abstract: According to some embodiments, systems, methods and media for dynamically generating scenario parameters for an autonomous driving vehicles (ADV) are described. In one embodiment, when an ADV enters a driving scenario, the ADV can invoke a map-based scenario checker to determine the type of scenario, and invokes a corresponding neural network model to generate a set of parameters for the scenario based on real-time environmental conditions (e.g., traffics) and vehicle status information (e.g., speed). The set of scenario parameters can be a set of extra constraints for configuring the ADV to drive in a driving mode corresponding to the scenario.