Abstract: Apparatuses, systems, and techniques to use one or more neural networks to generate data labels. In at least one embodiment, one or more neural networks is trained based, at least in part on, one or more labels, pseudo-labels, training data, and modified training data.

Abstract: A furnace may include an outer wall defining a chamber, the chamber including an internal cavity configured to receive one or more parts, at least one heater positioned within the chamber, the at least one heater being configured to generate temperatures of at least about 800 degrees Celsius within the internal cavity, and a vacuum pump configured to apply a vacuum to at least a portion of the chamber. The furnace may also include at least one layer of inner insulation and at least one layer of outer insulation disposed outward of the inner insulation with respect to the chamber, the at least one layer of outer insulation being sealed with respect to the at least one layer of inner insulation.

Abstract: Navigation instructions are determined using visual data or other sensory information. Individual frames can be extracted from video data, captured from passes through an environment, to generate a sequence of image frames. The frames are processed using a feature extractor to generate frame-specific feature vectors. Image triplets are generated, including a representative image frame (or corresponding feature vector), a similar image frame adjacent in the sequence, and a disparate image frame that is separated by a number of frames in the sequence. The embedding network is trained using the triplets. Image data for a current position and a target destination can then be provided as input to the trained embedding model, which outputs a navigation vector indicating a direction and distance over which the vehicle is to be navigated in the physical environment.

Abstract: Navigation instructions are determined using visual data or other sensory information. Individual frames can be extracted from video data, captured from passes through an environment, to generate a sequence of image frames. The frames are processed using a feature extractor to generate frame-specific feature vectors. Image triplets are generated, including a representative image frame (or corresponding feature vector), a similar image frame adjacent in the sequence, and a disparate image frame that is separated by a number of frames in the sequence. The embedding network is trained using the triplets. Image data for a current position and a target destination can then be provided as input to the trained embedding model, which outputs a navigation vector indicating a direction and distance over which the vehicle is to be navigated in the physical environment.

Abstract: Navigation instructions are determined using visual data or other sensory information. Individual frames can be extracted from video data, captured from passes through an environment, to generate a sequence of image frames. The frames are processed using a feature extractor to generate frame-specific feature vectors. Image triplets are generated, including a representative image frame (or corresponding feature vector), a similar image frame adjacent in the sequence, and a disparate image frame that is separated by a number of frames in the sequence. The embedding network is trained using the triplets. Image data for a current position and a target destination can then be provided as input to the trained embedding model, which outputs a navigation vector indicating a direction and distance over which the vehicle is to be navigated in the physical environment.

Abstract: A game ball of composite construction, having a spherical hollow shell, with a pattern of perforation or holes arranged on and through the surface thereof, to substantially increase the aerodynamic drag upon the composite, and thereby effective reduce the distance the composite can travel upon being struck by a bat or club. In order to modulate and better emulate or mimic the feel and track of the composite in play, the composite is further modified by the inclusion of a plurality of randomly ordered discrete spherical elements, within said hollow shell, to effectively displacing about 50% of the interior volume of the hollow shell, and thereby contributing from about 20% up to about 75% to the mass of said composite In the composite embodiment of this invention game ball is a golf ball.

Abstract: A method for the manufacture of metallically encapsulated ceramic armor through the use of low pressure processing methods in autoclave furnaces or brazing furnaces.

Abstract: A method for the manufacture through diffusion bonding of metallically encapsulated ceramic armor providing enhanced ballistic efficiency and physical durability.

Abstract: A method for the manufacture through diffusion bonding of metallically encapsulated ceramic armor providing enhanced ballistic efficiency and physical durability.

Abstract: A method for the manufacture of metallically encapsulated ceramic armor through the use of low pressure processing methods in autoclave furnaces or brazing furnaces.

Abstract: The pair of denim pants, particularly jeans, designed for added durability in areas of high stress. The areas of high stress include the waistband, the belt loops, pockets and yoke on a pair of denim. These areas of high stress are lined and reinforced with a nylon material. The denim material will encase the nylon material increasing the strength of pants when stitched in the final form.