Abstract: At least one computer processor can replace visual words of an unsupervised machine learning classification model with visual objects of an image. At least two co-occurring single visual objects adjacent to each other in pixels of the image can be combined to obtain a compound visual object. The unsupervised machine learning classification model can be augmented to model the image as a mixture of subjects, where each subject is represented through placements of the visual objects in a mixture of concentric spheres centering on a mixture of intersections on a mixture of horizontal layers. At least one processor can learn latent relationships between the placements of the visual objects in a three-dimensional space depicted in the image and image semantics. Learning the latent relationships trains the unsupervised machine learning classification model to perform image subject classification through the placements of the visual objects in a new image.

Abstract: Embodiments are generally directed to compression in machine learning and deep learning processing. An embodiment of an apparatus for compression of untyped data includes a graphical processing unit (GPU) including a data compression pipeline, the data compression pipeline including a data port coupled with one or more shader cores, wherein the data port is to allow transfer of untyped data without format conversion, and a 3D compression/decompression unit to provide for compression of untyped data to be stored to a memory subsystem and decompression of untyped data from the memory subsystem.

Abstract: Embodiments are generally directed to compression in machine learning and deep learning processing. An embodiment of an apparatus for compression of untyped data includes a graphical processing unit (GPU) including a data compression pipeline, the data compression pipeline including a data port coupled with one or more shader cores, wherein the data port is to allow transfer of untyped data without format conversion, and a 3D compression/decompression unit to provide for compression of untyped data to be stored to a memory subsystem and decompression of untyped data from the memory subsystem.

Abstract: Embodiments are generally directed to compression in machine learning and deep learning processing. An embodiment of an apparatus for compression of untyped data includes a graphical processing unit (GPU) including a data compression pipeline, the data compression pipeline including a data port coupled with one or more shader cores, wherein the data port is to allow transfer of untyped data without format conversion, and a 3D compression/decompression unit to provide for compression of untyped data to be stored to a memory subsystem and decompression of untyped data from the memory subsystem.

Abstract: Embodiments are generally directed to compression in machine learning and deep learning processing. An embodiment of an apparatus for compression of untyped data includes a graphical processing unit (GPU) including a data compression pipeline, the data compression pipeline including a data port coupled with one or more shader cores, wherein the data port is to allow transfer of untyped data without format conversion, and a 3D compression/decompression unit to provide for compression of untyped data to be stored to a memory subsystem and decompression of untyped data from the memory subsystem.

Abstract: A method for measuring a distance to a target in a multi-user environment, comprising: irradiating the environment by a series of light pulses, wherein this series of light pulses is emitted by a battery of at least two or a single light source device emitting on at least two different wavelengths, the light pulses being emitted at a determined repetition rate and with a determined randomly selected wavelength; collecting pulses reflected or scattered from the environment to at least one detector equipped with a wavelength filter whose pass band corresponds to the selected emitted wavelength; assigning a timestamp at the detection of a pulse by at least one chronometer connected to the detector, said timestamps corresponding to the time of arrival (TOA); determining the statistical distribution of said time of arrivals; determining the distance to the target from said statistical distribution.

Abstract: An image acquisition apparatus based on a miniature camera matrix, comprising an image generation circuit board and a miniature lens matrix, wherein the image generation circuit board comprises an image generation circuit matrix, various image generation circuits in the image generation circuit matrix all comprise a miniature photosensitive element, and various miniature photosensitive elements constitute a miniature photosensitive element matrix; and various miniature lenses in the miniature lens matrix are fixedly arranged on the image generation circuit board and correspond to the various miniature photosensitive elements on a one-to-one basis, and the axes of the various miniature lenses are respectively perpendicular to a plane where the image generation circuit board is located.

Abstract: An image acquisition apparatus based on a miniature camera matrix, comprising an image generation circuit board (200) and a miniature lens matrix, wherein the image generation circuit board (200) comprises an image generation circuit matrix, various image generation circuits (200) in the image generation circuit matrix all comprise a miniature photosensitive element (201), and various miniature photosensitive elements (201) constitute a miniature photosensitive element matrix; and various miniature lenses (203) in the miniature lens matrix are fixedly arranged on the image generation circuit board (200) and correspond to the various miniature photosensitive elements (201) on a one-to-one basis, and the axes of the various miniature lenses (203) are respectively perpendicular to a plane where the image generation circuit board (200) is located.

Abstract: A method for measuring a distance to a target in a multi-user environment, comprising: irradiating the environment by a series of light pulses, wherein this series of light pulses is emitted by a battery of at least two or a single light source device emitting on at least two different wavelengths, the light pulses being emitted at a determined repetition rate and with a determined randomly selected wavelength; collecting pulses reflected or scattered from the environment to at least one detector equipped with a wavelength filter whose pass band corresponds to the selected emitted wavelength; assigning a timestamp at the detection of a pulse by at least one chronometer connected to the detector, said timestamps corresponding to the time of arrival (TOA); determining the statistical distribution of said time of arrivals; determining the distance to the target from said statistical distribution.

Abstract: Provided is an image acquisition apparatus based on an industrial digital camera matrix, comprising a first substrate and a second substrate arranged in parallel. The first substrate is provided with a lens matrix, and axes of various lenses in the lens matrix are respectively perpendicular to a plane where the first substrate is located; and a surface, towards the first substrate, of the second substrate is provided with a photosensitive element matrix, and various photosensitive elements in the photosensitive element matrix are arranged in one-to-one correspondence with the various lenses.

Abstract: Embodiments are generally directed to compression in machine learning and deep learning processing. An embodiment of an apparatus for compression of untyped data includes a graphical processing unit (GPU) including a data compression pipeline, the data compression pipeline including a data port coupled with one or more shader cores, wherein the data port is to allow transfer of untyped data without format conversion, and a 3D compression/decompression unit to provide for compression of untyped data to be stored to a memory subsystem and decompression of untyped data from the memory subsystem.

Abstract: Methods, apparatus, systems and articles of manufacture for low precision training of a machine learning model are disclosed. An example apparatus includes a low precision converter to calculate an average magnitude of weighting values included in a tensor, the weighting values represented in a high precision format, the low precision converter to calculate a maximal magnitude of the weighting values included in the tensor, determine a squeeze factor and a shift factor based on the average magnitude and the maximal magnitude, and convert the weighting values from the high precision format into a low precision format based on the squeeze factor and the shift factor. A model parameter memory is to store the tensor as part of a machine learning model, the tensor including the weighting values represented in the low precision format, the shift factor, and squeeze factor. A model executor is to execute the machine learning model.

Abstract: Embodiments are generally directed to compression in machine learning and deep learning processing. An embodiment of an apparatus for compression of untyped data includes a graphical processing unit (GPU) including a data compression pipeline, the data compression pipeline including a data port coupled with one or more shader cores, wherein the data port is to allow transfer of untyped data without format conversion, and a 3D compression/decompression unit to provide for compression of untyped data to be stored to a memory subsystem and decompression of untyped data from the memory subsystem.

Abstract: Provided is an image acquisition apparatus based on an industrial digital camera matrix, comprising a first substrate and a second substrate arranged in parallel. The first substrate is provided with a lens matrix, and axes of various lenses in the lens matrix are respectively perpendicular to a plane where the first substrate is located; and a surface, towards the first substrate, of the second substrate is provided with a photosensitive element matrix, and various photosensitive elements in the photosensitive element matrix are arranged in one-to-one correspondence with the various lenses.

Abstract: Embodiments are generally directed to compression in machine learning and deep learning processing. An embodiment of an apparatus for compression of untyped data includes a graphical processing unit (GPU) including a data compression pipeline, the data compression pipeline including a data port coupled with one or more shader cores, wherein the data port is to allow transfer of untyped data without format conversion, and a 3D compression/decompression unit to provide for compression of untyped data to be stored to a memory subsystem and decompression of untyped data from the memory subsystem.

Abstract: An electronic reader system includes an electronic book and a stylus. The electronic book includes a battery and a cover. The stylus includes a pen body for writing on the electronic book to input data or commands into the electronic book, and an illumination member detachably attached to the pen body. The illumination member can be detached from the pen body and attached onto the electronic book, receiving electrical power from the book as it does so, to illuminate the electronic book.

Abstract: A battery charger includes a base, a plug formed on the base and electrically connected to an external power source, an interface formed on the base and electrically connecting an external device, a clamp formed on the base and fixing and electrically connecting a battery, and a controller embedded in the base and electrically connected to the plug, the clamp, and the interface. The controller is able to boost the battery voltage to a rated charging voltage of the external device and conducting the voltage to the interface to charge the external device.

Abstract: An electronic reader system includes an electronic book and a stylus. The electronic book includes a battery and a cover. The stylus includes a pen body for writing on the electronic book to input data or commands into the electronic book, and an illumination member detachably attached to the pen body. The illumination member can be detached from the pen body and attached onto the electronic book, receiving electrical power from the book as it does so, to illuminate the electronic book.

Abstract: An electronic device with power output function includes an interface for connecting with an external power supply or a small electronic device. The electronic device further includes a battery, a boost circuit, a switch circuit, and a charging circuit. When the external power supply is connected to the electronic device through a first cable, the switch circuit is closed to allow the external power supply to charge the battery through the first cable, the interface, the switch circuit, and the charging circuit. When the small electronic device is connected to the electronic device through a second cable, the boost circuit is enabled to boost the power voltage of the battery to allow the battery to charge the small electronic device through the boost circuit, the interface, and the second cable.

Abstract: A relatively moving surface is illuminated with a laser. Light from the laser is reflected by the surface into an array of photosensitive elements; the reflected light includes a speckle pattern. Sums are calculated for outputs of pixels perpendicular to a first dimension along which motion is to be determined. Motion along the first dimension is then determined based on spatial and temporal gradients of the calculated sums. Sums are also calculated for outputs of pixels perpendicular to a second dimension along which motion is to be determined. Motion along the second dimension is then determined based on spatial and temporal gradients of those sums. The array may be rectangular, or may contain arms separated by a pixel-free region.