Abstract: Disclosed herein includes a system, a method, and a device including shift circuitry and add circuitry for performing multiplication of a first value and a second value for a neural network. The first value has a predetermined format including a first bit, and two or more second bits to represent a value of zero or 2n where n is an integer greater than or equal to 0. The device shifts, when the two or more second bits represent the value of 2n, the second value by (n+1) bits via the shift circuitry to provide a first result, selectively outputs zero or the second value, based on a value of the first bit of the first value, to provide a second result, and adds the first result and the second results via the add circuitry to provide a result of the multiplication of the first and second values.

Abstract: Disclosed herein includes a system, a method, and a device for improving computational efficiency of deconvolution by reducing a number of dot products. In one aspect, an input image having a set of pixels is received. A first dot product may be performed on a subset of the set of pixels of the input image and a portion of a kernel, to generate a first pixel of an output image. A number of multiplications performed for the first dot product performed may be less than a number of elements of the kernel. A second dot product on a remaining portion of the kernel to generate the first pixel of the output image may be bypassed.

Abstract: Disclosed herein includes a system, a method, and a device for receiving input data to generate a plurality of outputs for a layer of a neural network. The plurality of outputs are arranged in a first array. Dimensions of the first array may be compared with dimensions of a processing unit (PE) array including a plurality of PEs. According to a result of the comparing, the first array is partitioned into subarrays by the processor. Each of the subarrays has dimensions less than or equal to the dimensions of the PE array. A first group of PEs in the PE array is assigned to a first one of the subarrays. A corresponding output of the plurality of outputs is generated using a portion of the input data by each PE of the first group of PEs assigned to the first one of the subarrays.

Abstract: The disclosed transparent electroactive systems may include at least one transparent electroactive material, a first electrode material disposed over a first surface of the transparent electroactive material, and a second electrode material disposed over a second, opposite surface of the transparent electroactive material. The first and second electrode materials may be configured to apply a sufficient voltage to the transparent electroactive material to deform the transparent electroactive material. At least the first electrode material may include conductive traces that are nonlinear. Various other methods and systems are also disclosed.

Abstract: A near-eye display (NED) includes a source assembly, a waveguide outside a field-of-view of a user, and a main optic within the field-of-view. The waveguide expands light emitted from the source assembly in at least one dimension and out-couple the expanded light. The main optic is partially transparent and is positioned such that the user of the NED looks through the main optic to view a local area surrounding the NED. The main optic receives light from the local area, combines the received light with the expanded light to generate combined light, and directs the combined light to the user's eye-box.

Abstract: A system includes a near eye display (NED) comprising a substantially transparent electronic display that is configured to display images in accordance with display instructions, and an imaging device configured to capture one or more images of a portions of a local area surrounding the NED. The system further includes a controller configured to determine a position of an object within the local area using the captured one or more images and location information associated with the object. The controller accesses supplemental information regarding the object, and updates the display instructions to cause the substantially transparent electronic display to display at least a portion of the supplemental information about the object. The display of the at least a portion of the supplemental information is positioned within a threshold distance of the determined position of the object in an augmented reality environment as presented via the substantially transparent electronic display.

Abstract: An audio system for customizing sound fields for increased user privacy. A microphone array of a headset detects sounds from one or more sound sources in a local area of the headset. The audio system estimates array transfer functions (ATFs) associated with the sounds, and determines determining sound field reproduction filters for a loudspeaker array of the headset using the ATFs. The audio system presents audio content, via the loudspeaker array, based in part on the sound field reproduction filters. The presented audio content has a sound field that has a reduced amplitude in a first damped region of the local area that includes a first sound source of the one or more sound sources.

Abstract: A head-mounted-display (HMD) includes a face plate, and a main body. The main body attaches the face plate to cover a front portion of the main body. The main body includes a front plate, a back plate, and a circuit board. The front plate includes at least one first air channel formed at a rear side of the front plate facing away from the face plate for conveying air from a bottom portion of the main body to an upper portion of the main body. Each of the front plates may be formed with windows for receiving an intake air or discharging an outflow air. The back plate is attached to the front plate and located further away from the face plate. The circuit board is located between the front plate and the back plate.

Abstract: In one embodiment, a system includes at least one projector configured to project a plurality of projected patterns onto a scene, the projected patterns including a first projected pattern that includes a plurality of first projected features, a camera configured to capture a plurality of images including a first detected pattern corresponding to a reflection of the first projected pattern, and one or more processors configured to: compute a depth map of the scene based on the first projected pattern, the first detected pattern, and relative positions of the camera and the at least one projector, project, using the projector, a second projected pattern comprising a plurality of second projected features onto a portion of the scene, where the second projected pattern is more sparse than the first projected pattern, and capture, using the camera, a second detected pattern corresponding to a reflection of the second projected pattern.

Abstract: A system generates an output audio signal for an object or virtual object using image data of a room to select a room impulse response from a database. A headset may include a depth camera assembly (DCA) and processing circuitry. The DCA generates depth image data of a room. The processing circuitry determines room parameters such as the dimensions of the room based on the depth image data. A room impulse response for the room is determined based on referencing a database of room impulse responses using the room parameters. An output audio signal is generated by convolving a source audio signal of an object with the room impulse response.

Abstract: A hand-held controller includes a handle extending in a longitudinal direction. The handle is shaped and dimensioned to be grasped by a user's hand. A ring is attached to an end of the handle and surrounds a thumb of the user when the handle is grasped by the user's hand. The ring has an annular surface that defines a plane that forms a predetermined angle with respect to the longitudinal direction, and a curved outer surface. Light sources are mounted on the curved outer surface and configured to emit light to be captured by an imaging device for tracking a position or orientation of the hand-held controller. The light is emitted in a direction normal to the curved outer surface.

Abstract: A display device for projecting light to a viewer may include (1) a plurality of subpixels, in which subpixels may emit light of differing spectral distributions, (2) at least one light deviator disposed optically downstream from the plurality of subpixels, and (3) and a controller. The light emitted from each of the plurality of subpixels may be transmitted through and laterally shifted by the least one light deviator towards a viewer and the at least one light deviator may be mechanically rotatable by a force applied to an outer circumferential region of the at least one light deviator. The controller may control illumination of at least a subset of the plurality of subpixels in synchronization with rotation of the at least one light deviator. Various other apparatus, systems, and methods are also disclosed.

Abstract: A facial-interface system for a head-mounted display may include a facial interface and a facial-interface adjustment apparatus. The facial-interface adjustment apparatus may include (1) a coupling mount for coupling the facial-interface adjustment apparatus to a head-mounted-display housing, (2) an adjustment protrusion that is movable between a plurality of holding positions relative to the coupling mount, the adjustment protrusion having a facial-interface coupling portion coupled to the facial interface, and (3) a spring that biases the adjustment protrusion in a biasing direction toward the facial interface such that the adjustment protrusion moves between separate holding positions of the plurality of holding positions following application of a compressing force to the adjustment protrusion compressing the spring in a compressing direction opposite the biasing direction. Various other systems, devices, and methods are also disclosed.

Abstract: Disclosed herein is a method for using a neural network across multiple devices. The method can include receiving, by a first device configured with a first one or more layers of a neural network, input data for processing via the neural network implemented across the first device and a second device. The method can include outputting, by the first one or more layers of the neural network implemented on the first device, a data set that is reduced in size relative to the input data while identifying one or more features of the input data for processing by a second one or more layers of the neural network. The method can include communicating, by the first device, the data set to the second device for processing via the second one or more layers of the neural network implemented on the second device.

Abstract: Disclosed herein includes a system, a method, and a device for asymmetrical scaling factor support for negative and positive values. A device can include a circuit having a shift circuitry and multiply circuitry. The circuit can be configured to perform computation for a neural network, including multiplying, via the multiply circuitry, a first value and a second value. The circuit can be configured to perform computation for a neural network, including shifting, via the shift circuitry, a result of the multiplying by a determined number of bits. The circuit can be configured to perform computation for a neural network, including outputting the result of the multiplying when a sign bit of the first value is negative, and a result of the shifting when the sign bit of the first value is positive.

Abstract: Disclosed herein includes a system, a method, and a device for multiply-accumulate operation. In one aspect, an input operand is received by control circuitry. In one aspect, the control circuitry determines a sparsity of the input operand, where the sparsity may indicate whether a value of the input operand has a predetermined value or not. In one aspect, the control circuitry determines a stationarity of the input operand, where the stationarity may indicate whether the value of the input operand changes over one or more clock cycles. In one aspect, the input operand is provided to multiply-accumulate circuitry as an input, according to the determined sparsity and stationarity of the input operand.

Abstract: Disclosed herein includes a system, a method, and a device for early-exit from convolution. In some embodiments, at least one processing element (PE) circuit is configured to perform, for a node of a neural network corresponding to a dot-product operation with a set of operands, computation using a subset of the set of operands to generate a dot-product value of the subset of the set of operands. The at least one PE circuit can compare the dot-product value of the subset of the set of operands, to a threshold value. The at least one PE circuit can determine whether to activate the node of the neural network, based at least on a result of the comparing.

Abstract: Disclosed herein includes a system, a method, and a device for pipelined parallelism to accelerate distributed learning network graph. First data for a first layer of a neural network may be stored in memory. First circuitry including a first plurality of processing element (PE) circuits may read the first data from the memory and perform computation for the first layer of the neural network using the first data to generate second data. The first circuitry includes a plurality of buffers for outputting the generated second data as input to second circuitry to perform computation for a second layer of the neural network. The second circuitry includes a second plurality of PE circuits configured to perform computation for the second layer of the neural network using the second data.

Abstract: Disclosed herein includes a system, a method, and a device for reading and writing sparse data in a neural network accelerator. A plurality of slices can be established to access a memory having an access size of a data word. A first slice can be configured to access a first side of the data word in memory. Circuitry can access a mask identifying byte positions within the data word having non-zero values. The circuitry can modify the data word to have non-zero byte values stored starting at an end of the first side, and any zero byte values stored in a remainder of the data word. A determination can be made whether a number of non-zero byte values is less than or equal to a first access size of the first slice. The circuitry can write the modified data word to the memory via at least the first slice.

Abstract: A speaker assembly presents audio content to an ear canal of a user. A speaker of the speaker assembly generates positive and negative acoustic pressure waves. A first vent assembly of the speaker assembly ports the positive acoustic pressure waves to an entrance of the ear canal of the user, whereas a second vent assembly ports the negative acoustic pressure waves to an area behind a pinna of the user. The first and second vent assembly are configured to provide improved audio content playback to the user.