Abstract: In examples there is a computing device comprising a processor, the processor having a memory management unit. The computing device also has a memory that stores instructions that, when executed by the processor, cause the memory management unit to receive a memory access instruction comprising a virtual memory address; translate the virtual memory address to a physical memory address of the memory, and obtain permission information associated with the physical memory address. Responsive to the permission information indicating that metadata is permitted to be associated with the physical memory address, a check is made of a metadata summary table stored in the physical memory to check whether metadata is compatible with the physical memory address. Responsive to the check being unsuccessful, a trap is sent to system software of the computing device in order to trigger dynamic allocation of physical memory for storing metadata associated with the physical memory address.

Abstract: A deconvolution can be decomposed into multiple convolutions. Results of the convolutions constitute an output of the deconvolution. Zeros may be added to an input tensor of the deconvolution to generate an upsampled input tensor. Subtensors having the same size as the kernel of the deconvolution may be identified from the upsampled input tensor. A subtensor may include one or more input activations and one or more zeros. Subtensors having same distribution patterns of input activations may be used to generate a reduced kernel. The reduced kernel includes a subset of the kernel. The position of a weight in the reduced kernel may be the same as the positions of an input activation in the subtensor. Multiple reduced kernels may be generated based on multiple subtensors having different distribution patterns of activations. Each of the convolutions may use the input tensor and a different one of the reduced kernels.

Abstract: A memory array of a compute tile may store activations or weights of a DNN. The memory array may include databanks for storing contexts, context MUXs, and byte MUXs. A databank may store a context with flip-flop arrays, each of which includes a sequence of flip-flops. A logic gate and an ICG unit may gate flip-flops and control whether states of the flip-flops can be changed. The data gating can prevent a context not selected for the databank from inadvertently toggling and wasting power A context MUX may read a context from different flip-flop arrays in a databank based on gray-coded addresses. A byte MUX can combine bits from different bytes in a context read by the context MUX. The memory array may be implemented with bit packing to reduce distance between the context MUX and byte MUX to reduce lengths of wires connecting the context MUXs and byte MUXs.

Abstract: Sparsity processing within a compute block can be done on unpacked data. The compute block includes a sparsity decoder that generates a combined sparsity vector from an activation sparsity vector and a weight sparsity vector. The activation sparsity vector indicates positions of non-zero valued activations in an activation context. The weight sparsity vector indicates positions of non-zero valued weights in a weight context. The combined sparsity vector comprises one or more zero valued bits and one or more non-zero valued bits. The sparsity decoder may determine the position of a non-zero valued bit in the combined sparsity vector and determine an address for the non-zero valued activation and the non-zero valued weight based on the position of the non-zero valued bit. The non-zero valued activation and the non-zero valued weight may be provided to a PE for performing MAC operations.

Abstract: A compute tile includes a WCB that receives a workload of writing an output tensor of a convolution into a local memory of the compute tile. The local memory may be a SRAM. The WCB receives write transactions. A write transaction includes a data block, which is a part of the output tensor, and metadata describing one or more attributes of the data block. The WCB may store write transactions in its internal buffers. The WCB may determine whether to combine two write transactions, e.g., based on an operation mode or metadata in the write transactions. In embodiments where the WCB determines to combine the two write transactions, the WCB may combine the two write transactions into a new write transaction and write the new write transaction into the local memory or an internal memory of the WCB. The total number of write transactions for the workload can be reduced.

Abstract: An DNN accelerator includes a DMA engine that can rearrange weight data layout. The DMA engine may read a weight tensor from a memory (e.g., DRAM). The weight tensor includes weights arranged in a 3D matrix. The DMA engine may partition the weight tensor into a plurality of virtual banks based on a structure of a PE array, e.g., based on the number of activated PE columns in the PE array. Then the DMA engine may partition a virtual bank into a plurality of virtual sub-banks. The DMA engine may also identify data blocks from different ones of the plurality of virtual sub-banks. A data block may include a plurality of input channels and may have a predetermined spatial size and storage size. The DMA engine form a linear data structure by interleaving the data blocks. The DMA engine can write the linear data structure into another memory (e.g., SRAM).

Abstract: An DNN accelerator may perform 1×N kernel decomposition to decompose a convolutional kernel into kernel vectors, each of which includes multiple weights. Through the kernel decomposition, a weight operand may be generated from a filter. The DNN accelerator converts an input tensor into input operands. An input operand includes activations and has the same size as the weight operand. The DNN accelerator may read a first activation in the input operand from memory to an internal memory of a first PE and read a second activation in the input operand from the memory to an internal memory of a second PE. The first PE may receive the second activation from the second PE through activation broadcasting between the two PEs and perform MAC operations on the input operand and weight operand. The second PE may perform MAC operations on another input operand in the input tensor and the weight operand.

Abstract: Devices and techniques are generally described for a speech processing routing architecture. In various examples, first data comprising a first feature definition is received. The first feature definition may include a first indication of first source data and first instructions for generating feature data using the first source data. In various examples, the feature data may be generated according to the first feature definition. In some examples, a speech processing system may receive a first request to process a first utterance. The feature data may be retrieved from a non-transitory computer-readable memory. The speech processing system may determine a first skill for processing the first utterance based at least in part on the feature data.

Abstract: Methods for optimizing Boundary Layer Control (BLC) systems and related systems (e.g. a Laminar Flow Control (LFC) system or systems, a Static Pressure Thrust (SPT) system or systems, a Boundary Layer Ingestion (BLI)/Wake Immersed Propulsion (WIP) system or systems, and/or low-dissipation BLC fluid-movement system or systems) to operate in concert with each other and a bellows air-moving system are disclosed.

Abstract: In accordance with some embodiments, an apparatus for privacy protection is provided. The apparatus includes a housing arranged to hold a personal communication device and a peripheral interface supported by the housing, where the peripheral interface is connectable to a supplemental functional device. The apparatus further includes a local communication device coupled to the peripheral interface and supported by the housing, where the local communication device includes a personal communication device interface modem operable to provide a communication channel between the peripheral interface and the personal communication device. The apparatus further includes a controller coupled to the peripheral interface and the local communication device, where the controller is operable to manage the communication channel between the supplemental functional device and the personal communication device.

Abstract: In accordance with some embodiments, an apparatus that controls sensor paths for privacy protection is provided. The apparatus includes a housing arranged to hold a second device. The apparatus obtains first sensor data that includes a biometric marker associated with a user. The apparatus controls sensor paths by obtaining the first sensor data using sensors on the second device, on the apparatus, and/or on a supplemental functional device. The apparatus further generates second sensor data by masking the biometric marker associated with the user in the first sensor data. The apparatus additionally controls the sensor paths by providing the second sensor data from the first apparatus to the second device.

Abstract: The invention is directed to a composite material comprising a biofabricated material and a secondary component. The secondary component may be a porous material, such as a sheet of paper, cellulose, or fabric that has been coated or otherwise contacted with the biofabricated material. The biofabricated material comprises a uniform network of crosslinked collagen fibrils and provides strength, elasticity and an aesthetic appearance to the composite material.

Abstract: A plurality of sensors and a controller are disposed in a marine seismic streamer. Each of the sensors comprises an enclosure having two opposing interior walls, first and second piezoelectric elements disposed on the opposing interior walls, a third piezoelectric element disposed on a flexible substrate within the enclosure between the opposing interior walls, a pressure signal output node and an acceleration signal output node disposed on the exterior surface of the enclosure. A combined pressure signal derived from the pressure signal output nodes of the plural sensors is coupled to a pressure signal input of the controller. A combined acceleration signal derived from the acceleration signal output nodes of the plural sensors is coupled to an acceleration signal input of the controller. The streamer may be towed, and the combined pressure and acceleration signals may be recorded in a computer-readable medium.

Abstract: A biofabricated material comprising a network of crosslinked collagen fibrils produced from recombinant collagen that contains substantially no 3-hydroxyproline residues is disclosed. This material is composed of collagen which is also a major component of natural leather and is produced by a process of fibrillation of collagen molecules into fibrils, crosslinking the fibrils and lubricating the crosslinked fibrils. Unlike natural leathers, this biofabricated material exhibits non-anisotropic (not directionally dependent) physical properties, for example, a sheet of biofabricated material can have substantially the same elasticity or tensile strength when stretched or stressed in different directions. Unlike natural leather, it has a uniform texture that facilitates uniform uptake of dyes and coatings. Aesthetically, it produces a uniform and consistent grain for ease of manufacturability.

Abstract: The present invention provides a compound of formula (I) or a pharmaceutically acceptable salt thereof; Wherein R1, R3-R6, X2 and X3 are as defined herein, a method for manufacturing the compounds of the invention, and its therapeutic uses. The present invention further provides a combination of pharmacologically active agents and a pharmaceutical composition.

Abstract: A shield system for an unmanned arial vehicle (AUV) is provided. The shield system includes inner shield members rigidly mounted to at least the rotor arms of the AUV and outer shield members shock mounted to the inner shield members. The shield system defines opening for sensors, payload or mechanical portions of the UAV.

Abstract: The invention is directed to a composite material comprising a biofabricated material and a secondary component. The secondary component may be a porous material, such as a sheet of paper, cellulose, or fabric that has been coated or otherwise contacted with the biofabricated material. The biofabricated material comprises a uniform network of crosslinked collagen fibrilsand provides strength, elasticity and an aesthetic appearance to the composite material.

Abstract: A computer system has a separation mechanism which enforces separation between at least two execution environments such that one execution environment is a gatekeeper which interposes on all communications of the other execution environment. The computer system has an attestation mechanism which enables the gatekeeper to attest to properties of the at least two execution environments. A first one of the execution environments runs application specific code which may contain security vulnerabilities. The gatekeeper is configured to enforce an input output policy on the first execution environment by interposing on all communication to and from the first execution environment by forwarding, modifying or dropping individual ones of the communications according to the policy. The gatekeeper provides evidence of attestation both for the application specific code and the policy.

Abstract: A system for converting an infant child seat to a stroller may include an infant child seat, a wheeled frame, and a handle. The handle is coupled to the infant child seat and configured to manipulate the wheeled frame when the infant child seat is mounted. A wheeled frame, being removably securable to an infant child seat, may include a frame movable between a folded storage configuration and an expanded stroller configuration. The frame is configured to removably receive an infant child seat and includes a plurality of wheels. The wheeled frame does not define a handle. A method of converting an infant child seat including a handle to a stroller may include coupling the seat to a wheeled frame. The method further includes manually actuating, by a first hand of a user, the handle into a stroller position. The seat coupled to the frame are operable as a stroller.

Abstract: An apparatus for supporting a resting or sleeping occupant may include a compressible structure configured to allow airflow through the compressible structure, and a cover removably coupled to the compressible structure and overlaying at least a portion of the compressible structure. The cover may include a foldover portion. The foldover portion may include a zipper along an edge of the foldover portion, and may be secured about a portion of the compressible structure. A method of assembling the apparatus may include inserting the cover through a compressible structure, and wrapping the foldover portion around an outer rim of the compressible structure. The method may further include manipulating the zipper along the foldover portion to secure the foldover portion about the compressible structure.