Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for selecting a neural network to perform a particular machine learning task while satisfying a set of constraints.

Abstract: A data service layer running on a storage director node generates a request to destage host data from a plurality of cache slots in a single back-end track. The destage request includes pointers to addresses of the cache slots and indicates an order in which the host application data in the cache slots is to be included in the back-end track. A back-end redundant array of independent drives (RAID) subsystem running on a drive adapter is responsive to the request to calculate parity information using the host application data in the cache slots. The back-end RAID subsystem assembles the single back-end track comprising the host application data from the plurality of cache slots of the request, and destages the single back-end track to a non-volatile drive in a single back-end input-output (IO) operation.

Abstract: The present disclosure discloses a display panel, a method for manufacturing the same, and a display device, which belong to the field of display technologies. The display panel includes: a first substrate and a second substrate which are oppositely disposed. The first substrate may include one or more light-emitting unit, and the second substrate may include two or more reflective electrodes. All first reflective electrodes in the display panel are capable of reflecting light emitted by the light-emitting unit to a first view zone, and all second reflective electrodes are capable of emitting the light emitted by the light-emitting unit to a second view zone. Therefore, a complete picture is viewed in the first view zone, and another complete picture can be viewed in the second view zone.

Abstract: A connector housing includes an accommodation space defined by four walls and formed with an insertion port. The four walls include a first wall extending in a first plane and a second wall extending in a second plane perpendicular to the first plane, the first wall is connected with the second wall at a corner of the connector housing. A first positioning groove is disposed in an edge of the first wall proximate to the insertion port and a first positioning tooth is disposed on an edge of the second wall. The first positioning tooth extends in the first plane by vertically bending and engaging within the first positioning groove. The first positioning tooth and the first positioning groove have a first locking feature preventing the first positioning tooth from being disengaged from the first positioning groove in a direction perpendicular to the second plane.

Abstract: Predicting patch displacement maps using a neural network is described. Initially, a digital image on which an image editing operation is to be performed is provided as input to a patch matcher having an offset prediction neural network. From this image and based on the image editing operation for which this network is trained, the offset prediction neural network generates an offset prediction formed as a displacement map, which has offset vectors that represent a displacement of pixels of the digital image to different locations for performing the image editing operation. Pixel values of the digital image are copied to the image pixels affected by the operation.

Abstract: An array substrate is provided. One of a first electrode layer and a second electrode layer in the array substrate includes at least one slit electrode. The slit electrode is disposed between two adjacent data leads in the array substrate, and includes an electrode connecting portion and a plurality of first strip-shaped sub-electrodes. The electrode connecting portion includes a first connecting section parallel to and adjacent to the data lead, and a distance between two adjacent first strip-shaped sub-electrodes in a direction parallel to an extending direction of the first connecting section gradually increases along a direction going away from the first connecting section.

Abstract: An electronic sensing apparatus and a method of producing the electronic sensing apparatus includes a triboelectric generator encapsulated between a bottom substrate and a top encapsulation layer, wherein the triboelectric generator is arranged to generate a triboelectric sensing signal in response to a deformation of the bottom substrate and/or the top encapsulation layer.

Abstract: Systems and methods can utilize a conformer model to process a data set for various data processing tasks, including, but not limited to, speech recognition, sound separation, protein synthesis determination, video or other image set analysis, and natural language processing. The conformer model can use feed-forward blocks, a self-attention block, and a convolution block to process data to learn global interactions and relative-offset-based local correlations of the input data.

Abstract: A storage system that supports multiple RAID levels presents storage objects with front-end tracks corresponding to back-end tracks on non-volatile drives and accesses the drives using a single type of back-end allocation unit that is larger than a back-end track. When the number of members of a protection group of a RAID level does not align with the back-end allocation unit, multiple back-end tracks are grouped and accessed using a single IO. The number of back-end tracks in a group is selected to align with the back-end allocation unit size. If the front-end tracks are variable size, then front-end tracks may be destaged into a smaller number of grouped back-end tracks in a single IO.

Abstract: Digital image completion by learning generation and patch matching jointly is described. Initially, a digital image having at least one hole is received. This holey digital image is provided as input to an image completer formed with a dual-stage framework that combines a coarse image neural network and an image refinement network. The coarse image neural network generates a coarse prediction of imagery for filling the holes of the holey digital image. The image refinement network receives the coarse prediction as input, refines the coarse prediction, and outputs a filled digital image having refined imagery that fills these holes. The image refinement network generates refined imagery using a patch matching technique, which includes leveraging information corresponding to patches of known pixels for filtering patches generated based on the coarse prediction. Based on this, the image completer outputs the filled digital image with the refined imagery.

Abstract: A computer-implemented method of training a streaming speech recognition model that includes receiving, as input to the streaming speech recognition model, a sequence of acoustic frames. The streaming speech recognition model is configured to learn an alignment probability between the sequence of acoustic frames and an output sequence of vocabulary tokens. The vocabulary tokens include a plurality of label tokens and a blank token. At each output step, the method includes determining a first probability of emitting one of the label tokens and determining a second probability of emitting the blank token. The method also includes generating the alignment probability at a sequence level based on the first probability and the second probability. The method also includes applying a tuning parameter to the alignment probability at the sequence level to maximize the first probability of emitting one of the label tokens.

Abstract: An automated speech recognition (ASR) model includes a first encoder, a second encoder, and a decoder. The first encoder receives, as input, a sequence of acoustic frames, and generates, at each of a plurality of output steps, a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The second encoder receives, as input, the first higher order feature representation generated by the first encoder at each of the plurality of output steps, and generates, at each of the plurality of output steps, a second higher order feature representation for a corresponding first higher order feature frame. The decoder receives, as input, the second higher order feature representation generated by the second encoder at each of the plurality of output steps, and generates, at each of the plurality of time steps, a first probability distribution over possible speech recognition hypotheses.

Abstract: Digital image completion using deep learning is described. Initially, a digital image having at least one hole is received. This holey digital image is provided as input to an image completer formed with a framework that combines generative and discriminative neural networks based on learning architecture of the generative adversarial networks. From the holey digital image, the generative neural network generates a filled digital image having hole-filling content in place of holes. The discriminative neural networks detect whether the filled digital image and the hole-filling digital content correspond to or include computer-generated content or are photo-realistic. The generating and detecting are iteratively continued until the discriminative neural networks fail to detect computer-generated content for the filled digital image and hole-filling content or until detection surpasses a threshold difficulty.

Abstract: Certain embodiments involve using an image completion neural network to perform user-guided image completion. For example, an image editing application accesses an input image having a completion region to be replaced with new image content. The image editing application also receives a guidance input that is applied to a portion of a completion region. The image editing application provides the input image and the guidance input to an image completion neural network that is trained to perform image-completion operations using guidance input. The image editing application produces a modified image by replacing the completion region of the input image with the new image content generated with the image completion network. The image editing application outputs the modified image having the new image content.

Abstract: Digital image completion by learning generation and patch matching jointly is described. Initially, a digital image having at least one hole is received. This holey digital image is provided as input to an image completer formed with a dual-stage framework that combines a coarse image neural network and an image refinement network. The coarse image neural network generates a coarse prediction of imagery for filling the holes of the holey digital image. The image refinement network receives the coarse prediction as input, refines the coarse prediction, and outputs a filled digital image having refined imagery that fills these holes. The image refinement network generates refined imagery using a patch matching technique, which includes leveraging information corresponding to patches of known pixels for filtering patches generated based on the coarse prediction. Based on this, the image completer outputs the filled digital image with the refined imagery.

Abstract: The embodiments of the present invention provide training and construction methods and apparatus of a neural network for object detection, an object detection method and apparatus based on a neural network and a neural network. The training method of the neural network for object detection, comprises: inputting a training image including a training object to the neural network to obtain a predicted bounding box of the training object; acquiring a first loss function according to a ratio of the intersection area to the union area of the predicted bounding box and a true bounding box, the true bounding box being a bounding box of the training object marked in advance in the training image; and adjusting parameters of the neural network by utilizing at least the first loss function to train the neural network.

Abstract: Digital image completion by learning generation and patch matching jointly is described. Initially, a digital image having at least one hole is received. This holey digital image is provided as input to an image completer formed with a dual-stage framework that combines a coarse image neural network and an image refinement network. The coarse image neural network generates a coarse prediction of imagery for filling the holes of the holey digital image. The image refinement network receives the coarse prediction as input, refines the coarse prediction, and outputs a filled digital image having refined imagery that fills these holes. The image refinement network generates refined imagery using a patch matching technique, which includes leveraging information corresponding to patches of known pixels for filtering patches generated based on the coarse prediction. Based on this, the image completer outputs the filled digital image with the refined imagery.

Abstract: Predicting patch displacement maps using a neural network is described. Initially, a digital image on which an image editing operation is to be performed is provided as input to a patch matcher having an offset prediction neural network. From this image and based on the image editing operation for which this network is trained, the offset prediction neural network generates an offset prediction formed as a displacement map, which has offset vectors that represent a displacement of pixels of the digital image to different locations for performing the image editing operation. Pixel values of the digital image are copied to the image pixels affected by the operation.

Abstract: Digital image completion using deep learning is described. Initially, a digital image having at least one hole is received. This holey digital image is provided as input to an image completer formed with a framework that combines generative and discriminative neural networks based on learning architecture of the generative adversarial networks. From the holey digital image, the generative neural network generates a filled digital image having hole-filling content in place of holes. The discriminative neural networks detect whether the filled digital image and the hole-filling digital content correspond to or include computer-generated content or are photo-realistic. The generating and detecting are iteratively continued until the discriminative neural networks fail to detect computer-generated content for the filled digital image and hole-filling content or until detection surpasses a threshold difficulty.

Abstract: Predicting patch displacement maps using a neural network is described. Initially, a digital image on which an image editing operation is to be performed is provided as input to a patch matcher having an offset prediction neural network. From this image and based on the image editing operation for which this network is trained, the offset prediction neural network generates an offset prediction formed as a displacement map, which has offset vectors that represent a displacement of pixels of the digital image to different locations for performing the image editing operation. Pixel values of the digital image are copied to the image pixels affected by the operation by: determining the vectors pixels that correspond to the image pixels affected by the image editing operation and mapping the pixel values of the image pixels represented by the determined offset vectors to the affected pixels.