Abstract: The present disclosure discloses a gradient single-crystal positive electrode material, which has a chemical formula of LiNixCoyA1-x-yO2@mLiaZbOc, wherein 0<x<1, 0<y<1, 0<x+y<1, 0<m<0.05, 0.3<a?10, 1?b<4, and 1?c<15, A is at least one of Mn, Zr, Sr, Ba, W, Ti, Al, Mg, Y, and Nb, and Z is at least one of B, Al, Co, W, Ti, Zr, and Si. The atomic ratio of the content of Co on the surface of the single-crystal positive electrode material particle to the content of Ni+Co+A on the surface is greater than 0.4 and less than 0.8, and the atomic ratio of Co at a depth 10% of the radius from the surface of the single crystal positive electrode material particle is not less than 0.3; and the single-crystal positive electrode material particle has a roundness of greater than 0.4, and is free from sharp corners.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing neural architecture search for machine learning models. In one aspect, a method comprises receiving training data for a machine learning, generating a plurality of candidate neural networks for performing the machine learning task, wherein each candidate neural network comprises a plurality of instances of a layer block composed of a plurality of layers, for each candidate neural network, selecting a respective type for each of the plurality of layers from a set of layer types that comprises, training the candidate neural network and evaluating performance scores for the trained candidate neural networks as applied to the machine learning task, and determining a final neural network for performing the machine learning task based at least on the performance scores for the candidate neural networks.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing machine learning task on a network input to generate a network output. In one aspect, one of the systems includes an attention neural network configured to perform the machine learning task, the attention neural network including one or more attention layers, each attention layer comprising an attention sub-layer and a feed-forward sub-layer. Some or all of the attention layers have a feed-forward sub-layer that applies conditional computation to the inputs to the sub-layer.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: Systems and methods for routing in mixture-of-expert models. In some aspects of the technology, a transformer may have at least one Mixture-of-Experts (“MoE”) layer in each of its encoder and decoder, with the at least one MoE layer of the encoder having a learned gating function configured to route each token of a task to two or more selected expert feed-forward networks, and the at least one MoE layer of the decoder having a learned gating function configured to route each task to two or more selected expert feed-forward networks.

Abstract: Disclosed is an application of Epigallocatechin Gallate (EGCG) or a compound synthesized with it as a lead compound in combination with one or more tyrosine kinase inhibitors in preparation of a cancer treatment drug, herein the cancer is an Epidermal Growth Factor Receptor (EGFR) wild-type tumor. This combined use may significantly inhibit the growth of the EGFR wild-type tumor, and reduce the toxic and side effects of an anticancer drug.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training giant neural networks. One of the methods includes obtaining data specifying a partitioning of the neural network into N composite layers that form a sequence of composite layers, wherein each composite layer comprises a distinct plurality of layers from the multiple network layers of the neural network; obtaining data assigning each of the N composite layers to one or more computing devices from a set of N computing devices; partitioning a mini-batch of training examples into a plurality of micro-batches; and training the neural network, comprising: performing a forward pass through the neural network until output activations have been computed for each micro-batch for a final composite layer in the sequence, and performing a backward pass through the neural network until output gradients have been computed for each micro-batch for the first composite layer in the sequence.

Abstract: Disclosed are a button battery and a manufacturing method therefor, the button battery includes an upper cover plate, a pole, an insulating sleeve, a bottom shell, a battery winding core and a sealing ball. The upper cover plate is provided with a liquid injection hole and a stepped through hole with a small upper portion and a large lower portion; upper and lower ends of the pole are respectively a cylindrical portion and a head portion; the insulating sleeve is sheathed on the pole that penetrates through the stepped through hole, the head and cylindrical portions correspond to lower and upper end of the stepped through hole respectively; the battery winding core is arranged in an inner cavity formed by the upper cover plate and the bottom shell; and the sealing ball is arranged on the upper cover plate and seals the liquid injection hole.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training giant neural networks. One of the methods includes obtaining data specifying a partitioning of the neural network into N composite layers that form a sequence of composite layers, wherein each composite layer comprises a distinct plurality of layers from the multiple network layers of the neural network; obtaining data assigning each of the N composite layers to one or more computing devices from a set of N computing devices; partitioning a mini-batch of training examples into a plurality of micro-batches; and training the neural network, comprising: performing a forward pass through the neural network until output activations have been computed for each micro-batch for a final composite layer in the sequence, and performing a backward pass through the neural network until output gradients have been computed for each micro-batch for the first composite layer in the sequence.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training giant neural networks. One of the methods includes obtaining data specifying a partitioning of the neural network into N composite layers that form a sequence of composite layers, wherein each composite layer comprises a distinct plurality of layers from the multiple network layers of the neural network; obtaining data assigning each of the N composite layers to one or more computing devices from a set of N computing devices; partitioning a mini-batch of training examples into a plurality of micro-batches; and training the neural network, comprising: performing a forward pass through the neural network until output activations have been computed for each micro-batch for a final composite layer in the sequence, and performing a backward pass through the neural network until output gradients have been computed for each micro-batch for the first composite layer in the sequence.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: Compositions and methods are provided for the treatment and/or prevention of an inflammatory and/or autoimmune disease or disorder.

Abstract: Compositions and methods are provided for the treatment and/or prevention of an inflammatory and/or autoimmune disease or disorder.

Abstract: Disclosed are various embodiments for a search query segmentation application. Search queries are broken into segments. Each of the segments is assigned a taxonomy node from a catalog of items. Search results are generated as those items included in the taxonomy nodes assigned to the search query segments.

Abstract: A wireless ultrasonic scanning system comprises an ultrasonic sensor, a motor, an ultrasonic transceiver, a high-speed data sampling module, a motor controller and a master control module. The ultrasonic sensor, which moves in accordance with control of the motor controller, is mounted on the motor. The ultrasonic transceiver activates the ultrasonic sensor and amplifies a received ultrasonic signal. The high-speed data sampling module wirelessly transmits radio-frequency ultrasonic data to the master control module. The master control module sets a scanning mode, initiates a scanning process, and wirelessly transmits control signals and control parameters to the high-speed data sampling module.