Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating an output sequence of audio data that comprises a respective audio sample at each of a plurality of time steps. One of the methods includes, for each of the time steps: providing a current sequence of audio data as input to a convolutional subnetwork, wherein the current sequence comprises the respective audio sample at each time step that precedes the time step in the output sequence, and wherein the convolutional subnetwork is configured to process the current sequence of audio data to generate an alternative representation for the time step; and providing the alternative representation for the time step as input to an output layer, wherein the output layer is configured to: process the alternative representation to generate an output that defines a score distribution over a plurality of possible audio samples for the time step.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: An off-policy reinforcement learning actor-critic neural network system configured to select actions from a continuous action space to be performed by an agent interacting with an environment to perform a task. An observation defines environment state data and reward data. The system has an actor neural network which learns a policy function mapping the state data to action data. A critic neural network learns an action-value (Q) function. A replay buffer stores tuples of the state data, the action data, the reward data and new state data. The replay buffer also includes demonstration transition data comprising a set of the tuples from a demonstration of the task within the environment. The neural network system is configured to train the actor neural network and the critic neural network off-policy using stored tuples from the replay buffer comprising tuples both from operation of the system and from the demonstration transition data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for assigning operations of a computational graph to a plurality of computing devices are disclosed. Data characterizing a computational graph is obtained. Context information for a computational environment in which to perform the operations of the computational graph is received. A model input is generated, which includes at least the context information and the data characterizing the computational graph. The model input is processed using the machine learning model to generate an output defining placement assignments of the operations of the computational graph to the plurality of computing devices. The operations of the computational graph are assigned to the plurality of computing device according to the defined placement assignments.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating output examples using neural networks. One of the methods includes receiving a request to generate an output example of a particular type, accessing dependency data, and generating the output example by, at each of a plurality of generation time steps: identifying one or more current blocks for the generation time step, wherein each current block is a block for which the values of the bits in all of the other blocks identified in the dependency for the block have already been generated; and generating the values of the bits in the current blocks for the generation time step conditioned on, for each current block, the already generated values of the bits in the other blocks identified in the dependency for the current block.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a text classification machine learning model. One of the methods includes training a model having a plurality of parameters and configured to generate a classification of a text sample comprising a plurality of words by processing a model input that includes a combined feature representation of the plurality of words in the text sample, wherein the training comprises receiving a text sample and a target classification for the text sample; generating a plurality of perturbed combined feature representations; determining, based on the plurality of perturbed combined feature representations, a region in the embedding space; and determining an update to the parameters based on an adversarial objective that encourages the model to assign the target classification for the text sample for all of the combined feature representations in the region in the embedding space.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for a neural network system comprising one or more gated linear networks. A system includes: one or more gated linear networks, wherein each gated linear network corresponds to a respective data value in an output data sample and is configured to generate a network probability output that defines a probability distribution over possible values for the corresponding data value, wherein each gated linear network comprises a plurality of layers, wherein the plurality of layers comprises a plurality of gated linear layers, wherein each gated linear layer has one or more nodes, and wherein each node is configured to: receive a plurality of inputs, receive side information for the node; combine the plurality of inputs according to a set of weights defined by the side information, and generate and output a node probability output for the corresponding data value.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for reinforcement learning. One of the methods includes selecting an action to be performed by the agent using both a slow updating recurrent neural network and a fast updating recurrent neural network that receives a fast updating input that includes the hidden state of the slow updating recurrent neural network.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a reinforcement learning system. The method includes: training an action selection policy neural network, and during the training of the action selection neural network, training one or more auxiliary control neural networks and a reward prediction neural network. Each of the auxiliary control neural networks is configured to receive a respective intermediate output generated by the action selection policy neural network and generate a policy output for a corresponding auxiliary control task. The reward prediction neural network is configured to receive one or more intermediate outputs generated by the action selection policy neural network and generate a corresponding predicted reward.

Abstract: We describe an artificial neural network comprising: an input layer of input neurons, one or more hidden layers of neurons in successive layers of neurons above the input layer, and at least one further, concept-identifying layer of neurons above the hidden layers. The neural network includes an activation memory coupled to an intermediate, hidden layer of neurons between the input concept-identifying layers to store a pattern of activation of the intermediate layer. The neural network further includes a system to determine an overlap between a plurality of the stored patterns of activation and to activate in the intermediate hidden layer an overlap pattern such that the concept-identifying layer of neurons is configured to identify features of the overlap patterns. We also describe related methods, processor control code, and computing systems for the neural network. Optionally further, higher level concept-identifying layers of neurons may be included.

Abstract: Methods, systems and apparatus, including computer programs encoded on computer storage media, for training an action selection neural network. One of the methods includes receiving an observation characterizing a current state of the environment; determining a target network output for the observation by performing a look ahead search of possible future states of the environment starting from the current state until the environment reaches a possible future state that satisfies one or more termination criteria, wherein the look ahead search is guided by the neural network in accordance with current values of the network parameters; selecting an action to be performed by the agent in response to the observation using the target network output generated by performing the look ahead search; and storing, in an exploration history data store, the target network output in association with the observation for use in updating the current values of the network parameters.

Abstract: A system including one or more computers and one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to implement a memory and memory-based neural network is described. The memory is configured to store a respective memory vector at each of a plurality of memory locations in the memory. The memory-based neural network is configured to: at each of a plurality of time steps: receive an input; determine an update to the memory, wherein determining the update comprising applying an attention mechanism over the memory vectors in the memory and the received input; update the memory using the determined update to the memory; and generate an output for the current time step using the updated memory.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for improving operational efficiency within a data center by modeling data center performance and predicting power usage efficiency. An example method receives a state input characterizing a current state of a data center. For each data center setting slate, the state input and the data center setting slate are processed through an ensemble of machine learning models. Each machine learning model is configured to receive and process the state input and the data center setting slate to generate an efficiency score that characterizes a predicted resource efficiency of the data center if the data center settings defined by the data center setting slate are adopted t. The method selects, based on the efficiency scores for the data center setting slates, new values for the data center settings.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a neural network. In one aspect, a method includes maintaining data specifying, for each of the network parameters, current values of a respective set of distribution parameters that define a posterior distribution over possible values for the network parameter. A respective current training value for each of the network parameters is determined from a respective temporary gradient value for the network parameter. The current values of the respective sets of distribution parameters for the network parameters are updated in accordance with the respective current training values for the network parameters. The trained values of the network parameters are determined based on the updated current values of the respective sets of distribution parameters.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for reinforcement learning. The embodiments described herein apply meta-learning (and in particular, meta-gradient reinforcement learning) to learn an optimum return function G so that the training of the system is improved. This provides a more effective and efficient means of training a reinforcement learning system as the system is able to converge on an optimum set of one or more policy parameters ? more quickly by training the return function G as it goes. In particular, the return function G is made dependent on the one or more policy parameters ? and a meta-objective function J? is used that is differentiated with respect to the one or more return parameters ? to improve the training of the return function G.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network to perform speech synthesis. One of the methods includes obtaining a training data set for training a first neural network to process a spectral representation of an audio sample and to generate a prediction of the audio sample, wherein, after training, the first neural network obtains spectral representations of audio samples from a second neural network; for a plurality of audio samples in the training data set: generating a ground-truth spectral representation of the audio sample; and processing the ground-truth spectral representation using a third neural network to generate an updated spectral representation of the audio sample; and training the first neural network using the updated spectral representations, wherein the third neural network is configured to generate updated spectral representations that resemble spectral representations generated by the second neural network.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing a machine learning task on a network input that is a sequence to generate a network output. In one aspect, one of the methods includes, for each particular sequence of layer inputs: for each attention layer in the neural network: maintaining episodic memory data; maintaining compressed memory data; receiving a layer input to be processed by the attention layer; and applying an attention mechanism over (i) the compressed representation in the compressed memory data for the layer, (ii) the hidden states in the episodic memory data for the layer, and (iii) the respective hidden state at each of the plurality of input positions in the particular network input to generate a respective activation for each input position in the layer input.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for neural programming. One of the methods includes processing a current neural network input using a core recurrent neural network to generate a neural network output; determining, from the neural network output, whether or not to end a currently invoked program and to return to a calling program from the set of programs; determining, from the neural network output, a next program to be called; determining, from the neural network output, contents of arguments to the next program to be called; receiving a representation of a current state of the environment; and generating a next neural network input from an embedding for the next program to be called and the representation of the current state of the environment.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training an actor neural network used to select actions to be performed by an agent interacting with an environment. One of the methods includes obtaining a minibatch of experience tuples; and updating current values of the parameters of the actor neural network, comprising: for each experience tuple in the minibatch: processing the training observation and the training action in the experience tuple using a critic neural network to determine a neural network output for the experience tuple, and determining a target neural network output for the experience tuple; updating current values of the parameters of the critic neural network using errors between the target neural network outputs and the neural network outputs; and updating the current values of the parameters of the actor neural network using the critic neural network.

Abstract: Methods, and systems, including computer programs encoded on computer storage media for generating data items. A method includes reading a glimpse from a data item using a decoder hidden state vector of a decoder for a preceding time step, providing, as input to a encoder, the glimpse and decoder hidden state vector for the preceding time step for processing, receiving, as output from the encoder, a generated encoder hidden state vector for the time step, generating a decoder input from the generated encoder hidden state vector, providing the decoder input to the decoder for processing, receiving, as output from the decoder, a generated a decoder hidden state vector for the time step, generating a neural network output update from the decoder hidden state vector for the time step, and combining the neural network output update with a current neural network output to generate an updated neural network output.