Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating a target sequence from a source sequence. In one aspect, the system includes a recurrent neural network configured to, at each time step, receive an input for the time step and process the input to generate a progress score and a set of output scores; and a subsystem configured to, at each time step, generate the recurrent neural network input and provide the input to the recurrent neural network; determine, from the progress score, whether or not to emit a new output at the time step; and, in response to determining to emit a new output, select an output using the output scores and emit the selected output as the output at a next position in the output order.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for sample-efficient reinforcement learning. One of the methods includes maintaining an ensemble of Q networks, an ensemble of transition models, and an ensemble of reward models; obtaining a transition; generating, using the ensemble of transition models, M trajectories; for each time step in each of the trajectories: generating, using the ensemble of reward models, N rewards for the time step, generating, using the ensemble of Q networks, L Q values for the time step, and determining, from the rewards, the Q values, and the training reward, L*N candidate target Q values for the trajectory and for the time step; for each of the time steps, combining the candidate target Q values; determining a final target Q value; and training at least one of the Q networks in the ensemble using the final target Q value.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating a target sequence from a source sequence. In one aspect, the system includes a recurrent neural network configured to, at each time step, receive an input for the time step and process the input to generate a progress score and a set of output scores; and a subsystem configured to, at each time step, generate the recurrent neural network input and provide the input to the recurrent neural network; determine, from the progress score, whether or not to emit a new output at the time step; and, in response to determining to emit a new output, select an output using the output scores and emit the selected output as the output at a next position in the output order.

Abstract: The present disclosure is directed to the generation of a compact artificial neural network by removing individual nodes from the artificial neural network. Individual nodes of the artificial neural network may be deactivated randomly and/or selectively during training of the artificial neural network. In some embodiments, a particular node may be randomly deactivated approximately half of the time during processing of a set of training data inputs. Based on the accuracy of the results obtained when the node is deactivated compared to the accuracy of the results obtained when the node is activated, an activation probability may be generated. Nodes can then be selectively removed from the artificial neural network based on the activation probability.

Abstract: Systems and methods for detecting similar audio being received by separate voice activated electronic devices, and ignoring those commands, is described herein. In some embodiments, a voice activated electronic device may be activated by a wakeword that is output by the additional electronic device, such as a television or radio, may capture audio of sound subsequently following the wakeword, and may send audio data representing the sound to a backend system. Upon receipt, the backend system may, in parallel to performing automated speech recognition processing to the audio data, generate a sound profile of the audio data, and may compare that sound profile to sound profiles of recently received audio data and/or flagged sound profiles. If the generated sound profile is determined to match another sound profiles, then the automated speech recognition processing may be stopped, and the voice activated electronic device may be instructed to return to a keyword spotting mode.