The use of sparse neural networks has seen rapid growth in recent years,...
We investigate the optimal model size and number of tokens for training ...
The performance of a language model has been shown to be effectively mod...
In this paper we propose a new generative model of text, Step-unrolled
D...
We enhance auto-regressive language models by conditioning on document c...
Sparse neural networks are becoming increasingly important as the field ...
It has long been argued that minibatch stochastic gradient descent can
g...
Scientific workloads have traditionally exploited high levels of sparsit...
Neural networks have historically been built layerwise from the set of
f...
Current methods for training recurrent neural networks are based on
back...
Modern text-to-speech synthesis pipelines typically involve multiple
pro...
Sparse neural networks have been shown to be more parameter and compute
...
Historically, the pursuit of efficient inference has been one of the dri...
Generative adversarial networks have seen rapid development in recent ye...
We investigate the difficulties of training sparse neural networks and m...
In this work we show that Evolution Strategies (ES) are a viable method ...
We rigorously evaluate three state-of-the-art techniques for inducing
sp...
Generating musical audio directly with neural networks is notoriously
di...
Sequential models achieve state-of-the-art results in audio, visual and
...
The recently-developed WaveNet architecture is the current state of the ...
We consider the problem of transcribing polyphonic piano music with an
e...
Deep neural networks have enabled progress in a wide variety of applicat...
Recurrent Neural Networks (RNN) are widely used to solve a variety of
pr...
Modern deep neural networks have a large number of parameters, making th...
We show that an end-to-end deep learning approach can be used to recogni...
We present a state-of-the-art speech recognition system developed using
...