Neural networks are powerful and flexible models that work well for many difficult
learning tasks in image, speech and natural language understanding. Despite
their success, neural networks are still hard to design. In this paper, we use a recurrent network to generate the model descriptions of neural networks and train
this RNN with reinforcement learning to maximize the expected accuracy of the
generated architectures on a validation set. On the CIFAR-10 dataset, our method,
starting from scratch, can design a novel network architecture that rivals the best
human-invented architecture in terms of test set accuracy. Our CIFAR-10 model
achieves a test error rate of 3:65, which is 0:09 percent better and 1.05x faster than
the previous state-of-the-art model that used a similar architectural scheme. On
the Penn Treebank dataset, our model can compose a novel recurrent cell that outperforms the widely-used LSTM cell, and other state-of-the-art baselines. Our cell
achieves a test set perplexity of 62.4 on the Penn Treebank, which is 3.6 perplexity
better than the previous state-of-the-art model. The cell can also be transferred
to the character language modeling task on PTB and achieves a state-of-the-art
perplexity of 1.214.
%0 Journal Article
%1 journals/corr/ZophL16
%A Zoph, Barret
%A Le, Quoc V.
%D 2016
%J CoRR
%K AI NeuralNetwork
%T Neural Architecture Search with Reinforcement Learning.
%U http://dblp.uni-trier.de/db/journals/corr/corr1611.html#ZophL16
%V abs/1611.01578
%X Neural networks are powerful and flexible models that work well for many difficult
learning tasks in image, speech and natural language understanding. Despite
their success, neural networks are still hard to design. In this paper, we use a recurrent network to generate the model descriptions of neural networks and train
this RNN with reinforcement learning to maximize the expected accuracy of the
generated architectures on a validation set. On the CIFAR-10 dataset, our method,
starting from scratch, can design a novel network architecture that rivals the best
human-invented architecture in terms of test set accuracy. Our CIFAR-10 model
achieves a test error rate of 3:65, which is 0:09 percent better and 1.05x faster than
the previous state-of-the-art model that used a similar architectural scheme. On
the Penn Treebank dataset, our model can compose a novel recurrent cell that outperforms the widely-used LSTM cell, and other state-of-the-art baselines. Our cell
achieves a test set perplexity of 62.4 on the Penn Treebank, which is 3.6 perplexity
better than the previous state-of-the-art model. The cell can also be transferred
to the character language modeling task on PTB and achieves a state-of-the-art
perplexity of 1.214.
@article{journals/corr/ZophL16,
abstract = {Neural networks are powerful and flexible models that work well for many difficult
learning tasks in image, speech and natural language understanding. Despite
their success, neural networks are still hard to design. In this paper, we use a recurrent network to generate the model descriptions of neural networks and train
this RNN with reinforcement learning to maximize the expected accuracy of the
generated architectures on a validation set. On the CIFAR-10 dataset, our method,
starting from scratch, can design a novel network architecture that rivals the best
human-invented architecture in terms of test set accuracy. Our CIFAR-10 model
achieves a test error rate of 3:65, which is 0:09 percent better and 1.05x faster than
the previous state-of-the-art model that used a similar architectural scheme. On
the Penn Treebank dataset, our model can compose a novel recurrent cell that outperforms the widely-used LSTM cell, and other state-of-the-art baselines. Our cell
achieves a test set perplexity of 62.4 on the Penn Treebank, which is 3.6 perplexity
better than the previous state-of-the-art model. The cell can also be transferred
to the character language modeling task on PTB and achieves a state-of-the-art
perplexity of 1.214.},
added-at = {2017-03-24T12:15:28.000+0100},
author = {Zoph, Barret and Le, Quoc V.},
biburl = {https://www.bibsonomy.org/bibtex/2da12131341e17fd937ce03a11bf27f33/ristephens},
ee = {http://arxiv.org/abs/1611.01578},
interhash = {39ee88dbc3d50930a602050b3684b6d6},
intrahash = {da12131341e17fd937ce03a11bf27f33},
journal = {CoRR},
keywords = {AI NeuralNetwork},
timestamp = {2017-04-04T09:30:04.000+0200},
title = {Neural Architecture Search with Reinforcement Learning.},
url = {http://dblp.uni-trier.de/db/journals/corr/corr1611.html#ZophL16},
volume = {abs/1611.01578},
year = 2016
}