Efficient Neural Architecture Search via Parameter Sharing
ICML
Hieu Pham, Melody Y. Guan, Barret Zoph, Quoc V. Le, Jeff Dean
We propose Efficient Neural Architecture Search (ENAS), a fast and inexpensive approach for automatic model design. In ENAS, a controller learns to discover neural network architectures by searching for an optimal subgraph within a large computational graph. The controller is trained with policy gradient to select a subgraph that maximizes the expected reward on the validation set. Meanwhile the model corresponding to the selected subgraph is trained to minimize a canonical cross entropy loss. Thanks to parameter sharing between child models, ENAS is fast: it delivers strong empirical performances using much fewer GPU-hours than all existing automatic model design approaches, and notably, 1000x less expensive than standard Neural Architecture Search. On the Penn Treebank dataset, ENAS discovers a novel architecture that achieves a test perplexity of 55.8, establishing a new state-of-the-art among all methods without post-training processing. On the CIFAR-10 dataset, ENAS designs novel architectures that achieve a test error of 2.89%, which is on par with NASNet (Zoph et al., 2018), whose test error is 2.65%.
@article{DBLP:journals/corr/abs-1802-03268, author = {Hieu Pham and Melody Y. Guan and Barret Zoph and Quoc V. Le and Jeff Dean}, title = {Efficient Neural Architecture Search via Parameter Sharing}, journal = {CoRR}, volume = {abs/1802.03268}, year = {2018}, url = {http://arxiv.org/abs/1802.03268}, archivePrefix = {arXiv}, eprint = {1802.03268}, timestamp = {Mon, 13 Aug 2018 16:47:58 +0200}, biburl = {https://dblp.org/rec/journals/corr/abs-1802-03268.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} }