Sample-Efficient Neural Architecture Search by Learning Action Space for Monte Carlo Tree Search

1 Jan 2019  ·  Linnan Wang, Saining Xie, Teng Li, Rodrigo Fonseca, Yuandong Tian ·

Neural Architecture Search (NAS) has emerged as a promising technique for automatic neural network design. However, existing MCTS based NAS approaches often utilize manually designed action space, which is not directly related to the performance metric to be optimized (e.g., accuracy), leading to sample-inefficient explorations of architectures. To improve the sample efficiency, this paper proposes Latent Action Neural Architecture Search (LaNAS), which learns actions to recursively partition the search space into good or bad regions that contain networks with similar performance metrics. During the search phase, as different action sequences lead to regions with different performance, the search efficiency can be significantly improved by biasing towards the good regions. On three NAS tasks, empirical results demonstrate that LaNAS is at least an order more sample efficient than baseline methods including evolutionary algorithms, Bayesian optimizations and random search. When applied in practice, both one-shot and regular LaNAS consistently outperforms existing results. Particularly, LaNAS achieves 99.0\% accuracy on CIFAR-10 and 80.8\% top1 accuracy at 600 MFLOPS on ImageNet in only 800 samples, significantly outperforming AmoebaNet with $33\times$ fewer samples.

PDF

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Image Classification CIFAR-10 LaNet Percentage correct 99.03 # 18
PARAMS 44.1M # 227

Methods


No methods listed for this paper. Add relevant methods here