Neural Architecture Search via Combinatorial Multi-Armed Bandit
This work significantly improves the search efficiency and accuracy of tree-search methods for Neural Architecture Search, making them more competitive for researchers and practitioners in deep learning.
This paper formulates Neural Architecture Search (NAS) as a Combinatorial Multi-Armed Bandit (CMAB) problem, enabling the decomposition of large search spaces. Using Nested Monte-Carlo Search, their approach achieves a state-of-the-art comparable error rate on CIFAR-10 in 0.58 GPU days, which is 20 times faster than existing tree-search methods.
Neural Architecture Search (NAS) has gained significant popularity as an effective tool for designing high performance deep neural networks (DNNs). NAS can be performed via policy gradient, evolutionary algorithms, differentiable architecture search or tree-search methods. While significant progress has been made for both policy gradient and differentiable architecture search, tree-search methods have so far failed to achieve comparable accuracy or search efficiency. In this paper, we formulate NAS as a Combinatorial Multi-Armed Bandit (CMAB) problem (CMAB-NAS). This allows the decomposition of a large search space into smaller blocks where tree-search methods can be applied more effectively and efficiently. We further leverage a tree-based method called Nested Monte-Carlo Search to tackle the CMAB-NAS problem. On CIFAR-10, our approach discovers a cell structure that achieves a low error rate that is comparable to the state-of-the-art, using only 0.58 GPU days, which is 20 times faster than current tree-search methods. Moreover, the discovered structure transfers well to large-scale datasets such as ImageNet.