When NAS Meets Trees: An Efficient Algorithm for Neural Architecture Search
This work addresses the efficiency problem in NAS for machine learning researchers and practitioners, offering a novel approach that is incremental but provides strong specific gains.
The paper tackles the challenge of efficiently exploring the large search space in neural architecture search (NAS) by proposing TNAS, a method that uses architecture and binary operation trees to reduce exploration size, achieving a test accuracy of 94.37% on CIFAR-10 in four GPU hours and outperforming state-of-the-art methods.
The key challenge in neural architecture search (NAS) is designing how to explore wisely in the huge search space. We propose a new NAS method called TNAS (NAS with trees), which improves search efficiency by exploring only a small number of architectures while also achieving a higher search accuracy. TNAS introduces an architecture tree and a binary operation tree, to factorize the search space and substantially reduce the exploration size. TNAS performs a modified bi-level Breadth-First Search in the proposed trees to discover a high-performance architecture. Impressively, TNAS finds the global optimal architecture on CIFAR-10 with test accuracy of 94.37\% in four GPU hours in NAS-Bench-201. The average test accuracy is 94.35\%, which outperforms the state-of-the-art. Code is available at: \url{https://github.com/guochengqian/TNAS}.