Differential Evolution for Neural Architecture Search
This work provides an improved and more robust search strategy for researchers and practitioners working on neural architecture search, particularly for tabular benchmarks.
This paper introduces differential evolution as a search strategy for neural architecture search (NAS). When combined with full evaluations, it outperforms regularized evolution and Bayesian optimization across 13 tabular NAS benchmarks.
Neural architecture search (NAS) methods rely on a search strategy for deciding which architectures to evaluate next and a performance estimation strategy for assessing their performance (e.g., using full evaluations, multi-fidelity evaluations, or the one-shot model). In this paper, we focus on the search strategy. We introduce the simple yet powerful evolutionary algorithm of differential evolution to the NAS community. Using the simplest performance evaluation strategy of full evaluations, we comprehensively compare this search strategy to regularized evolution and Bayesian optimization and demonstrate that it yields improved and more robust results for 13 tabular NAS benchmarks based on NAS-Bench-101, NAS-Bench-1Shot1, NAS-Bench-201 and NAS-HPO bench.