Guided Evolution for Neural Architecture Search

TL;DR

G-EA introduces a guided evolutionary approach to neural architecture search that efficiently explores the search space using a zero-proxy estimator, achieving state-of-the-art results on multiple benchmarks.

Contribution

The paper proposes G-EA, a novel guided evolutionary NAS method that evaluates architectures at initialization to improve search efficiency and performance.

Findings

Achieves state-of-the-art accuracy on NAS-Bench-201 benchmarks.

Effectively balances exploration and exploitation during search.

Demonstrates improved search efficiency without increased computation.

Abstract

Neural Architecture Search (NAS) methods have been successfully applied to image tasks with excellent results. However, NAS methods are often complex and tend to converge to local minima as soon as generated architectures seem to yield good results. In this paper, we propose G-EA, a novel approach for guided evolutionary NAS. The rationale behind G-EA, is to explore the search space by generating and evaluating several architectures in each generation at initialization stage using a zero-proxy estimator, where only the highest-scoring network is trained and kept for the next generation. This evaluation at initialization stage allows continuous extraction of knowledge from the search space without increasing computation, thus allowing the search to be efficiently guided. Moreover, G-EA forces exploitation of the most performant networks by descendant generation while at the same time forcing exploration by parent mutation and by favouring younger architectures to the detriment of older ones. Experimental results demonstrate the effectiveness of the proposed method, showing that G-EA achieves state-of-the-art results in NAS-Bench-201 search space in CIFAR-10, CIFAR-100 and ImageNet16-120, with mean accuracies of 93.98%, 72.12% and 45.94% respectively.