Adaptive Stochastic Natural Gradient Method for One-Shot Neural Architecture Search

TL;DR

This paper introduces an adaptive stochastic natural gradient method for neural architecture search that is fast, robust, and broadly applicable, enabling simultaneous optimization of weights and architecture without extensive tuning.

Contribution

It develops a generic, differentiable NAS framework using stochastic relaxation and proposes a novel adaptive stochastic natural gradient method with theoretical backing.

Findings

Achieved near state-of-the-art performance on image classification.

Demonstrated robustness across different search spaces.

Operated efficiently with low computational budgets.

Abstract

High sensitivity of neural architecture search (NAS) methods against their input such as step-size (i.e., learning rate) and search space prevents practitioners from applying them out-of-the-box to their own problems, albeit its purpose is to automate a part of tuning process. Aiming at a fast, robust, and widely-applicable NAS, we develop a generic optimization framework for NAS. We turn a coupled optimization of connection weights and neural architecture into a differentiable optimization by means of stochastic relaxation. It accepts arbitrary search space (widely-applicable) and enables to employ a gradient-based simultaneous optimization of weights and architecture (fast). We propose a stochastic natural gradient method with an adaptive step-size mechanism built upon our theoretical investigation (robust). Despite its simplicity and no problem-dependent parameter tuning, our method exhibited near state-of-the-art performances with low computational budgets both on image classification and inpainting tasks.