Efficient Automation of Neural Network Design: A Survey on Differentiable Neural Architecture Search

TL;DR

This survey reviews recent advancements in Differentiable Neural Architecture Search (DNAS), highlighting its efficiency, novel classification taxonomy, and impact on automating neural network design, while discussing future research directions.

Contribution

It provides a comprehensive review of DNAS methods, introduces a new challenge-based taxonomy, and offers insights into future research avenues in the field.

Findings

DNAS is significantly faster and more resource-efficient than previous NAS methods.

Recent DNAS approaches have advanced the automation of neural network design.

The survey identifies key challenges and future directions for DNAS research.

Abstract

In the past few years, Differentiable Neural Architecture Search (DNAS) rapidly imposed itself as the trending approach to automate the discovery of deep neural network architectures. This rise is mainly due to the popularity of DARTS, one of the first major DNAS methods. In contrast with previous works based on Reinforcement Learning or Evolutionary Algorithms, DNAS is faster by several orders of magnitude and uses fewer computational resources. In this comprehensive survey, we focus specifically on DNAS and review recent approaches in this field. Furthermore, we propose a novel challenge-based taxonomy to classify DNAS methods. We also discuss the contributions brought to DNAS in the past few years and its impact on the global NAS field. Finally, we conclude by giving some insights into future research directions for the DNAS field.