Regularizing Differentiable Architecture Search with Smooth Activation
This addresses performance collapse problems in neural architecture search for researchers and practitioners, offering an incremental improvement over existing regularization techniques.
The paper tackles robustness and generalization issues in Differentiable Architecture Search (DARTS) by proposing Smooth Activation DARTS (SA-DARTS), which uses a smooth activation function as an auxiliary loss to mitigate skip dominance and achieve state-of-the-art results on NAS-Bench-201, classification, and super-resolution tasks.
Differentiable Architecture Search (DARTS) is an efficient Neural Architecture Search (NAS) method but suffers from robustness, generalization, and discrepancy issues. Many efforts have been made towards the performance collapse issue caused by skip dominance with various regularization techniques towards operation weights, path weights, noise injection, and super-network redesign. It had become questionable at a certain point if there could exist a better and more elegant way to retract the search to its intended goal -- NAS is a selection problem. In this paper, we undertake a simple but effective approach, named Smooth Activation DARTS (SA-DARTS), to overcome skip dominance and discretization discrepancy challenges. By leveraging a smooth activation function on architecture weights as an auxiliary loss, our SA-DARTS mitigates the unfair advantage of weight-free operations, converging to fanned-out architecture weight values, and can recover the search process from skip-dominance initialization. Through theoretical and empirical analysis, we demonstrate that the SA-DARTS can yield new state-of-the-art (SOTA) results on NAS-Bench-201, classification, and super-resolution. Further, we show that SA-DARTS can help improve the performance of SOTA models with fewer parameters, such as Information Multi-distillation Network on the super-resolution task.