Search Space Adaptation for Differentiable Neural Architecture Search in Image Classification
This work addresses the challenge of optimizing search spaces in NAS for image classification, but it appears incremental as it builds on existing methods like ProxylessNAS.
The paper tackles the problem of improving neural architecture search (NAS) performance by proposing a search space adaptation scheme with a search scope, and demonstrates its effectiveness using ProxylessNAS for image classification, providing insights through visualization of architecture parameter updates.
As deep neural networks achieve unprecedented performance in various tasks, neural architecture search (NAS), a research field for designing neural network architectures with automated processes, is actively underway. More recently, differentiable NAS has a great impact by reducing the search cost to the level of training a single network. Besides, the search space that defines candidate architectures to be searched directly affects the performance of the final architecture. In this paper, we propose an adaptation scheme of the search space by introducing a search scope. The effectiveness of proposed method is demonstrated with ProxylessNAS for the image classification task. Furthermore, we visualize the trajectory of architecture parameter updates and provide insights to improve the architecture search.