Discretization-aware architecture search
作者：
Highlights：
• We propose discretization-aware architecture search (DA2S), and target at pushing the super-network towards the configuration of desired topol- ogy. DA2S is implemented with an entropy-based loss term, which can be regularized to differentiable architecture search in a plug-and-play fashion.
• The regularization for architecture search is controlled by elaborated continuation functions, so that discretization is adaptive to the dynamic change of edges and operations.
• Experiments on standard image classification benchmarks demonstrate the effectiveness of our approach, in particular, under imbalanced network configurations that were not studied before.
摘要
•We propose discretization-aware architecture search (DA2S), and target at pushing the super-network towards the configuration of desired topol- ogy. DA2S is implemented with an entropy-based loss term, which can be regularized to differentiable architecture search in a plug-and-play fashion.•The regularization for architecture search is controlled by elaborated continuation functions, so that discretization is adaptive to the dynamic change of edges and operations.•Experiments on standard image classification benchmarks demonstrate the effectiveness of our approach, in particular, under imbalanced network configurations that were not studied before.
论文关键词：Neural architecture search,Weight-sharing,Discretization-aware,Imbalanced network configuration
论文评审过程：Received 7 December 2020, Revised 4 July 2021, Accepted 14 July 2021, Available online 22 July 2021, Version of Record 28 July 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2021.108186
