2020/02/11 by Aloïs Pourchot, Pourchot, Aloïs, Alexis Ducarouge +3 · 1 citation
Agricultural and Biological Sciences · Business, Management and Accounting · Psychology · #Computer Vision and Pattern Recognition (cs.CV) #Eating Disorders and Behaviors #FOS: Computer and information sciences #Food Waste Reduction and Sustainability #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Neural and Evolutionary Computing (cs.NE) #Service and Product Innovation
paper · pdf · doi:10.48550/arxiv.2002.04289
openalex publication_date 2020/02/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Weight-sharing (WS) has recently emerged as a paradigm to accelerate the automated search for efficient neural architectures, a process dubbed Neural Architecture Search (NAS). Although very appealing, this framework is not without drawbacks and several works have started to question its capabilities on small hand-crafted benchmarks. In this paper, we take advantage of the \nasbench dataset to challenge the efficiency of WS on a representative search space. By comparing a SOTA WS approach to a plain random search we show that, despite decent correlations between evaluations using weight-sharing and standalone ones, WS is only rarely significantly helpful to NAS. In particular we highlight the impact of the search space itself on the benefits.