LGCVNEMLFeb 11, 2020

To Share or Not To Share: A Comprehensive Appraisal of Weight-Sharing

arXiv:2002.04289v20.0022 citations
AI Analysis25

This work addresses the efficiency of weight-sharing for researchers and practitioners in automated neural architecture design, revealing it as incremental with limited practical advantages.

The paper investigates the effectiveness of weight-sharing in Neural Architecture Search (NAS) using the NASBench dataset, finding that despite decent correlations, weight-sharing rarely provides significant benefits compared to random search, with search space characteristics influencing outcomes.

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.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes