Self-Supervised Neural Architecture Search for Imbalanced Datasets

Abstract

Neural Architecture Search (NAS) provides stateof-the-art results when trained on well-curated datasets with annotated labels. However, annotating data or even having balanced number of samples can be a luxury for practitioners from different scientific fields, e.g., in the medical domain.To that end, we propose a NAS-based framework that bears the threefold contributions: (a) we focus on the self-supervised scenario, i.e., where no labels are required to determine the architecture, and (b) we assume the datasets are imbalanced, (c) we design each component to be able to run on a resource constrained setup, i.e., on a single GPU (e.g. Google Colab). Our components build on top of recent developments in self-supervised learning (Zbontar et al., 2021), self-supervised NAS (Kaplan & Giryes, 2020) and extend them for the case of imbalanced datasets. We conduct experiments on an (artificially) imbalanced version of CIFAR-10 and we demonstrate our proposed method outperforms standard neural networks, while using 27× less parameters. To validate our assumption on a naturally imbalanced dataset, we also conduct experiments on ChestM-NIST and COVID-19 X-ray. The results demonstrate how the proposed method can be used in imbalanced datasets, while it can be fully run on a single GPU. Code is available here. Preliminaries2.1. Neural Architecture Search Neural Architecture Search (NAS) can be roughly separated into three components (Elsken et al., 2019): search space, search strategy, and performance estimation strategy. The first component defines the set of architectures to be