Unsupervised Domain Adaptation Architecture Search with Self-Training for Land Cover Mapping

TL;DR

This paper introduces a novel framework combining neural architecture search and self-training for unsupervised domain adaptation in land cover mapping, resulting in lightweight models with state-of-the-art accuracy on remote sensing datasets.

Contribution

It is the first to integrate NAS with self-training UDA specifically for land cover mapping, focusing on resource-efficient neural networks.

Findings

Lightweight networks with less than 2M parameters achieve state-of-the-art performance.

The proposed method outperforms existing approaches on OpenEarthMap and FLAIR #1 datasets.

Self-training with energy-based pseudo-labeling improves adaptation accuracy.

Abstract

Unsupervised domain adaptation (UDA) is a challenging open problem in land cover mapping. Previous studies show encouraging progress in addressing cross-domain distribution shifts on remote sensing benchmarks for land cover mapping. The existing works are mainly built on large neural network architectures, which makes them resource-hungry systems, limiting their practical impact for many real-world applications in resource-constrained environments. Thus, we proposed a simple yet effective framework to search for lightweight neural networks automatically for land cover mapping tasks under domain shifts. This is achieved by integrating Markov random field neural architecture search (MRF-NAS) into a self-training UDA framework to search for efficient and effective networks under a limited computation budget. This is the first attempt to combine NAS with self-training UDA as a single framework for land cover mapping. We also investigate two different pseudo-labelling approaches (confidence-based and energy-based) in self-training scheme. Experimental results on two recent datasets (OpenEarthMap & FLAIR #1) for remote sensing UDA demonstrate a satisfactory performance. With only less than 2M parameters and 30.16 GFLOPs, the best-discovered lightweight network reaches state-of-the-art performance on the regional target domain of OpenEarthMap (59.38% mIoU) and the considered target domain of FLAIR #1 (51.19% mIoU). The code is at https://github.com/cliffbb/UDA-NAS}{https://github.com/cliffbb/UDA-NAS.