SimpleDG: Simple Domain Generalization Baseline without Bells and Whistles
This work addresses domain generalization for machine learning practitioners, but it is incremental as it builds on existing baselines with minor enhancements.
The paper tackles domain generalization by proposing SimpleDG, a baseline method that achieved second place in two tracks of the NICO CHALLENGE 2022, building on findings that ERM is strong and adding simple designs to boost performance.
We present a simple domain generalization baseline, which wins second place in both the common context generalization track and the hybrid context generalization track respectively in NICO CHALLENGE 2022. We verify the founding in recent literature, domainbed, that ERM is a strong baseline compared to recent state-of-the-art domain generalization methods and propose SimpleDG which includes several simple yet effective designs that further boost generalization performance. Code is available at https://github.com/megvii-research/SimpleDG