Domain Generalisation

Domain generalization (DG) aims to train machine learning models that generalize well to unseen data distributions, addressing the limitations of models trained on a single domain. Current research focuses on improving generalization performance with limited data, exploring techniques like meta-learning and adaptive task sampling to leverage existing knowledge effectively, and enhancing robustness against adversarial attacks by treating different attack types as separate domains. These advancements are crucial for deploying reliable machine learning systems in real-world applications where data variability is inevitable, impacting fields like computer vision and beyond.