Feature-Critic Networks for Heterogeneous Domain Generalization
This addresses the domain shift issue for machine learning models, enabling better generalization to unseen domains without target data, though it appears incremental as it builds on existing domain generalization techniques.
The paper tackles the problem of domain generalization, where models degrade on unseen domains, by proposing a learning-to-learn approach that learns an auxiliary loss for generalization, and it outperforms state-of-the-art methods in both conventional and heterogeneous settings.
The well known domain shift issue causes model performance to degrade when deployed to a new target domain with different statistics to training. Domain adaptation techniques alleviate this, but need some instances from the target domain to drive adaptation. Domain generalisation is the recently topical problem of learning a model that generalises to unseen domains out of the box, and various approaches aim to train a domain-invariant feature extractor, typically by adding some manually designed losses. In this work, we propose a learning to learn approach, where the auxiliary loss that helps generalisation is itself learned. Beyond conventional domain generalisation, we consider a more challenging setting of heterogeneous domain generalisation, where the unseen domains do not share label space with the seen ones, and the goal is to train a feature representation that is useful off-the-shelf for novel data and novel categories. Experimental evaluation demonstrates that our method outperforms state-of-the-art solutions in both settings.