Keywords: Domain Generalization, Data-Free Learning, Knowledge Distillation, Knowledge Amalgamation
TL;DR: We define and investigate the novel and practical problem setting of data-free domain generalization, and propose a first and strong approach for it, which outperforms ensemble and data-free knowledge distillation baselines.
Abstract: In this work, we investigate the unexplored intersection of domain generalization and data-free learning. In particular, we address the question: How can knowledge contained in models trained on different source data domains be merged into a single model that generalizes well to unseen target domains, in the absence of source and target domain data? Machine learning models that can cope with domain shift are essential for for real-world scenarios with often changing data distributions. Prior domain generalization methods typically rely on using source domain data, making them unsuitable for private decentralized data. We define the novel problem of Data-Free Domain Generalization (DFDG), a practical setting where models trained on the source domains separately are available instead of the original datasets, and investigate how to effectively solve the domain generalization problem in that case. We propose DEKAN, an approach that extracts and fuses domain-specific knowledge from the available teacher models into a student model robust to domain shift. Our empirical evaluation demonstrates the effectiveness of our method which achieves first state-of-the-art results in DFDG by significantly outperforming ensemble and data-free knowledge distillation baselines.