Moment Matching for Multi-Source Domain Adaptation
This work addresses the practical scenario of adapting models across multiple data sources, which is incremental but important for real-world applications like image recognition.
The authors tackled the problem of multi-source domain adaptation by introducing DomainNet, a large dataset with six domains and 0.6 million images, and proposed M3SDA, a method that aligns feature distribution moments to transfer knowledge from multiple labeled sources to an unlabeled target, achieving state-of-the-art performance in experiments.
Conventional unsupervised domain adaptation (UDA) assumes that training data are sampled from a single domain. This neglects the more practical scenario where training data are collected from multiple sources, requiring multi-source domain adaptation. We make three major contributions towards addressing this problem. First, we collect and annotate by far the largest UDA dataset, called DomainNet, which contains six domains and about 0.6 million images distributed among 345 categories, addressing the gap in data availability for multi-source UDA research. Second, we propose a new deep learning approach, Moment Matching for Multi-Source Domain Adaptation M3SDA, which aims to transfer knowledge learned from multiple labeled source domains to an unlabeled target domain by dynamically aligning moments of their feature distributions. Third, we provide new theoretical insights specifically for moment matching approaches in both single and multiple source domain adaptation. Extensive experiments are conducted to demonstrate the power of our new dataset in benchmarking state-of-the-art multi-source domain adaptation methods, as well as the advantage of our proposed model. Dataset and Code are available at \url{http://ai.bu.edu/M3SDA/}.