Supervised Domain Adaptation: A Graph Embedding Perspective and a Rectified Experimental Protocol
This work addresses reproducibility and benchmarking problems for researchers in domain adaptation, but it is incremental as it refines existing methods rather than introducing new ones.
The paper shows that three existing supervised domain adaptation methods can be reformulated as graph embedding problems, and it identifies reproducibility issues in their evaluation, proposing a rectified protocol with updated benchmarks on datasets like Office31, Digits, and VisDA.
Domain Adaptation is the process of alleviating distribution gaps between data from different domains. In this paper, we show that Domain Adaptation methods using pair-wise relationships between source and target domain data can be formulated as a Graph Embedding in which the domain labels are incorporated into the structure of the intrinsic and penalty graphs. Specifically, we analyse the loss functions of three existing state-of-the-art Supervised Domain Adaptation methods and demonstrate that they perform Graph Embedding. Moreover, we highlight some generalisation and reproducibility issues related to the experimental setup commonly used to demonstrate the few-shot learning capabilities of these methods. To assess and compare Supervised Domain Adaptation methods accurately, we propose a rectified evaluation protocol, and report updated benchmarks on the standard datasets Office31 (Amazon, DSLR, and Webcam), Digits (MNIST, USPS, SVHN, and MNIST-M) and VisDA (Synthetic, Real).