TY - JOUR
T1 - Supervised Domain Adaptation: A Graph Embedding Perspective and a Rectified Experimental Protocol
AU - Hedegaard Morsing, Lukas
AU - Sheikh-Omar, Omar Ali
AU - Iosifidis, Alexandros
PY - 2021/10
Y1 - 2021/10
N2 - 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).
AB - 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).
KW - domain adaptation
KW - few-shot learning
KW - graph embedding
KW - supervised domain adaptation
KW - transfer learning
UR - http://www.scopus.com/inward/record.url?scp=85117745897&partnerID=8YFLogxK
U2 - 10.1109/TIP.2021.3118978
DO - 10.1109/TIP.2021.3118978
M3 - Journal article
C2 - 34648445
SN - 1057-7149
VL - 30
SP - 8619
EP - 8631
JO - IEEE Transactions on Image Processing
JF - IEEE Transactions on Image Processing
ER -