TY - GEN
T1 - Domain-Invariant Feature Alignment Using Variational Inference for Partial Domain Adaptation
AU - Choudhuri, Sandipan
AU - Adeniye, Suli
AU - Sen, Arunabha
AU - Venkateswara, Hemanth
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - The standard closed-set domain adaptation approaches seek to mitigate distribution discrepancies between two domains under the constraint of both sharing identical label sets. However, in realistic scenarios, finding an optimal source domain with identical label space is a challenging task. Partial domain adaptation alleviates this problem of procuring a labeled dataset with identical label space assumptions and addresses a more practical scenario where the source label set subsumes the target label set. This, however, presents a few additional obstacles during adaptation. Samples with categories private to the source domain thwart relevant knowledge transfer and degrade model performance. In this work, we try to address these issues by coupling variational information and adversarial learning with a pseudo-labeling technique to enforce class distribution alignment and minimize the transfer of superfluous information from the source samples. The experimental findings in numerous cross-domain classification tasks demonstrate that the proposed technique delivers superior and comparable accuracy to existing methods.
AB - The standard closed-set domain adaptation approaches seek to mitigate distribution discrepancies between two domains under the constraint of both sharing identical label sets. However, in realistic scenarios, finding an optimal source domain with identical label space is a challenging task. Partial domain adaptation alleviates this problem of procuring a labeled dataset with identical label space assumptions and addresses a more practical scenario where the source label set subsumes the target label set. This, however, presents a few additional obstacles during adaptation. Samples with categories private to the source domain thwart relevant knowledge transfer and degrade model performance. In this work, we try to address these issues by coupling variational information and adversarial learning with a pseudo-labeling technique to enforce class distribution alignment and minimize the transfer of superfluous information from the source samples. The experimental findings in numerous cross-domain classification tasks demonstrate that the proposed technique delivers superior and comparable accuracy to existing methods.
UR - http://www.scopus.com/inward/record.url?scp=85150223435&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85150223435&partnerID=8YFLogxK
U2 - 10.1109/IEEECONF56349.2022.10052012
DO - 10.1109/IEEECONF56349.2022.10052012
M3 - Conference contribution
AN - SCOPUS:85150223435
T3 - Conference Record - Asilomar Conference on Signals, Systems and Computers
SP - 349
EP - 355
BT - 56th Asilomar Conference on Signals, Systems and Computers, ACSSC 2022
A2 - Matthews, Michael B.
PB - IEEE Computer Society
T2 - 56th Asilomar Conference on Signals, Systems and Computers, ACSSC 2022
Y2 - 31 October 2022 through 2 November 2022
ER -