Reproduce “Open Set Domain Adaptation: Theoretical Bound and Algorithm” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Zhen Fang, Jie Lu, Feng Liu, Junyu Xuan, Guangquan Zhang

DAOD addresses open-set unsupervised domain adaptation by rejecting unknown target examples while aligning source data with known target examples. Its bound includes an open set difference, alongside source risk, distribution discrepancy, and joint error; the implemented objective uses projected MMD. Across 38 Office-31, Office-Home, and PIE tasks, DAOD averaged 69.3% Acc(OS), versus OSBP’s 65.9%, but led on 24 tasks.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.