OT Score: An OT based Confidence Score for Prototype-Assisted Source Free Unsupervised Domain Adaptation

TMLR Paper6988 Authors

Abstract: We address the computational and theoretical limitations of current distributional alignment methods for source-free unsupervised domain adaptation (SFUDA) using source class-mean features. In particular, we focus on estimating classification performance and confidence in the absence of target labels. Current theoretical frameworks for these methods often yield computationally intractable quantities and fail to adequately reflect the properties of the alignment algorithms employed. To overcome these challenges, we introduce the Optimal Transport (OT) score, a confidence metric derived from a novel theoretical analysis that exploits the flexibility of decision boundaries induced by Semi-Discrete Optimal Transport alignment. The proposed OT score is intuitively interpretable and theoretically rigorous. It provides principled uncertainty estimates for any given set of target pseudo-labels. Experimental results demonstrate that OT score outperforms existing confidence scores. Moreover, it improves SFUDA performance through training-time reweighting and provides a reliable, label-free proxy for model performance.

Submission Type: Regular submission (no more than 12 pages of main content)

Changes Since Last Submission:

We emphasize that our setting requires class mean feature beyond the source-free unsupervised domain adaptation setting.
We added missing mathematical notations.
Additional benchmarks (DomainNet-126) and comparisons with more recent leading methods are added.
Finally, we added a section 5.4 for sensitivity analysis of entropic regularization and dual optimization stability.

Assigned Action Editor: ~Sanghyuk_Chun1

Submission Number: 6988