MDA learns a domain-invariant feature transformation minimizing within-class domain divergence while maximizing class separability and compactness, with learning-theoretic bounds on excess risk and generalization error.
Geodesic flow kernel for unsupervised domain adapta- tion
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Domain Generalization via Multidomain Discriminant Analysis
MDA learns a domain-invariant feature transformation minimizing within-class domain divergence while maximizing class separability and compactness, with learning-theoretic bounds on excess risk and generalization error.