Domain Generalization Under Posterior Drift

Yilun Zhu, Naihao Deng, Naichen Shi +2

Domain generalization (DG) is the problem of generalizing from several distributions (or domains), for which labeled training data are available, to a new test domain for which no…