A General Upper Bound for Unsupervised Domain Adaptation

Dexuan Zhang, Tatsuya Harada

In this work, we present a novel upper bound of target error to address the problem for unsupervised domain adaptation. Recent studies reveal that a deep neural network can learn t…