Joint distribution matching

Learning domain invariant representations by joint Wasserstein distance minimization

Domain shifts in the training data are common in practical applications of machine learning; they occur for instance when the data is …