Description
DistanceNet is a learning algorithm for multi-source domain adaptation that uses various distance measures, or a mixture of these distance measures, as an additional loss function to be minimized jointly with the task's loss function, so as to achieve better unsupervised domain adaptation.