Sample-to-Sample Correspondence for Unsupervised Domain Adaptation

Debasmit Das, C. S. George Lee

The assumption that training and testing samples are generated from the same distribution does not always hold for real-world machine-learning applications. The procedure of tackli…