Learning Noise Transition Matrix from Only Noisy Labels
via Total Variation Regularization
Yivan Zhang 1 2 Gang Niu 2 Masashi Sugiyama 2 1
Abstract Background. Early studies on learning from noisy labels
can be traced back to the random classification noise (RCN)
Many weakly supervised classification methods model for binary classification (Angluin & Laird, 1988;
employ a nois ...
附件列表