Deep Learning under Privileged Information Using Heteroscedastic Dropout

Open in new window