Corruption is frequently observed in collected data and has been extensi...
Distribution shift (DS) may have two levels: the distribution itself cha...
Recent years have witnessed a great success of supervised deep learning,...
Supervised federated learning (FL) enables multiple clients to share the...
A key assumption in supervised learning is that training and test data f...
To cope with high annotation costs, training a classifier only from weak...
Ordinary (pointwise) binary classification aims to learn a binary classi...
A default assumption in many machine learning scenarios is that the trai...
Under distribution shift (DS) where the training data distribution diffe...
From two unlabeled (U) datasets with different class priors, we can trai...
Empirical risk minimization (ERM), with proper loss function and
regular...