Better generalization with less data using robust gradient descent.

Matthew J. Holland, Kazushi Ikeda

Browse the full ICML paper archive.