Learning with Maximum A-Posteriori Perturbation Models.

Andreea Gane, Tamir Hazan, Tommi S. Jaakkola

Browse the full AISTATS paper archive.