Learning Maximum-A-Posteriori Perturbation Models for Structured Prediction in Polynomial Time.

Asish Ghoshal, Jean Honorio

Browse the full ICML paper archive.