More data means less inference: A pseudo-max approach to structured learning

Open in new window