Convex structure learning in log-linear models beyond pairwise potentials

Previous work has examined structure learning in log-linear models with L1-regularization, largely focusing on the case of pairwise potentials. In this work we consider the case of models with potenti