Unifying Non-Maximum Likelihood Learning Objectives with Minimum KL Contraction

Open in new window