Discriminative Log-Linear Grammars with Latent Variables

We demonstrate that log-linear grammars with latent variables can be practically trained using discriminative methods. Central to efficient discriminative training is a hierarchical pruning procedure which allows feature expectations to be efficiently approximated in a gradient-based procedure. We compare L1 and L2 regularization and show that L1 regularization is superior, requiring fewer iterations to converge, and yielding sparser solutions. On full-scale treebank parsing experiments, the discriminative latent models outperform both the comparable generative latent models as well as the discriminative non-latent baselines.

• Type: Conference
• Year: 2007
• Where: NIPS
• Authors: Slav Petrov, Dan Klein
• Keywords: Discriminative Latent Models, Discriminative Methods, Efficient Discriminative Training, Information Technology, NIPS 2007