Diversity in Spectral Learning for Natural Language Parsing
This provides incremental improvements in parsing accuracy for English and German, benefiting NLP applications.
The paper tackles the problem of improving natural language parsing by creating diverse predictions with spectral learning of latent-variable PCFGs, achieving F1 scores of 90.18 for English and 83.38 for German.
We describe an approach to create a diverse set of predictions with spectral learning of latent-variable PCFGs (L-PCFGs). Our approach works by creating multiple spectral models where noise is added to the underlying features in the training set before the estimation of each model. We describe three ways to decode with multiple models. In addition, we describe a simple variant of the spectral algorithm for L-PCFGs that is fast and leads to compact models. Our experiments for natural language parsing, for English and German, show that we get a significant improvement over baselines comparable to state of the art. For English, we achieve the $F_1$ score of 90.18, and for German we achieve the $F_1$ score of 83.38.