Effective Inference for Generative Neural Parsing

2017/07/27 by Mitchell Stern, Daniel Fried, Stern, Mitchell +3 · 3 citations
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Natural Language Processing Techniques #Speech Recognition and Synthesis #Topic Modeling

paper · pdf · doi:10.48550/arxiv.1707.08976

openalex publication_date 2017/07/27 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Generative neural models have recently achieved state-of-the-art results for constituency parsing. However, without a feasible search procedure, their use has so far been limited to reranking the output of external parsers in which decoding is more tractable. We describe an alternative to the conventional action-level beam search used for discriminative neural models that enables us to decode directly in these generative models. We then show that by improving our basic candidate selection strategy and using a coarse pruning function, we can improve accuracy while exploring significantly less of the search space. Applied to the model of Choe and Charniak (2016), our inference procedure obtains 92.56 F1 on section 23 of the Penn Treebank, surpassing prior state-of-the-art results for single-model systems.

Cited by

• A Systematic Study of Compositional Syntactic Transformer Language Models
• Representation of Constituents in Neural Language Models: Coordination Phrase as a Case Study
• Structural Supervision Improves Few-Shot Learning and Syntactic Generalization in Neural Language Models

Related

• Improving Neural Parsing by Disentangling Model Combination and Reranking Effects
• What's Going On in Neural Constituency Parsers? An Analysis
• A Minimal Span-Based Neural Constituency Parser
• Cross-Domain Generalization of Neural Constituency Parsers
• Neural Combinatory Constituency Parsing