An empirical study of compound PCFGs

Open Access
Authors
Publication date 2021
Host editors
  • E. Ben-David
  • S. Cohen
  • R. McDonald
  • B. Plank
  • R. Reichart
  • G. Rotman
  • Y.. Ziser
Book title The Second Workshop on Domain Adaptation for NLP
Book subtitle Adap-NLP 2021 : Proceedings of the Workshop : April 20, 2021
ISBN (electronic)
  • 9781954085084
Event 2nd Workshop on Domain Adaptation for NLP, Adapt-NLP 2021
Pages (from-to) 166-171
Number of pages 6
Publisher Stroudsburg, PA: Association for Computational Linguistics
Organisations
  • Interfacultary Research - Institute for Logic, Language and Computation (ILLC)
Abstract

Compound probabilistic context-free grammars (C-PCFGs) have recently established a new state of the art for phrase-structure grammar induction. However, due to the high time-complexity of chart-based representation and inference, it is difficult to investigate them comprehensively. In this work, we rely on a fast implementation of C-PCFGs to conduct evaluation complementary to that of Kim et al. (2019). We highlight three key findings: (1) C-PCFGs are data-efficient, (2) C-PCFGs make the best use of global sentence-level information in preterminal rule probabilities, and (3) the best configurations of C-PCFGs on English do not always generalize to morphology-rich languages.

Document type Conference contribution
Language English
Published at https://aclanthology.org/2021.adaptnlp-1.17
Other links https://github.com/zhaoyanpeng/xcfg https://www.scopus.com/pages/publications/85115691535
Downloads
2021.adaptnlp-1.17 (Final published version)
Permalink to this page
Back