pyBART: Evidence-based Syntactic Transformations for IE

What it is about

pyBART is an open-source Python library that converts English UD trees into a representation (BART) that makes event structure and many content-word relations explicit, going beyond Enhanced UD. The relations were chosen by comparing paths between annotated entities in ACE05 and QA-SRL with the shorter paths that the corpora imply, so each added edge has evidence behind it. Edges that are plausible but uncertain are kept and marked (ALT and UNC flags) rather than dropped. Copulas get a STATE node, and finite auxiliaries and evidential or aspectual verbs become event modifiers (ev). On a pattern-based relation extraction task on TACRED, BART reaches F1 49.15, against 43.77 for UD and 45.69 for Enhanced UD, and it reaches 30.7% recall with 52 patterns, where UD needs 112.

Ideas

• I1. prag: add a STATE-style event node for copular clauses and an ev-style modifier for evidential and aspectual verbs, and measure the number of event frames and their precision on EWT dev against a hand check of 100 sentences.
• I2. Uncertainty flags in MISC: mark the propagated edges that the paper calls uncertain (nested subjects, reduced relatives, shared modifiers), and report precision and recall of world facts with and without them on a relation dev set from annot.
• I3. world induce: report the pattern-economy curve (recall against number of induced patterns), as the paper does, for the rules learned from Mova trees, and compare gold and en trees on dev.
• I4. Use the paper's filter, which drops any pattern that lowers dev F1 for its relation, as a gate in world induce; measure test F1 once per release.

For Mova

The paper's view that finite verbs can be event modifiers fits the event structure that prag and world already build. The ideas touch 21. The simple clause, 23-complementation.md (control) and 20-syntax-overview.md. The TACRED data is LDC-licensed, so it is used for test runs only and never shipped.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/2020.acl-demos.7 · arXiv 2005.01306

@inproceedings{tiktinsky2020pybart,
  title = {{pyBART: Evidence-based Syntactic Transformations for IE}},
  author = {Tiktinsky, Aryeh and Goldberg, Yoav and Tsarfaty, Reut},
  booktitle = {Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations},
  year = {2020},
  doi = {10.18653/v1/2020.acl-demos.7},
  eprint = {2005.01306},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.18653/v1/2020.acl-demos.7}
}

References

18 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Omri Abend and Ari Rappoport. 2013. Universal conceptual cognitive annotation (ucca). In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 228–238. find our card
2. Martha Palmer, Daniel Gildea, and Nianwen Xue. 2010. Semantic role labeling. Synthesis Lectures on Human Language Technologies, 3(1):1–103. find
3. Laura Banarescu, Claire Bonial, Shu Cai, Madalina Georgescu, Kira Griffitt, Ulf Hermjakob, Kevin Knight, Philipp Koehn, Martha Palmer, and Nathan Schneider. 2013. Abstract meaning representation for sembanking. In Proceedings of the 7th linguistic annotation workshop and interoperability with discourse, pages 178–186. find our card
4. Siva Reddy, Oscar Täckström, Michael Collins, Tom Kwiatkowski, Dipanjan Das, Mark Steedman, and Mirella Lapata. 2016. Transforming dependency structures to logical forms for semantic parsing. Transactions of the Association for Computational Linguistics, 4:127–140. find
5. Marie Candito, Bruno Guillaume, Guy Perrier, and Djamé Seddah. 2017. Enhanced UD dependencies with neutralized diathesis alternation. In Proceedings of the Fourth International Conference on Dependency Linguistics (Depling 2017), pages 42–53, Pisa,Italy. Linköping University Electronic Press. find our card
6. Siva Reddy, Oscar Tckstrm, Slav Petrov, Mark Steedman, and Mirella Lapata. 2017. Universal semantic parsing. find
7. Rachel Rudinger and Benjamin Van Durme. 2014. Is the Stanford dependency representation semantic? In Proceedings of the Second Workshop on EVENTS: Definition, Detection, Coreference, and Representation, pages 54–58, Baltimore, Maryland, USA. Association for Computational Linguistics. find
8. Marie-Catherine De Marneffe and Christopher D Manning. 2008. Stanford typed dependencies manual. Technical report, Technical report, Stanford University. find
9. Sebastian Schuster and Christopher D Manning. 2016. Enhanced english universal dependencies: An improved representation for natural language understanding tasks. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC’16), pages 2371–2378. find our card
10. Benjamin Durme and Lenhart Schubert. 2008. Open knowledge extraction through compositional language processing. Proceedings of the 2008 Conference on Semantics in Text Processing. find
11. Angus Forbes, Kristine Lee, Gus Hahn-Powell, Marco A. Valenzuela-Escrcega, and Mihai Surdeanu. 2018. Text annotation graphs: Annotating complex natural language phenomena. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC’18), Miyazaki, Japan. European Language Resources Association (ELRA). find
12. Vered Shwartz and Chris Waterson. 2018. Olive oil is made of olives, baby oil is made for babies: Interpreting noun compounds using paraphrases in a neural model. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers), pages 218–224, New Orleans, Louisiana. Association for Computational Linguistics. find
13. Luheng He, Mike Lewis, and Luke Zettlemoyer. 2015. Question-answer driven semantic role labeling: Using natural language to annotate natural language. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 643–653, Lisbon, Portugal. Association for Computational Linguistics. find our card
14. Zeno Vendler. 1957. Verbs and times. The Philosophical Review, 66(2):143–160. find
15. Matthew Honnibal and Ines Montani. 2017. spaCy 2: Natural language understanding with Bloom embeddings, convolutional neural networks and incremental parsing. To appear. find
16. Christopher Walker, Stephanie Strassel, Julie Medero, and Kazuaki Maeda. 2006. Ace 2005 multilingual training corpus. Linguistic Data Consortium, Philadelphia, LDC2006T06. find
17. Gabriel Stanovsky, Jessica Ficler, Ido Dagan, and Yoav Goldberg. 2016. Getting more out of syntax with props. arXiv preprint arXiv:1603.01648. 53 7 ↗
18. Aaron Steven White, Drew Reisinger, Keisuke Sakaguchi, Tim Vieira, Sheng Zhang, Rachel Rudinger, Kyle Rawlins, and Benjamin Van Durme. 2016. Universal Decompositional Semantics on Universal Dependencies. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 1713–1723, Austin, Texas. Association for Computational Linguistics. find our card

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/ud-2020-tiktinsky-pybart-evidence-based-syntactic-transformations/