Elamur Research⌕
☰
home / papers / 2010 / Generating Entailment Rules from FrameNet

Generating Entailment Rules from FrameNet

Summary

Semantic inference tasks such as QA and IE need entailment rules between predicates, for example "cure X ⇒ X recover". WordNet, the usual manual source for these rules, has two gaps. It lacks many predicate relations, and it gives no argument mappings, so only synonymy and hypernymy can safely become rules. The earlier FrameNet-based resource, LexPar, took its entailment relations mostly from WordNet.

The authors present FRED, which derives rules from FrameNet alone in three stages:

  • Templates: it extracts unary templates from FrameNet's annotated sentences, parsed with Minipar.
  • Lexical relations: it infers relations between lexical units from morphological derivations (CATVAR), from "dominant" LUs that carry a frame's core meaning, and from the Inheritance, Cause and Perspective frame relations, then takes the transitive closure.
  • Rules: it builds rules by aligning templates that share a frame element.

The authors tested the rules on an IE-style entailment testbed built from ACE 2005. FRED reached the best F1, well above LexPar, and kept precision close to the no-rules baseline. It also contributed rules that WordNet lacks, which makes it a useful companion to WordNet.

Key points

  • Manual evaluation of 250 randomly sampled FRED rules found 69% of them correct. Rules mapping adjunct-like frame elements, such as Time and Location, were filtered out to improve quality.

  • On 26 ACE 2005 events, macro-averaged F1 was:

    ConfigurationF1
    FRED24.6 (precision 55.1, recall 17.6)
    No-Rules20.9
    WordNet17.8
    LexPar17.4

    FRED's F1 gain was significant under a two-sided Wilcoxon test at the 0.01 level.

  • WordNet rules raised recall but lost much precision, mainly because of rare senses when no WSD module is used. FrameNet covers mostly common senses, which keeps FRED's precision high.

  • Combining FRED with WordNet raised recall by 19% over WordNet alone, reaching 21.8. FRED adds argument mappings for non-substitutable relations, such as noun–verb pairs, and new relations like "ambush ⇒ attack".

  • Only 7.5% of incorrect extractions came from faulty FRED rules. The rule-set is publicly released by Bar-Ilan University's NLP group.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Topics: Semantic role labeling

Cited by

21 papers of our library cite it (of the 140673 whose bibliographies we read); 3 of them have a page here, 18 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{aharon2010generating,
  title = {{Generating Entailment Rules from FrameNet}},
  author = {Aharon, Roni Ben and Szpektor, Idan and Dagan, Ido},
  booktitle = {Proceedings of the ACL 2010 Conference Short Papers},
  year = {2010},
  url = {https://aclanthology.org/P10-2045/}
}

References

13 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Collin Baker, Charles Fillmore, and John Lowe. 1998. The berkeley framenet project. In Proceedings of COLING-ACL, Montreal, Canada. find our card
  2. Roy Bar-Haim, Ido Dagan, Bill Dolan, Lisa Ferro, Danilo Giampiccolo, Bernardo Magnini, and Idan Szpektor. 2006. The second pascal recognising textual entailment challenge. In Second PASCAL Challenge Workshop for Recognizing Textual Entailment. find
  3. Roni Ben Aharon. 2010. Generating entailment rules from framenet. Master’s thesis, Bar-Ilan University. find
  4. Robert Coyne and Owen Rambow. 2009. Lexpar: A freely available english paraphrase lexicon automatically extracted from framenet. In Proceedings of the Third IEEE International Conference on Semantic Computing. find
  5. Ido Dagan, Oren Glickman, and Bernardo Magnini. 2006. The pascal recognising textual entailment challenge. In Lecture Notes in Computer Science, volume 3944, pages 177–190. find
  6. Christiane Fellbaum, editor. 1998. WordNet: An Electronic Lexical Database. MIT Press, Cambridge, Massachusetts. find
  7. Danilo Giampiccolo, Bernardo Magnini, Ido Dagan, and Bill Dolan. 2007. The third pascal recognizing textual entailment challenge. In Proceedings of the ACL-PASCAL Workshop on Textual Entailment and Paraphrasing. find
  8. Nizar Habash and Bonnie Dorr. 2003. A categorial variation database for english. In Proceedings of the North American Association for Computational Linguistics (NAACL ’03), pages 96–102, Edmonton, Canada. Association for Computational Linguistics. find
  9. Dekang Lin. 1998. Dependency-based evaluation of minipar. In Proceedings of the Workshop on Evaluation of Parsing Systems at LREC. find
  10. Dan Moldovan and Adrian Novischi. 2002. Lexical chains for question answering. In Proceedings of COLING. find
  11. Idan Szpektor and Ido Dagan. 2008. Learning entailment rules for unary templates. In Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008), pages 849–856, Manchester, UK, August. find
  12. Idan Szpektor and Ido Dagan. 2009. Augmenting wordnet-based inference with argument mapping. In Proceedings of the 2009 Workshop on Applied Textual Inference, pages 27–35, Suntec, Singapore, August. find
  13. Frank Wilcoxon. 1945. Individual comparisons by ranking methods. Biometrics Bulletin, 1(6):80–83. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/aharon-2010-generating-entailment-rules-framenet/