Elamur Research⌕
☰
home / papers / 2015 / FrameNet+: Fast Paraphrastic Tripling of FrameNet

FrameNet+: Fast Paraphrastic Tripling of FrameNet

Summary

FrameNet is a manually built lexical-semantic resource of nearly 1,000 frames. Its practical value is limited because many words in real text are not listed as lexical units (LUs) for any frame. In the authors' New York Times sample, fewer than half of the targets could be matched to a correct frame. To address this, the authors expand FrameNet in two steps. First, they automatically rewrite FrameNet's frame-annotated fulltext sentences using single-word paraphrase rules from PPDB-XL, ranked with a 5-gram Kneser-Ney language model. Second, Amazon Mechanical Turk workers rate each rewrite so that substitutions that change the evoked frame are removed. The result, released as FrameNet+, adds about 22K human-vetted frame/LU pairs, roughly tripling the resource. An expert-annotated evaluation on NYT words shows a 40% improvement in lexical coverage. Unlike earlier expansion work tied to specific tools, this is a reusable public resource. It can also supply extra training data for frame-semantic parsers.

Key points

  • Paraphrase generation: PPDB-XL contributes over 370K lexical paraphrases. Applying them yielded 188,061 paraphrased sentences across 686 frames.
  • Crowdsourced vetting: Each paraphrase was rated by three Turkers on a 5-point scale, with WordNet-based control items. Workers scored 76% on the controls, and inter-annotator correlation was ρ = 0.49. About 44% of paraphrases averaged a rating of 3 or higher, giving 85K approved sentences and 22K new LU/frame mappings.
  • Coverage evaluation: The test set was 300 NYT word types (100 nouns, 100 verbs, 100 adjectives), labeled by two experts with κ = 0.56. At threshold t = 3, FN+ covered 60% of targets, 40% more than the original FN. Using all unvetted paraphrases (t = 0) reached 71% coverage.
  • Lexicon density: FN+ averages more than 40 LUs per frame, compared with roughly 10 in the original FrameNet. Including unvetted paraphrases raises this to nearly 70.
  • Release contents: About 31K LUs in total, most of them nouns (12,786), verbs (10,862) and adjectives (6,195). The release also includes the 85K approved paraphrased fulltext sentences, far more than the original 4K.

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 · Machine translation

Topics: Semantic role labeling

Cited by

30 papers of our library cite it (of the 140673 whose bibliographies we read); 4 of them have a page here, 26 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.3115/v1/p15-2067

@inproceedings{pavlick2015framenet,
  title = {{FrameNet+: Fast Paraphrastic Tripling of FrameNet}},
  author = {Pavlick, Ellie and Wolfe, Travis and Rastogi, Pushpendre and Callison-Burch, Chris and Dredze, Mark and Durme, Benjamin Van},
  booktitle = {Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 2: Short Papers)},
  year = {2015},
  doi = {10.3115/v1/p15-2067},
  url = {https://doi.org/10.3115/v1/p15-2067}
}

References

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

  1. Roni Ben Aharon, Idan Szpektor, and Ido Dagan. 2010. Generating entailment rules from framenet. In ACL, pages 241–246. find our card
  2. Collin F Baker, Charles J Fillmore, and John B Lowe. 1998. The Berkeley FrameNet Project. In COLING. Aljoscha Burchardt and Anette Frank. Approaching textual entailment with LFG and FrameNet frames. In Proceedings of the Second PASCAL RTE Challenge Workshop. Citeseer. find our card
  3. Olivia Buzek, Philip Resnik, and Benjamin B Bederson. 2010. Error driven paraphrase annotation using mechanical turk. In Proceedings of the NAACL HLT 2010 Workshop on Creating Speech and Language Data with Amazon’s Mechanical Turk, pages 217– 221. Association for Computational Linguistics. find
  4. Chris Callison-Burch. 2008. Syntactic constraints on paraphrases extracted from parallel corpora. In EMNLP, pages 196–205. Association for Computational Linguistics. find our card
  5. Janara Christensen, Stephen Soderland, Oren Etzioni, et al. 2010. Semantic role labeling for open information extraction. In Proceedings of the NAACL HLT 2010 First International Workshop on Formalisms and Methodology for Learning by Reading, pages 52–60. Association for Computational Linguistics. find
  6. Dipanjan Das and Noah A Smith. 2011. Semi-supervised frame-semantic parsing for unknown predicates. In ACL, pages 1435–1444. find our card
  7. Dipanjan Das and Noah A Smith. 2012. Graph-based lexicon expansion with sparsity-inducing penalties. In NAACL. ↗
  8. Michael Ellsworth and Adam Janin. 2007. Mutaphrase: Paraphrasing with framenet. In Proceedings of the ACL-PASCAL Workshop on Textual Entailment and Paraphrasing, pages 143–150. find our card
  9. Oscar Ferrández, Michael Ellsworth, Rafael Munoz, and Collin F Baker. 2010. Aligning FrameNet and WordNet based on semantic neighborhoods. In LREC. find our card
  10. Charles J Fillmore and Collin F Baker. 2001. Frame semantics for text understanding. In Proceedings of WordNet and Other Lexical Resources Workshop, NAACL. Association for Computational Linguistics. find
  11. Charles Fillmore. 1982. Frame semantics. Linguistics in the morning calm. find
  12. Marco Fossati, Claudio Giuliano, and Sara Tonelli. 2013. Outsourcing FrameNet to the crowd. In ACL, pages 742–747. find our card
  13. Juri Ganitkevitch, Benjamin Van Durme, and Chris Callison-Burch. 2013. Ppdb: The paraphrase database. In HLT-NAACL. find our card
  14. Daniel Gildea and Daniel Jurafsky. 2002. Automatic labeling of semantic roles. Computational linguistics, 28(3):245–288. ↗ our card
  15. Michelle L Gregory, Liam McGrath, Eric Belanga Bell, 2006. find
  16. Kelly O’Hara, and Kelly Domico. 2011. Domain independent knowledge base population from structured and unstructured data sources. In FLAIRS Conference. find
  17. Kenneth Heafield, Ivan Pouzyrevsky, Jonathan H Clark, and Philipp Koehn. 2013. Scalable modified kneser-ney language model estimation. In ACL. find our card
  18. Karl Moritz Hermann, Dipanjan Das, Jason Weston, and Kuzman Ganchev. 2014. Semantic frame identification with distributed word representations. In ACL. find
  19. Richard Johansson and Pierre Nugues. 2007. Using WordNet to extend FrameNet coverage. In Building Frame Semantics Resources for Scandinavian and Baltic Languages, pages 27–30. Department of Computer Science, Lund University. find
  20. Behrang Mohit and Srini Narayanan. 2003. Semantic extraction with wide-coverage lexical resources. In NAACL-HLT, pages 64–66. find
  21. Alexis Palmer and Caroline Sporleder. 2010. Evaluating FrameNet-style semantic parsing: The role of coverage gaps in framenet. In COLING. Association for Computational Linguistics. find our card
  22. Marco Pennacchiotti, Diego De Cao, Roberto Basili, Danilo Croce, and Michael Roth. 2008. Automatic induction of FrameNet lexical units. In EMNLP. find our card
  23. Pushpendre Rastogi and Benjamin Van Durme. 2014. Augmenting FrameNet via PPDB. In Proceedings of the 2nd Workshop on Events: Definition, Detection, Coreference, and Representation. Association of Computational Linguistics. find our card
  24. Dan Shen and Mirella Lapata. 2007. Using semantic roles to improve question answering. In EMNLP-CoNLL. find
  25. Lei Shi and Rada Mihalcea. 2005. Putting pieces together: Combining FrameNet, VerbNet and Word-Net for robust semantic parsing. In Computational linguistics and intelligent text processing. Springer. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/pavlick-2015-framenet-fast-paraphrastic-tripling-framenet/