SemLink 2.0: Chasing Lexical Resources

What it is about

SemLink maps between PropBank rolesets and arguments, VerbNet classes and thematic roles, FrameNet frames and frame elements, and OntoNotes sense groups, and holds a corpus of verb instances tagged in all of them. The resources keep changing (VerbNet 3.2 → 3.3, FrameNet 1.5 → 1.7, PropBank's unified frame files), so the static links rot. SemLink 2.0 regenerates the PropBank→VerbNet mappings from the links inside PropBank's frame files, tracks verbs that moved between VerbNet classes (soft subclass matching, WordNet keys), checks every VerbNet→FrameNet mapping for classes, roles and frames that still exist (2,387 valid, 160 re-annotated by hand), and marks every mapping as automatic or manual. The instance corpus nearly doubles (75k → 149k), mostly from OntoNotes 5.0 and 5,300 hand VerbNet-tagged verbs. It adds VerbNet class embeddings and, per class, the most frequent subjects and objects from a parsed Wikipedia, plus software to redo all this when a resource changes.

Ideas

• I1. A per-class argument profile from Mova's own lessons, the paper's «common arguments» done deterministically: for each VerbNet class the reader picks (vn::best), count the head lemmas and their animacy/category in agent and patient positions over the courses. Use it as a soft score in vn::best next to the selectional restrictions and the vnfreq.tsv preference (a tie-breaker in integer units), and print it in the role explanation. Measure frame bench F1 and the course test counts, as the vnfreq step was measured.
• I2. A version gate in prag/src/vnconv.rs: when prag/data/verbnet.tsv is regenerated from a new VerbNet, check that every class, frame pattern and role named elsewhere (prag/data/vnfreq.tsv, the vn-class/vn-frame atoms in prag/data/rules.tsv) still exists, and fail with the list, the way SemLink 2.0 validates its mappings. Its negative control: a rule naming a removed class must make the gate red.
• I3. Use SemLink's hand VerbNet-tagged instances (tests only, outside the build, like PropBank) to measure how often vn::best picks the gold class for a verb with several classes. That is the one step of the VerbNet reader that has no direct number today, only the downstream frame F1.

For Mova

13. Verbs and 23-complementation.md (verb classes and frames), 30-pragmatics-overview.md §5. Code: prag/src/vn.rs (vn::best), prag/src/vnconv.rs, prag/data/verbnet.tsv, prag/data/vnfreq.tsv, prag/data/rules.tsv; world frames vnfreq, world frames lexicon. Measures: frame bench F1, course tests/exams, class-choice accuracy on SemLink instances.

Topics: Semantic role labeling

Cite

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

@inproceedings{stowe2021semlink,
  title = {{SemLink 2.0: Chasing Lexical Resources}},
  author = {Stowe, Kevin and Preciado, Jenette and Conger, Kathryn and Brown, Susan Windisch and Kazeminejad, Ghazaleh and Gung, James and Palmer, Martha},
  booktitle = {Proceedings of the 14th International Conference on Computational Semantics (IWCS)},
  year = {2021},
  url = {https://aclanthology.org/2021.iwcs-1.21/}
}

References

22 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Fillmore C.J. Baker, C. F. and J.B. Lowe. 1998. The Berkeley FrameNet project. pages 86–90, Montreal, QC. COLING-ACL ’98. find our card
2. Claire Bonial, William Corvey, Martha Palmer, Volha V Petukhova, and Harry Bunt. 2011. A hierarchical unification of lirics and verbnet semantic roles. In 2011 IEEE Fifth International Conference on Semantic Computing, pages 483–489. IEEE. find
3. Claire Bonial, Kevin Stowe, and Martha Palmer. 2013. Renewing and revising SemLink. In Proceedings of the 2nd Workshop on Linked Data in Linguistics (LDL-2013): Representing and linking lexicons, terminologies and other language data, pages 9 – 17, Pisa, Italy. Association for Computational Linguistics. find
4. Susan Windisch Brown, Julia Bonn, James Gung, Annie Zaenen, James Pustejovsky, and Martha Palmer. 2019. Verbnet representations: Subevent semantics for transfer verbs. In Proceedings of the First International Workshop on Designing Meaning Representations, pages 154–163. sense tagging trenches. Essays in Lexical Semantics and Computational Lexicography - In honor of Adam Kilgarriff. find our card
5. Susan Windisch Brown, James Pustejovsky, Annie Zaenen, and Martha Palmer. 2018. Integrating generative lexicon event structures into verbnet. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC-2018). find
6. Jeffrey Pennington, Richard Socher, and Christopher Manning. 2014. GloVe: Global vectors for word representation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1532–1543, Doha, Qatar. Association for Computational Linguistics. ↗
7. Danqi Chen and Christopher Manning. 2014. A fast and accurate dependency parser using neural networks. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 740–750, Doha, Qatar. Association for Computational Linguistics. find our card
8. Matthew Peters, Mark Neumann, Mohit Iyyer, Matt Gardner, Christopher Clark, Kenton Lee, and Luke Zettlemoyer. 2018. Deep contextualized word representations. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 2227–2237, New Orleans, Louisiana. Association for Computational Linguistics. find
9. Silvana Hartmann, Éva Mújdricza-Maydt, Ilia Kuznetsov, Iryna Gurevych, and Anette Frank. 2017. Assessing SRL frameworks with automatic training data expansion. In Proceedings of the 11th Linguistic Annotation Workshop, pages 115–121, Valencia, Spain. Association for Computational Linguistics. find
10. Daisuke Kawahara, Daniel W. Peterson, and Martha Palmer. 2014. A step-wise usage-based method for inducing polysemy-aware verb classes. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1030–1040, Baltimore, Maryland. Association for Computational Linguistics. find
11. K Kipper-Schuler. 2005. VerbNet: A broad-coverage, comprehensive verb lexicon. find
12. Ilia Kuznetsov and Iryna Gurevych. 2018. Corpus-driven thematic hierarchy induction. In Proceedings of the 22nd Conference on Computational Natural Language Learning, pages 54–64, Brussels, Belgium. Association for Computational Linguistics. find
13. Ilia Kuznetsov and Iryna Gurevych. 2020. A matter of framing: The impact of linguistic formalism on probing results. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 171–182, Online. Association for Computational Linguistics. find
14. Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013. Efficient estimation of word representations in vector space. CoRR, abs/1301.3781. ↗
15. Gildea D. Palmer, M. and P. Kingsbury. 2005. The Proposition Bank: An Annotated Corpus of Semantic Roles. volume 31, pages 71–106. find our card
16. M. Palmer. 2009. Semlink: Linking PropBank, Verb-Net and FrameNet. Pisa, Italy. Proceedings of the Generative Lexicon Conference. find
17. Martha Palmer, James Gung, Claire Bonial, Jinho Choi, Orin Hargraves, Derek Palmer, and Kevin Stowe. 2017. The Pitfalls of Shortcuts: Tales from the word Daniel Peterson, Jordan Boyd-Graber, Martha Palmer, and Daisuke Kawahara. 2016. Leveraging VerbNet to build corpus-specific verb clusters. In Proceedings of the Fifth Joint Conference on Lexical and Computational Semantics, pages 102–107, Berlin, Germany. Association for Computational Linguistics. find
18. Daniel Peterson, Susan Brown, and Martha Palmer. 2020. Verb class induction with partial supervision. In Proceedings of the Thirty-fourth AAAI Conference on Artificial Intelligence, New York City, NY. find
19. Drew Reisinger, Rachel Rudinger, Francis Ferraro, Craig Harman, Kyle Rawlins, and Benjamin Van Durme. 2015. Semantic proto-roles. Transactions of the Association for Computational Linguistics, 3:475–488. find
20. Jennifer Sikos and Sebastian Padó. 2018. Using embeddings to compare FrameNet frames across languages. In Proceedings of the First Workshop on Linguistic Resources for Natural Language Processing, pages 91–101, Santa Fe, New Mexico, USA. Association for Computational Linguistics. find
21. Kevin Stowe. 2019. Syntactic and semantic improvements to computational metaphor processing. find
22. R. Weischedel, E. Hovy, M. Marcus, M. Palmer, R. Belvin, S. Pradan, L. Ramshaw, and X. Nianwen. 2011. OntoNotes: A Large Training Corpus for Enhanced Processing. Handbook of Natural Language Processing and Machine Translation: Global Automatic Language Exploitation, pages 53–63. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/stowe-2021-semlink-2-0-chasing-lexical/