Elamur Research⌕
☰
home / papers / 2012 / Semantic Role Labeling with the Swedish FrameNet

Semantic Role Labeling with the Swedish FrameNet

Written for the Research library (claude-opus-5-5) from the paper text, in our own words: the license "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)" does not allow redistributing the text itself. Original — aclanthology.org/L12-1241.pdf

Summary

This paper reports the first semantic role labeling (SRL) experiments built on the Swedish FrameNet (SweFN), a frame-semantic lexicon still under construction. SweFN mirrors the Berkeley FrameNet and draws its lexical units from SALDO. Its annotated example sentences provide only a small training set. The authors build a two-step pipeline on top of dependency trees from the LTH parser, which was trained on a Swedish treebank. A linear SVM first decides which tree nodes form argument spans, and an online-trained multiclass classifier then assigns frame element labels. Features are chosen by greedy forward selection. Experiments cover cross-validation over sentences and over frames, a comparison of parsers, cross-frame sharing of role labels, Brown cluster features, and a medical-domain test set. Accuracy is limited by the scarce data, but sharing labels and adding clusters both help. The argument segmenter holds up well on frames it has never seen. This makes the system a plausible aid for lexicographers who annotate new frames semi-automatically.

Key points

  • Resource: As of October 2011, SweFN had 519 frames, about 18,000 lexical units from SALDO, and 2,974 example sentences, which served as the training data.
  • Main results (5-fold CV over sentences):
    • Segmentation reached P 70.6 / R 64.8.
    • Labeling accuracy was 64.3, against a most-frequent-label baseline of 30.4.
    • Core roles scored 71.9 and non-core roles 47.2.
  • Robustness to new frames and domains:
    • On unseen frames, segmentation barely changed (P 70.6 / R 63.9), while labeling fell to 49.4.
    • Using frame-specific labels instead of shared ones lowered accuracy to 52.5.
    • Brown cluster features raised labeling accuracy from 64.3 to 65.0.
  • Parser choice: Replacing the LTH parser with MaltParser reduced results to P 64.2 / R 54.0 and labeling accuracy 61.2, showing that parse quality matters a great deal.
  • Medical domain: On 211 manually annotated sentences, the system scored P 67.3 / R 61.1 and accuracy 52.6, indicating a need for domain adaptation. Feature selection favored structural features for segmentation and lexical-plus-structural features for labeling.

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

8 papers of our library cite it (of the 140673 whose bibliographies we read).

Cite

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

@inproceedings{johansson2012semantic,
  title = {{Semantic Role Labeling with the Swedish FrameNet}},
  author = {Johansson, Richard and Heppin, Karin Friberg and Kokkinakis, Dimitrios},
  booktitle = {Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)},
  year = {2012},
  doi = {10.63317/2np8hazov8rt},
  url = {https://doi.org/10.63317/2np8hazov8rt}
}

References

20 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Lars Borin, Dana Dannélls, Markus Forsberg, Maria Toporowska Gronostaj, and Dimitrios Kokkinakis. 2010. The past meets the present in the Swedish FrameNet++. In Proceedings of EURALEX. find
  2. Lars Borin. 2010. Med Zipf mot framtiden - en integrerad lexikonresurs för svensk språkteknologi. LexicoNordica, 17. ↗
  3. Peter F. Brown, Peter V. deSouza, Robert L. Mercer, Vincent J. Della Pietra, and Jenifer C. Lai. 1992. Class-based n-gram models of natural language. Computational Linguistics, 18(4):467–479. find
  4. Ronan Collobert and Jason Weston. 2007. Fast semantic extraction using a novel neural network architecture. In Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics, pages 560–567, Prague, Czech Republic. find
  5. Dipanjan Das, Nathan Schneider, Desai Chen, and Noah A. Smith. 2010. Probabilistic frame-semantic parsing. In Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 948–956, Los Angeles, United States, June. find our card
  6. Charles Fillmore, Christopher Johnson, and Miriam Petruck. 2003. Background to FrameNet. International Journal of Lexicography, 16(3):235–250. ↗
  7. Hagen Fürstenau and Mirella Lapata. 2009. Semi-supervised semantic role labeling. In Proceedings of the 12th Conference of the European Chapter of the ACL (EACL 2009), pages 220–228, Athens, Greece, March. find
  8. Daniel Gildea and Daniel Jurafsky. 2002. Automatic labeling of semantic roles. Computational Linguistics, 28(3):245–288. ↗ our card
  9. Daniel Gildea and Martha Palmer. 2002. The necessity of syntactic parsing for predicate argument recognition. In Proceedings of the 40th Annual Conference of the Association for Computational Linguistics, pages 239–246, Philadelphia, United States. find
  10. Richard Johansson and Pierre Nugues. 2006. A FrameNet-based semantic role labeler for Swedish. In Proceedings of Coling/ACL, pages 436–443, Sydney, Australia, July 17-21. find our card
  11. Richard Johansson and Pierre Nugues. 2007. Semantic structure extraction using nonprojective dependency trees. In Proceedings of SemEval-2007, pages 227–230, Prague, Czech Republic, June 23-24. find
  12. Richard Johansson and Pierre Nugues. 2008a. Dependency-based syntactic–semantic analysis with PropBank and NomBank. In Proceedings of the CoNLL Shared Task, pages 183–187, Manchester, United Kingdom. find
  13. Richard Johansson and Pierre Nugues. 2008b. The effect of syntactic representation on semantic role labeling. In Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008), pages 393–400, Manchester, United Kingdom. find
  14. Richard Johansson. 2008. Dependency-based Semantic Analysis of Natural-language Text. Ph.D. thesis, Lund University. find
  15. Terry Koo, Xavier Carreras, and Michael Collins. 2008. Simple semi-supervised dependency parsing. In Proceedings of ACL-08: HLT, pages 595–603, Columbus, United States. find our card
  16. Lluı́s Màrquez, Xavier Carreras, Ken Litkowski, and Suzanne Stevenson. 2008. Semantic role labeling: An introduction to the special issue. Computational Linguistics, 34(2):145–159. ↗
  17. Yuichiroh Matsubayashi, Naoaki Okazaki, and Jun’ichi Tsujii. 2009. A comparative study on generalization of semantic roles in FrameNet. In Proceedings of ACL/IJCNLP, pages 19–27, Singapore. find our card
  18. Jens Nilsson, Johan Hall, and Joakim Nivre. 2005. MAMBA meets TIGER: Reconstructing a Swedish treebank from antiquity. In Proceedings of NODALIDA Special Session on Treebanks. find
  19. Joakim Nivre, Johan Hall, Jens Nilsson, Atanas Chanev, Gülşen Eryiğit, Sandra Kübler, Svetoslav Marinov, and Erwin Marsi. 2007. MaltParser: A language-independent system for data-driven dependency parsing. Natural Language Engineering, 13(2):95–135. ↗
  20. Nianwen Xue and Martha Palmer. 2004. Calibrating features for semantic role labeling. In Proceedings of the 2004 Conference on Empirical Methods in Natural Language Processing, pages 88–94, Barcelona, Spain. find

↑ back to top