Semi-automatic Annotation of Event Structure, Argument Structure, and Opposition Structure to WordNet by Using Event Structure Frame

What it is about

The paper links an event structure frame (ESF) to each WordNet synset of English verbs. An ESF has three parts: a pre-state, a process and a post-state, with semantic roles in each. The frame is assigned by GESL, an earlier rule-based tagger, which first classifies the verb by lexical aspect, then by semantic class, then by event type. GESL is run on the example sentence of each synset, and the result is corrected by hand, because many examples are incomplete and some synsets have none. For the test, 91 motion verbs with 952 synsets were annotated, using WordNet 2.1 through NLTK. About 30.6% of the synsets (291 of 952) fall in motion or change-of-location classes. The verb run alone has 41 synsets with 12 different frame types, so the frame has to be attached to the sense and not to the lemma. The paper proposes 44 verb classes, up from 23 in the earlier tagger.

Ideas

• I1. world/src/events.rs: write pre-state, process and post-state frames for the change-of-location verbs (arrive, leave, pass, transfer) in the lessons, and check whether the post-state inference answers the story questions. Measure the share of correct entailed-state answers on the lesson stories.
• I2. Key the frames by sense and not by lemma: world/src/dictionary.rs should carry the frame id per sense, as the run example shows. Compare lemma-level and sense-level frame accuracy on the same sentences.
• I3. Use the example sentences of the lesson exercises as the first pass, then correct by hand and keep the corrections as a test set. Report how many corrections each frame class needed; a class with many corrections is a rule to revisit.

For Mova

The chapter is 13. Verbs, with 28. Voice and the passive for the argument side. The code is world/src/events.rs and world/src/story.rs, and the frame reader in prag/src/frames.rs. The measures: post-state answer accuracy on the lessons, sense-level frame accuracy against lemma-level, and the correction count per frame class. The frame inventory is the idea to borrow; GESL's neural-free rules show that such frames can be applied deterministically.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/2019.gwc-1.35

@inproceedings{im2019semi,
  title = {{Semi-automatic Annotation of Event Structure, Argument Structure, and Opposition Structure to WordNet by Using Event Structure Frame}},
  author = {Im, Seohyun},
  booktitle = {Proceedings of the 10th Global Wordnet Conference},
  year = {2019},
  doi = {10.18653/v1/2019.gwc-1.35},
  url = {https://doi.org/10.18653/v1/2019.gwc-1.35}
}

References

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

1. Fillmore, Charles J. 1976. Frame Semantics and the Nature of Language. Annals of the New York Academy of Sciences: Conference on the Origin and Development of Language and Speech 280: 20-32. find
2. Im, Seohyun. 2013. The Generator of the Event Structure Lexicon (GESL): Automatic Annotation of Event Structure for Textual Inference Tasks. PhD Dissertation, Brandeis University, MA, USA. find
3. Im, Seohyun and James Pustejovsky. 2010. Annotating Lexically Entailed Subevents for Textual Inference Tasks. In the Proceedings of FLAIRS 23, Daytona Beach, Florida, USA, 2010. find
4. Im, Seohyun and James Pustejovsky. 2009. Annotating Event Implicatures for Textual Inference Tasks. In the Proceedings of the 5 th International Conference on Generative Approaches to the Lexicon, Pisa, Italy, 2009. find
5. Kipper, Karin Schuler. 2005. VerbNet: A Broad-coverage, Comprehensive Verb Lexicon. PhD Dissertation. University of Pennsilvenia. find
6. Miller, George A. 1995. Wordnet: A Lexical Database for English. Communications of the ACM 38, no. 11. find our card
7. Osswald, Rainer and Jr. Robert D. Van Valin. 2012. FrameNet, Fame Structure, and the Syntax-Semantics Interface (draft). find
8. Pustejovsky, James. 2000. Events and the Semantics of Opposition, Pustejovsky and Tenny (eds.) Events as Grammatical Objects. CSLI Publications. Pustejovsky. find
9. James. 1995. The Generative Lexicon . The MIT Press. Zaenen, Annie, Cleo Condoravi, and Danny. find
10. Bobrow. 2008. The Encoding of Lexical Implications in VerbNet. Proceedings of LREC 2008. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/im-2019-semi-automatic-annotation-event-structure/