DisSim: A Discourse-Aware Syntactic Text Simplification Framework for English and German

What it is about

DisSim is a system demo (4 pages) that splits a complex sentence into minimal propositions and arranges them in a hierarchy of core facts and context sentences, labelled with rhetorical relations (based on Rhetorical Structure Theory). The English version uses 35 hand-written transformation rules on a phrasal parse; the German version uses 29 rules on spaCy dependency parses. The output is a discourse tree in JSON, with a Java API, a REST service and a browser interface. As an extrinsic test, placing DisSim before Open IE systems raised precision by up to 346% and recall by up to 52%. On the RST Discourse Treebank, the hierarchy (nucleus/satellite) was assigned with almost 90% precision and the relations with about 70% on average; the German evaluation is said to be in progress.

Ideas

• I1. en/src/parse.rs output and prag/src/frames.rs: a rule pass over UD trees that marks advcl, ccomp, acl and conj clauses as satellites or cores, using the dependency labels rather than phrase labels. Measure the share of sentences split and the share of core propositions that keep the subject and verb.
• I2. A cue-word table for relations (however, because, then, although) as a sorted map in prag, keyed by enum values; check it against a hand-labelled set of 100 sentences, held out until the release test.
• I3. The Open IE gain is a clean deterministic measure: count the propositions world extracts from a text before and after splitting, and report precision on a small gold set. Run a negative control with the sentence order shuffled; the gain must vanish.

For Mova

The English code is en/src/parse.rs (UD trees) and prag (frames, relations). The chapters are 23. Complementation, 24-coordination-subordination.md and 36-discourse-relations.md. The German rule set is not for Mova now. The paper gives a design, not a corpus, so the measure has to be built: hand-labelled English sentences with cores, satellites and cue words.

Topics: Discourse analysis

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/w19-8662

@inproceedings{niklaus2019dissim,
  title = {{DisSim: A Discourse-Aware Syntactic Text Simplification Framework for English and German}},
  author = {Niklaus, Christina and Cetto, Matthias and Freitas, André and Handschuh, Siegfried},
  booktitle = {Proceedings of the 12th International Conference on Natural Language Generation},
  year = {2019},
  doi = {10.18653/v1/w19-8662},
  url = {https://doi.org/10.18653/v1/w19-8662}
}

References

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

1. Michele Banko, Michael J. Cafarella, Stephen Soderland, Matt Broadhead, and Oren Etzioni. 2007. Open information extraction from the web. In Proceedings of the 20th International Joint Conference on Artifical Intelligence, pages 2670–2676, San Francisco, CA, USA. Morgan Kaufmann Publishers Inc. find
2. Hannah Bast and Elmar Haussmann. 2013. Open information extraction via contextual sentence decomposition. In 2013 IEEE Seventh International Conference on Semantic Computing, pages 154–159. IEEE. find
3. Lynn Carlson, Mary Ellen Okurowski, and Daniel Marcu. 2002. RST discourse treebank. Linguistic Data Consortium, University of Pennsylvania. find
4. William C Mann and Sandra A Thompson. 1988. Rhetorical structure theory: Toward a functional theory of text organization. Text-Interdisciplinary Journal for the Study of Discourse, 8(3):243–281. find our card
5. Shashi Narayan, Claire Gardent, Shay B. Cohen, and Anastasia Shimorina. 2017. Split and rephrase. In Proceedings of the 2017 Conference on EMNLP, pages 606–616. ACL. find
6. Christina Niklaus, Matthias Cetto, André Freitas, and Siegfried Handschuh. 2018. A survey on open information extraction. In Proceedings of the 27th International Conference on Computational Linguistics, pages 3866–3878, Santa Fe, New Mexico, USA. Association for Computational Linguistics. find
7. Christina Niklaus, Matthias Cetto, André Freitas, and Siegfried Handschuh. 2019. Transforming complex sentences into a semantic hierarchy. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3415–3427, Florence, Italy. Association for Computational Linguistics. find
8. Advaith Siddharthan. 2014. A survey of research on text simplification. ITL-International Journal of Applied Linguistics, 165(2):259–298. find
9. Richard Socher, John Bauer, Christopher D. Manning, and Andrew Y. Ng. 2013. Parsing With Compositional Vector Grammars. In ACL. find
10. Gabriel Stanovsky, Julian Michael, Luke Zettlemoyer, and Ido Dagan. 2018. Supervised open information extraction. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), volume 1, pages 885–895. find
11. Maite Taboada and Debopam Das. 2013. Annotation upon annotation: Adding signalling information to a corpus of discourse relations. D&D, 4(2):249–281. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/niklaus-2019-dissim-discourse-aware-syntactic-text/