Elamur Research
☰
home / papers / 2013 / Universal Conceptual Cognitive Annotation (UCCA)

Universal Conceptual Cognitive Annotation (UCCA)

What it is about

UCCA is a semantic annotation scheme that tries to abstract away from syntactic form. Each sentence is a directed acyclic graph whose leaves are words, and whose units mark scenes, argument structure of verbs, nouns and adjectives, and the links between them. The argument is that syntactic schemes separate sentences that mean the same thing, such as taking a shower and showering, while treating some sentences that differ in meaning alike, such as taking a shower and taking someone's book, which the Penn Treebank annotates identically. The foundational layer covers every word, so each word belongs to at least one unit, and further layers can be added on top. Results: on 30 PropBank sentences, the converted UCCA structure agrees with PropBank with an unlabelled F-score of 89.4%. The authors also built a UCCA-annotated corpus, report that annotators without a linguistics background learn the scheme quickly, and name parsing and English–French translation as future work.

Ideas

For Mova

The code is tools/translate/meta.rs, with the layering in tools/roundtrip. The base-layer idea bears on open problem 5 (number and person from UD features), and I3 is a direct test of whether the meta-world is insensitive to paraphrase. The annotation procedure itself is not for Mova.

Library summary

The Research library's own summary of the paper, in our words — not the paper's text.

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/P13-1023.pdf

Summary

Syntactic annotation schemes such as the Penn Treebank or CoNLL-style dependencies mainly reflect the form of an utterance rather than its content. As a result, semantically similar sentences may receive different annotations, while sentences that differ in meaning may receive identical ones. For example, "took a shower" and "took my book" are annotated identically. The authors propose UCCA, a multi-layered semantic annotation scheme that abstracts away from specific syntactic constructions. It builds on Basic Linguistic Theory and cognitive linguistics. The scheme represents the structure of a text as a directed acyclic graph (DAG) and places categories on its edges. The paper describes the foundational layer in detail: it covers the argument structure of verbal, nominal and adjectival predicates, as well as the relations between "scenes". The authors compiled a corpus annotated with UCCA. They showed that people without linguistic training learn the scheme quickly and annotate as well as experts. The scheme is also largely insensitive to paraphrasing and cross-linguistic differences. This matters because it lowers the cost of building semantic resources for machine translation, QA and other applications.

Key points

  • The foundational layer has 12 categories. Scene elements: Process, State, Participant, Adverbial. Elements of non-scene units: Center, Elaborator, Connector, Relator. Inter-scene relations: Parallel Scene, Linker, Ground. Function stands apart.
  • A unit may have several Centers, so the scheme does not depend on the single-head assumption on which syntactic schemes rest.
  • The authors annotated the first 30 sentences of section 02 of the PropBank corpus in UCCA and converted the result into PropBank-style annotation. The unlabeled F-score was 89.4%, i.e., PropBank annotations can largely be derived from UCCA.
  • Unlike PropBank/NomBank, UCCA annotates copular clauses, multi-word predicates ("take a shower") and inter-scene relations. Unlike FrameNet, it provides full coverage of naturally occurring text.
  • The authors plan a parser for UCCA, an application to English–French machine translation, and porting the scheme to dynamic domains, including social media.

Topics: Machine translation

Cite

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

@inproceedings{abend2013universal,
  title = {{Universal Conceptual Cognitive Annotation (UCCA)}},
  author = {Abend, Omri and Rappoport, Ari},
  booktitle = {Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2013},
  url = {https://aclanthology.org/P13-1023/}
}

References

38 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Omri Abend and Ari Rappoport. 2013. UCCA: A semantics-based grammatical annotation scheme. In IWCS ’13, pages 1–12. find
  2. Ron Artstein and Massimo Poesio. 2008. Inter-coder agreement for computational linguistics. Computational Linguistics, 34(4):555–596. find
  3. Collin F. Baker, Charles J. Fillmore, and John B. Lowe. 1998. The Berkeley Framenet project. In ACL-COLING ’98, pages 86–90. find our card
  4. Laura Banarescu, Claire Bonial, Shu Cai, Madalina Georgescu, Kira Griffitt, Ulf Hermjakob, Kevin Knight, Philipp Koehn, Martha Palmer, and Nathan Schneider. 2012. Abstract meaning representation (AMR) 1.0 specification. http://www.isi.edu/natural-language/people/amr-guidelines-10-31-12.pdf. ↗
  5. Valerio Basile, Johan Bos, Kilian Evang, and Noortje Venhuizen. 2012. Developing a large semantically annotated corpus. In LREC ’12, pages 3196–3200. find
  6. Alena Böhmová, Jan Hajič, Eva Hajičová, and Barbora Hladká. 2003. The Prague Dependency Treebank. Treebanks, pages 103–127. find
  7. Alistair Butler and Kei Yoshimoto. 2012. Banking meaning representations from treebanks. Linguistic Issues in Language Technology, 7(1). find
  8. Alexander Clark and Shalom Lappin. 2010. Linguistic Nativism and the Poverty of the Stimulus. Wiley-Blackwell. find
  9. Robert M. W. Dixon. 2005. A Semantic Approach To English Grammar. Oxford University Press. find
  10. Robert M. W. Dixon. 2010a. Basic Linguistic Theory: Methodology, volume 1. Oxford University Press. find
  11. Robert M. W. Dixon. 2010b. Basic Linguistic Theory: Grammatical Topics, volume 2. Oxford University Press. find
  12. Robert M. W. Dixon. 2012. Basic Linguistic Theory: Further Grammatical Topics, volume 3. Oxford University Press. find
  13. Bill Dolan, Chris Quirk, and Chris Brockett. 2004. Unsupervised construction of large paraphrase corpora: Exploiting massively parallel news sources. In COLING ’04, pages 350–356. find our card
  14. Bonnie Dorr, Rebecca Passonneau, David Farwell, Rebecca Green, Nizar Habash, Stephen Helmreich, Edward Hovy, Lori Levin, Keith Miller, Teruko Mitamura, Owen Rambow, and Advaith Siddharthan. 2010. Interlingual annotation of parallel text corpora: A new framework for annotation and evaluation. Natural Language Engineering, 16(3):197– 243. find
  15. Angelina Ivanova, Stephan Oepen, Lilja Øvrelid, and Dan Flickinger. 2012. Who did what to whom?: A contrastive study of syntacto-semantic dependencies. In LAW ’12, pages 2–11. find
  16. Bevan Jones, Jacob Andreas, Daniel Bauer, Karl Moritz Hermann, and Kevin Knight. 2012. Semantics-based machine translation with hyperedge replacement grammars. In COLING ’12, pages 1359–1376. find our card
  17. Aravind K. Joshi and Yves Schabes. 1997. Tree-adjoining grammars. Handbook Of Formal Languages, 3:69–123. find
  18. Ronald M. Kaplan and Joan Bresnan. 1981. Lexical-Functional Grammar: A Formal System For Grammatical Representation. Massachusetts Institute Of Technology, Center For Cognitive Science. find
  19. Kevin Knight and Daniel Marcu. 2000. Statistics-based summarization – step one: Sentence compression. In AAAI ’00, pages 703–710. find
  20. Tom Kwiatkowski, Luke Zettlemoyer, Sharon Goldwater, and Mark Steedman. 2010. Inducing probabilistic CCG grammars from logical form with higher-order unification. In EMNLP ’10, pages 1223–1233. find
  21. R.W. Langacker. 2008. Cognitive Grammar: A Basic Introduction. Oxford University Press, USA. find
  22. Mitchell P. Marcus, Mary Ann Marcinkiewicz, and Beatrice Santorini. 1993. Building a large annotated corpus of English: The Penn Treebank. Computational Linguistics, 19(2):313–330. find our card
  23. Adam Meyers, Ruth Reeves, Catherine Macleod, Rachel Szekely, Veronika Zielinska, Brian Young, and Ralph Grishman. 2004. Annotating noun argument structure for Nombank. In LREC ’04, pages 803–806. find
  24. Jason Naradowsky, Sebastian Riedel, and David Smith. 2012. Improving NLP through marginalization of hidden syntactic structure. In EMNLP ’12, pages 810–820. find
  25. Stephan Oepen, Dan Flickinger, Kristina Toutanova, and Christopher D Manning. 2004. Lingo redwoods. Research on Language and Computation, 2(4):575–596. find
  26. Martha Palmer, Daniel Gildea, and Paul Kingsbury. 2005. The proposition bank: An annotated corpus of semantic roles. Computational Linguistics, 31(1):145–159. find our card
  27. Carl Pollard and Ivan A. Sag. 1994. Head-driven Phrase Structure Grammar. University Of Chicago Press. find
  28. Ivan A Sag. 2010. Sign-based construction grammar: An informal synopsis. Sign-based Construction Grammar. CSLI Publications, Stanford, pages 39–170. find
  29. Federico Sangati and Chiara Mazza. 2009. An English dependency treebank à la Tesnière. In TLT ’09, pages 173–184. find
  30. Roy Schwartz, Omri Abend, Roi Reichart, and Ari Rappoport. 2011. Neutralizing linguistically problematic annotations in unsupervised dependency parsing evaluation. In ACL-HLT ’11, pages 663– 672. find
  31. Advaith Siddharthan. 2006. Syntactic simplification and text cohesion. Research on Language & Computation, 4(1):77–109. find
  32. Mark Steedman. 2001. The Syntactic Process. MIT Press. find
  33. Mihai Surdeanu, Richard Johansson, Adam Meyers, Lluı́s Màrquez, and Joakim Nivre. 2008. The CoNLL-2008 shared task on joint parsing of syntactic and semantic dependencies. In CoNLL ’08, pages 159–177. find our card
  34. Jun Suzuki, Hideki Isozaki, Xavier Carreras, and Michael Collins. 2009. An empirical study of semi-supervised structured conditional models for dependency parsing. In EMNLP ’09, pages 551–560. find
  35. Hiroshi Uchida and Meiying Zhu. 2001. The universal networking language beyond machine translation. In International Symposium on Language in Cyberspace, pages 26–27. find
  36. David Vadas and James R Curran. 2011. Parsing noun phrases in the Penn Treebank. Computational Linguistics, 37(4):753–809. find our card
  37. Robert D. Van Valin Jr. 2005. Exploring The Syntax-semantics Interface. Cambridge University Press. find
  38. Mengqiu Wang, Noah A. Smith, and Teruko Mitamura. 2007. What is the Jeopardy model? A quasi-synchronous grammar for QA. In EMNLP-CoNLL ’07, pages 22–32. find