Squib: Expressive Power of Abstract Meaning Representations
What it is about
The squib gives AMR a model-theoretic meaning by translating it into first-order logic. Basic AMRs, with concepts as predicates, roles as binary relations and role inversion, map to conjunctions of atomic formulas, and recurrent variables are handled as reuse of the same entity. Negation is added as a polarity flag on a node. Universal quantification is not expressible in the basic language. The author shows that pulling a concept out and marking it with polarity can express a sentence with one universal subject, but only one universal per graph works this way, so a sentence with two universal noun phrases cannot be represented. He proposes an extension in which a quantification relation marks a variable as universal, and the scope is then fixed by the order of application. The conclusion is that AMR is less expressive than DRS or MRS, because conjunction is implicit and scope is not in the graph.
Ideas
- I1. Polarity as a feature: store negation on the meta token as a flag, not as a node. Check this on Bible verses with the negators not, no and neither, which differ across translations, and count how often the flag disagrees between authors.
- I2. Role inversion as a canonical form: when the same relation is read from a passive and an active clause, store one direction in the meta-world. This removes a source of attachment differences (open problem 1), and the test is whether the residual drops.
- I3. A universal flag on the meta token for all and every. Before building it, count the verses where scope matters, so that the work is spent only where it changes an output.
For Mova
The code is tools/translate/meta.rs. I1 and I2 touch open problem 1, and I1 is also the author-specific negation case of the delta. The paper's model-theoretic argument is not needed for Mova, but the counts in I3 decide whether scope is worth adding to the meta-world at all.
Cite
@article{bos2016squib,
title = {{Squib: Expressive Power of Abstract Meaning Representations}},
author = {Bos, Johan},
journal = {Computational Linguistics},
year = {2016},
doi = {10.1162/coli\_a\_00257},
url = {https://doi.org/10.1162/coli_a_00257}
}References
- Artzi, Yoav, Kenton Lee, and Luke Zettlemoyer. 2015. Broad-coverage CCG semantic parsing with AMR. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 1699–1710, Lisbon. find
- Banarescu, Laura, Claire Bonial, Shu Cai, Madalina Georgescu, Kira Griffitt, Ulf Hermjakob, Kevin Knight, Philipp Koehn, Martha Palmer, and Nathan Schneider. 2015. Abstract Meaning Representation (AMR) 1.2.2 Specification. github.com/ amrisi/amr-guidelines/blob/master/ amr.md. find
- Bender, Emily M., Dan Flickinger, Stephan Oepen, Woodley Packard, and Ann Copestake. 2015. Layers of interpretation: On grammar and compositionality. In Proceedings of the 11th International Conference on Computational Semantics, pages 239–249, London. find
- Blackburn, Patrick and Johan Bos. 2005. Representation and Inference for Natural Language. A First Course in Computational Semantics. CSLI, Stanford, CA. find
- Bos, Johan. 2010. Economical discourse representation theory. In CNL 2009 Workshop, pages 121–134, Marettimo Island. find
- Cai, Shu and Kevin Knight. 2013. Smatch: An evaluation metric for semantic feature structures. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 748–752, Sofia. find
- Chen, Wei-Te. 2015. Learning to map dependency parses to abstract meaning representations. In Proceedings of the ACL-IJCNLP 2015 Student Research Workshop, pages 41–46, Beijing. find
- Copestake, Ann, Dan Flickinger, Ivan Sag, and Carl Pollard. 2005. Minimal recursion semantics: An introduction. Journal of Research on Language and Computation, 3(2–3):281–332. find
- Kamp, Hans and Uwe Reyle. 1993. From Discourse to Logic; An Introduction to Modeltheoretic Semantics of Natural Language, Formal Logic and DRT. Kluwer, Dordrecht. find
- Langkilde, Irene and Kevin Knight. 1998. Generation that exploits corpus-based statistical knowledge. In Proceedings of the 36th Annual Meeting of the ACL and 17th International Conference on Computational Linguistics, pages 704–710, Montreal. find
- Mortimer, Michael. 1975. On languages with two variables. Mathematical Logic Quarterly, 21(1):135–140. find
- Peng, Xiaochang, Linfeng Song, and Daniel Gildea. 2015. A synchronous hyperedge replacement grammar based approach for AMR parsing. In Proceedings of the Nineteenth Conference on Computational Natural Language Learning, pages 32–41, Beijing. find
- Potts, Christopher. 2005. The Logic of Conventional Implicatures. Oxford Studies in Theoretical Linguistics. Oxford University Press, Oxford. find
- Pust, Michael, Ulf Hermjakob, Kevin Knight, Daniel Marcu, and Jonathan May. 2015. Parsing English into abstract meaning representation using syntax-based machine translation. In Proceedings of the EMNLP 2015, pages 1143–1154, Lisbon. find
- Sawai, Yuichiro, Hiroyuki Shindo, and Yuji Matsumoto. 2015. Semantic structure analysis of noun phrases using abstract meaning representation. In Proceedings of the 53rd Annual Meeting of the ACL (Volume 2: Short Papers), pages 851–856, Beijing. find
- Van der Sandt, Rob A. 1992. Presupposition projection as anaphora resolution. Journal of Semantics, 9:333–377. find
- Wang, Chuan, Nianwen Xue, and Sameer Pradhan. 2015. Boosting transition-based AMR parsing with refined actions and auxiliary analyzers. In Proceedings of the 53rd Annual Meeting of the ACL (Volume 2: Short Papers), pages 857–862, Beijing. find
- Werling, Keenon, Gabor Angeli, and Christopher D. Manning. 2015. Robust subgraph generation improves abstract meaning representation parsing. In Proceedings of the 53rd Annual Meeting of the ACL (Volume 1: Long Papers), pages 982–991, Beijing. find
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/bos-2016-squib-expressive-power-abstract-meaning/