Automatic Enrichment of Abstract Meaning Representations

Yuxin Ji, Gregor Williamson, Jinho D. Choi


Abstract Meaning Representation (AMR) is a semantic graph framework which inadequately represent a number of important semantic features including number, (in)definiteness, quantifiers, and intensional contexts. Several proposals have been made to improve the representational adequacy of AMR by enriching its graph structure. However, these modifications are rarely added to existing AMR corpora due to the labor costs associated with manual annotation. In this paper, we develop an automated annotation tool which algorithmically enriches AMR graphs to better represent number, (in)definite articles, quantificational determiners, and intensional arguments. We compare our automatically produced annotations to gold-standard manual annotations and show that our automatic annotator achieves impressive results. All code for this paper, including our automatic annotation tool, is publicly available at

Venue / Year

Proceedings of the LREC Workshop on Linguistic Annotation / 2022


Anthology | Paper | Presentation | BibTeX | GitHub