From Syntax to Semantics: Introducing UMR for NLP Annotation
What it is about
A short beginner's tutorial on Uniform Meaning Representation (UMR), the cross-linguistic successor to AMR. The aim is to show how Universal Dependencies structures guide the building of minimal UMR graphs, with Portuguese examples. Participants are expected to leave knowing what UMR is, how it relates to syntax and semantic roles, how to draw a minimal graph, and how Portuguese UD treebanks can support annotation. The tutorial starts from the basics of UD (heads, dependents and relation labels) and then shows how those relations guide the choice of UMR roles and concepts. The authors come from translation studies, computational linguistics and corpus annotation, and the examples are drawn from Portuguese treebanks, not English. The paper is a teaching abstract with author notes. The text available for this card has no worked examples, no data and no measurements, so the summary above is limited to what the abstract states.
Ideas
None for Mova. The text available is the abstract and the author notes, with no worked examples or results to test. The UD-to-UMR conversion that the tutorial teaches is covered in measured form by gamba-2025-bootstrapping-umrs-universal-dependencies-scalable.
For Mova
Nothing to apply from this tutorial alone. Its subject is the UD-to-UMR step, which matters for tools/roundtrip and for the frames in prag. The measured version of that step is the reference for any work in that direction.
Cite
@inproceedings{pagano2026syntax,
title = {{From Syntax to Semantics: Introducing UMR for NLP Annotation}},
author = {Pagano, Adriana S. and Duran, Magali Sanches and Gamba, Federica},
booktitle = {Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 2},
year = {2026},
url = {https://aclanthology.org/2026.propor-2.41/}
}This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/pagano-2026-syntax-semantics-introducing-umr-nlp/