Evaluating Large-Scale Construction Grammars on the Tasks of Semantic Frame Extraction and Semantic Role Labeling

Thomas Moerman; Paul Van Eecke; Katrien Beuls

doi:10.24338/cons-651

Evaluating Large-Scale Construction Grammars on the Tasks of Semantic Frame Extraction and Semantic Role Labeling

Constructions, vol. 16, 2024

Construction Grammar

Evaluates construction grammars on frame extraction and role labeling, comparing three parsing heuristics.

Authors

Thomas Moerman

Paul Van Eecke

Katrien Beuls

Published

January 1, 2024

An evaluation of computational construction grammars in Fluid Construction Grammar on semantic frame extraction and semantic role labeling, using OntoNotes and PropBank. It compares three heuristics for resolving competition between constructions and finds that a network-based heuristic performs best, while a purely frequency-based one is harmful.

DOI

Research theme: Construction grammar

Citation

BibTeX citation:

@article{moerman2024,
  author = {Moerman, Thomas and Van Eecke, Paul and Beuls, Katrien},
  title = {Evaluating {Large-Scale} {Construction} {Grammars} on the
    {Tasks} of {Semantic} {Frame} {Extraction} and {Semantic} {Role}
    {Labeling}},
  journal = {Constructions},
  date = {2024-01-01},
  doi = {10.24338/cons-651},
  langid = {en}
}

For attribution, please cite this work as:

Moerman, Thomas, Paul Van Eecke, and Katrien Beuls. 2024. “Evaluating Large-Scale Construction Grammars on the Tasks of Semantic Frame Extraction and Semantic Role Labeling.” Constructions, accepted, January 1. https://doi.org/10.24338/cons-651.