Mova0.5 pre-release⌕
☰
home / docs / papers / Resolving Pronoun References

Resolving Pronoun References

Summary

The paper presents what became known as the Hobbs algorithm, a purely syntactic procedure for finding the antecedent of a pronoun. It walks the parse tree of the current sentence upward from the pronoun and searches candidate noun phrases breadth-first, left to right, while respecting constraints that prevent a pronoun from taking a co-argument as its antecedent; if nothing suitable is found it continues into previous sentences, searching their trees in the same order. Candidates must agree in gender and number. Hobbs evaluated the procedure by hand on several texts and found it surprisingly accurate given that it uses no semantics, and he contrasted it with a more knowledge-based semantic approach. The "naive" algorithm became a long-standing baseline for anaphora resolution.

How Mova uses it

Effectiveness in Mova

Measured on GUM dev with gold trees (the development notes of coref), Hobbs order versus pure recency: pronoun precision 64.2% vs 55.4%; CoNLL F1 73.43 / 63.34 vs 72.23 / 61.33 (with / without singletons).


👨‍🔬💥

Topics: Coreference resolution

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.1016/0024-3841(78)90006-2

@inproceedings{hobbs1978resolving,
  title = {{Resolving Pronoun References}},
  author = {Hobbs, Jerry R.},
  booktitle = {Lingua 44(4), 311-338},
  year = {1978},
  doi = {10.1016/0024-3841(78)90006-2},
  url = {https://doi.org/10.1016/0024-3841(78)90006-2}
}

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/hobbs-1978-resolving-pronoun-references/