AmbiCoref: Evaluating Human and Model Sensitivity to Ambiguous Coreference

TL;DR

This paper introduces AmbiCoref, a diagnostic corpus to evaluate whether coreference resolution models are sensitive to ambiguity in pronoun references, revealing that models often ignore ambiguity unlike humans.

Contribution

The paper presents AmbiCoref, a novel dataset inspired by psycholinguistics, to test model sensitivity to ambiguity in coreference resolution tasks.

Findings

Humans are less certain of referents in ambiguous sentences.

Most models show little difference in handling ambiguous vs. unambiguous cases.

AmbiCoref enables testing of model-human sensitivity to ambiguity.

Abstract

Given a sentence "Abby told Brittney that she upset Courtney", one would struggle to understand who "she" refers to, and ask for clarification. However, if the word "upset" were replaced with "hugged", "she" unambiguously refers to Abby. We study if modern coreference resolution models are sensitive to such pronominal ambiguity. To this end, we construct AmbiCoref, a diagnostic corpus of minimal sentence pairs with ambiguous and unambiguous referents. Our examples generalize psycholinguistic studies of human perception of ambiguity around particular arrangements of verbs and their arguments. Analysis shows that (1) humans are less sure of referents in ambiguous AmbiCoref examples than unambiguous ones, and (2) most coreference models show little difference in output between ambiguous and unambiguous pairs. We release AmbiCoref as a diagnostic corpus for testing whether models treat ambiguity similarly to humans.