Talk: Multilingual Mention Detection for Coreference Resolution (O. Uryupina, 21/02/2014)

Speaker: Olga Uryupina
University of Trento
Data: February 21th 2014, Friday
Time: 12:00
Room: 3.2 computer Science Faculty (UPV/EHU)
Title: Multilingual Mention Detection for Coreference Resolution

Abstract:

Accurate mention detection is a vital prerequisite for a variety of
Natural Language Processing tasks, in particular, for Relation
Extraction and Coreference Resolution. If a toolkit cannot extract
mentions reliably, it will obviously be unable to assign them to
relations or entities.

We propose a novel algorithm for multilingual mention detection: we
extract mentions from parse trees via kernel-based SVM learning. Our
approach allows for straightforward mention detection for any language
where (not necessary perfect) parsing resources are available, without
any complex language-specific rule engineering.  We also investigate
possibilities for incorporating automatically acquired mentions into
an end-to-end coreference resolution system. We evaluate our approach
on the Arabic and Chinese portions of the CoNLL-2012 dataset, showing
a significant improvement over the system with the baseline mention
detection.

(Joint work with Alessandro Moschitti)