Improving pairwise coreference models through feature space hierarchy learning - Archive ouverte HAL Access content directly
Conference Papers Year : 2013

Improving pairwise coreference models through feature space hierarchy learning

(1) , (2)
1
2

Abstract

This paper proposes a new method for significantly improving the performance of pairwise coreference models. Given a set of indicators, our method learns how to best separate types of mention pairs into equivalence classes for which we construct distinct classification models. In effect, our approach finds an optimal feature space (derived from a base feature set and indicator set) for discriminating coreferential mention pairs. Although our approach explores a very large space of possible feature spaces, it remains tractable by exploiting the structure of the hierarchies built from the indicators. Our experiments on the CoNLL-2012 Shared Task English datasets (gold mentions) indicate that our method is robust relative to different clustering strategies and evaluation metrics, showing large and consistent improvements over a single pairwise model using the same base features. Our best system obtains a competitive 67:2 of average F1 over MUC, B3 , and CEAF which, despite its simplicity, places it above the mean score of other systems on these datasets.
Fichier principal
Vignette du fichier
main.pdf (166.64 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-00838192 , version 1 (25-06-2013)

Identifiers

  • HAL Id : hal-00838192 , version 1

Cite

Emmanuel Lassalle, Pascal Denis. Improving pairwise coreference models through feature space hierarchy learning. ACL 2013 - Annual meeting of the Association for Computational Linguistics, Association for Computational Linguistics, Aug 2013, Sofia, Bulgaria. ⟨hal-00838192⟩
333 View
96 Download

Share

Gmail Facebook Twitter LinkedIn More