Paper: Training a Korean SRL System with Rich Morphological Features

ACL ID P14-2104
Title Training a Korean SRL System with Rich Morphological Features
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 2014
Authors

In this paper we introduce a semantic role labeler for Korean, an agglutinative lan- guage with rich morphology. First, we create a novel training source by semanti- cally annotating a Korean corpus contain- ing fine-grained morphological and syn- tactic information. We then develop a su- pervised SRL model by leveraging mor- phological features of Korean that tend to correspond with semantic roles. Our model also employs a variety of latent morpheme representations induced from a larger body of unannotated Korean text. These elements lead to state-of-the-art per- formance of 81.07% labeled F1, represent- ing the best SRL performance reported to date for an agglutinative language.