Paper: Multilingual Semantic Role Labeling

ACL ID W09-1206
Title Multilingual Semantic Role Labeling
Venue International Conference on Computational Natural Language Learning
Session shared task
Year 2009

This paper describes our contribution to the semantic role labeling task (SRL-only) of the CoNLL-2009 shared task in the closed chal- lenge (Hajiˇc et al., 2009). Our system con- sists of a pipeline of independent, local clas- sifiers that identify the predicate sense, the ar- guments of the predicates, and the argument labels. Using these local models, we carried out a beam search to generate a pool of candi- dates. We then reranked the candidates using a joint learning approach that combines the lo- cal models and proposition features. To address the multilingual nature of the data, we implemented a feature selection procedure that systematically explored the feature space, yielding significant gains over a standard set of features. Our system achieved the second best semantic score ove...