Duetodifficultyofconstructinglarge-scalesemanticrolelabeled corpus for training, unsupervised learning for SRL has been given interest recently ( Titov and Klementiev, 2011; Lang and Lapata, 2011 ). But these unsupervised methods showed significantly lower performance compared to supervised methods.

Titov & Klementiev (2011)[1], Lang & Lapata (2011)[2]

References Edit

  1. Titov, I., Klementiev, A., 2011. A Bayesian model for unsupervised semantic parsing. In: Proc. ACL 2011.
  2. Lang, J., Lapata, M., 2011. Unsupervised semantic role induction with graph partitioning. In: Proc. EMNLP 2011, pp.1320–1331.