Semi-supervised Training of a Statistical Parser from Unlabeled Partially-bracketed Data

Add to your list(s) Download to your calendar using vCal

• John Carroll - Department of Informatics, University of Sussex
• Friday 15 June 2007, 15:00-16:00
• SW01 Computer Laboratory.

If you have a question about this talk, please contact nlip-seminars.

We compare the accuracy of a statistical parse ranking model trained from a fully-annotated portion of the Susanne treebank with one trained from unlabeled partially-bracketed sentences derived from this treebank and from the Penn Treebank. We demonstrate that confidence-based semi-supervised techniques similar to self-training outperform expectation maximization when both are constrained by partial bracketing. Both methods based on partially-bracketed training data outperform the fully supervised technique, and both can, in principle, be applied to any statistical parser whose output is consistent with such partial-bracketing. We also explore tuning the model to a different domain and the effect of in-domain data in the semi-supervised training processes.

This talk is part of the NLIP Seminar Series - Lent 2007 series.

This talk is included in these lists:

• SW01 Computer Laboratory

Note that ex-directory lists are not shown.