University of Cambridge > > NLIP Seminar Series > Error-Aware Probabilistic Parsing

Error-Aware Probabilistic Parsing

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

If you have a question about this talk, please contact Johanna Geiss.

Given the human propensity to err, a parser must be able to produce accurate analyses for sentences which are deviant according to human standards, yet which we routinely interpret correctly. State-of-the- art probabilistic parsers are generally robust to errors, and they will return analyses for most ungrammatical sentences. However, these robust analyses are not necessarily correct because they do not always reflect the meanings of the ill-formed sentences. I present a two- stage “error-aware” probabilistic parsing architecture which uses two versions of a probabilistic parser, one trained on a normal treebank and the other trained on an automatically created ungrammatical version of the original treebank. A binary classifier is used to decide which version to employ. I present the results of experiments carried out using the “error-aware” probabilistic parsing architecture and the Penn Treebank. I also present the results of experiments carried out with various grammatical/ungrammatical classifiers and the BNC .

This talk is part of the NLIP Seminar Series series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


© 2006-2023, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity