University of Cambridge > Talks.cam > Computer Laboratory Wednesday Seminars > Statistical Parsing of Natural Language with Lexicalised Grammars

Statistical Parsing of Natural Language with Lexicalised Grammars

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

If you have a question about this talk, please contact Mateja Jamnik.

In this talk I will describe the problem of natural language parsing, and explain why it is such a difficult problem. I will characterise the problem in terms of three main questions: 1) how to obtain a grammar for natural language; 2) how to obtain all possible parses for an input sentence; and 3) how to use machine learning to obtain a model of parse plausibility. The answer to 1) will focus on Combinatory Categorial Grammar (CCG), a ‘lexicalised’ grammar formalism with a long history in theoretical linguistics. As well as having some desirable linguistic and formal properties, I will show that that the use of CCG leads to a surprisingly efficient practical parser. The answer to 2) will focus on standard chart-parsing techniques. The answer to 3) will focus on the perceptron, a simple but highly effective method which can be applied to complex structured learning problems.

This talk is part of the Computer Laboratory Wednesday Seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2020 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity