|Date||May 05, 2014|
|Title||Towards Universal Natural Language Processing|
|Abstract||In this talk I will first describe our efforts towards a universalrepresentation of syntax that makes it possible to model syntax acrosslanguages in a consistent way. I will then highlight several examples of how wehave successfully used syntax in different applications and discuss how we haveused indirect signals for domain and task adaptation.|
Slav Petrov is a researcher in Google's New York office, leading a team thatworks on syntactic parsing and its applications to information extraction,question answering and machine translation. He holds a PhD degree from UCBerkeley, where he worked with Dan Klein. Before that he completed a Master'sdegree at the Free University of Berlin and was a member of the FU-Fightersteam that won the RoboCup world championship in 2004. His work on fast andaccurate multilingual syntactic analysis has been recognized with best paperawards at ACL 2011 and NAACL 2012. Slav also teaches Statistical NaturalLanguage Processing at New York University.