Thursday, April 28, 2011
6:15 PM
Meliora Hall, room 366
Regina Barzilay
Computer Massachusetts Institute of Technology
Embracing Language Diversity: Unsupervised Multilingual Learning
For centuries, the deep connection between human languages has fascinated scholars, and driven many important discoveries in linguistics and anthropology. In this talk, I will show that this connection can empower unsupervised methods for language analysis. The key insight is that joint learning from several languages reduces uncertainty about the linguistic structure of each individual language.

First, I will present a new class of probabilistic unsupervised model that use these multilingual links as a form of naturally occurring supervision for language analysis. Then I will show that these methods also enable us to solve a novel task, beyond the scope of past techniques: the automatic decipherment of ancient lost languages. I will present results on the unsupervised translation of ancient Ugaritic tablets using data from related Semitic languages.

This is joint work with Benjamin Snyder, Tahira Naseem and Jacob Eisenstein.

RSVP: Because we need to provide enough food, please let Kathy Corser ( know if you will be attending (and if you have dietary restrictions). Thanks.