In this talk, Robert Östling will discuss his work on large-scale Bayesian models for word alignment, that is, matching words across the languages in a translated text.

Robert Östling describes this as "a problem at the core of machine translation systems as well as other applications in linguistics."