Machine translation offers new opportunities for crossing language barriers in our globalised world, the results are often lacking in quality. One core problem is that hierarchical statistical machine translation (SMT) uses little context when composing rules into translations, leading to independently made and poorly coordinated ordering decisions. The aim of Gideon Maillette de Buy’s research is to improve hierarchical SMT using the hierarchical translation equivalence relations induced from word alignments.
G.E. Maillette De Buij Wenniger: Aligning the Foundations of Hierarchical Statistical Machine Translation.
Prof. K. Sima'an
This event is open to the public.