Exploring the correspondence between languages for machine learning
Ekaterina Garmash explores common and distinguishing properties of languages in the context of machine translation. Focusing on the properties common across languages first, she develops and tests bilingual syntactic language models to aid machine translation. She then looks at differences between languages, designing ensembles of neural machine translation systems that share a target language but have a different source language. She then compares these systems' translation performance with that of monolingual ensembles.
E. Garmash: Exploring the Correspondence between Languages for Machine Translation.
Prof. M. de Rijke
Dr C. Monz
Oudezijds Voorburgwal 229 - 231 | 1012 EZ AmsterdamGo to detailpage
+31 (0)20 525 2362
This event is open to the public.