News

Philip part of major IARPA grant

Congratulations to Philip Resnik, who is part of Maryland's team on a multi-institutional IARPA grant to develop systems for multi-lingual translation, information retrieval and summarization systems that is effective even with 'low resource' languages, for which there is little data that is digitally accessible. The Terp team, which also includes Marine Carpuat, Hal Daumé III and its head Doug Oard, will join four other institutions on the grant: Edinburgh, Cambridge, Yale and Columbia, its lead institution. The system they are collectively building – called SCRIPTS, or System for Cross Language Information Processing, Translation and Summarization – will take advantage of recent developments in deep learning neural networks to sift through large amounts of linguistic data and extract structural patterns. Importantly, it is designed to incorporate four key areas of language processing – speech recognition, machine translation, cross-language retrieval, and information summarization – into a single robust platform. Says Kathleen McKeown, Professor of Computer Science at Columbia and leader of the group, “intelligence analysts have come to meetings and told us precisely the kind of system they need to be more efficient." Two languages of interest mentioned in press releases are Hausa and Uyghur.