In its first decade, Cross-Language Information Retrieval (CLIR) was a specialized topic pursued mainly by information retrieval researchers. Today, it is an integrating technology that draws together a broad range of language technology. In this talk, we will describe our work in the IARPA MATERIAL program, in which our goal is to integrate text processing, speech processing, machine translation, information retrieval and summarization to make it possible for people who know one language to find relevant content in another. After describing our architecture and our most recent results, we will use the rest of our time to discuss open research questions that arise in the context of this driving application.
Douglas Oard is a Professor in the iSchool and UMIACS. He is perhaps best known in this context for having introduced the idea of “surprise language” evaluations, an idea so compelling that after the first series of surprise language evaluations his colleagues made him promise never to suggest such a thing again. Marine Carpuat is an Assistant Professor in the Computer Science department and UMIACS. Petra Galuscakova is a postdoc in UMIACS. Suraj Nair, Aquia Richburg, and Weijia Xu are Ph.D. students in Computer Science. Rui Zhang is a Ph.D. student at Yale.