Linguistics Department

Colloquium - Matt Lease "Crowdsourcing for Natural Language Processing"

Mon, February 28, 2011 | UTC 4.134

3:00 PM - 4:30 PM

The recent emergence of crowdsourcing and coupled re-discovery of human computation are radically transforming practice in academic research and industry alike. For example, one of the most prominent examples of crowdsourcing, Amazon's "Mechanical Turk", has in the five years since its inception quickly become a sensation in academic research across disciplines. To date, crowdsourcing research has largely broken into two camps: those studying the phenomenon, and those utilizing it.

Human language technology and computational linguistics have largely fallen into the latter camp, with crowdsourcing having altered the methodological landscape for annotating and collecting linguistic data, conducting behavioral user studies, efficiently training language systems, and designing new language applications which blend automation with human computation in real-time.

In this talk, I will briefly introduce crowdsourcing, its connections to language processing applications and studies, and future trends and challenges.

Matt Lease is an Assistant Professor in the School of Information at the University of Texas at Austin.  His research spans information retrieval (IR) and natural language processing (NLP) along with crowdsourcing. Lease's dissertation studied search queries posed naturally as questions, and his earlier NLP work analyzed text and spontaneous speech for syntax and dysfluency. He holds a Ph.D. and M.Sc. in Computer Science from Brown University.

Bookmark and Share