CIS Colloquium, Nov 23, 2011, 11:00AM - 12:00PM, Wachman 447

CIS Colloquium, Nov 23, 2011, 11:00AM - 12:00PM, Wachman 447

Moving Between Tasks and Domains

Hal Daumé III, University of Maryland, College Park

Often it is easier to learn to solve a task once you've already learned to solve other, related tasks. In one case, perhaps the task you want to solve stays the same (eg., natural language parsing) but you move from one domain to another (eg., news to twitter feeds). In another case, the task itself might change (eg., going from parsing to named entity recognition). I will describe several recent approaches to these problems and their generalizations, particularly addressing issues having to do with active and semi-supervised learning, as well as learning when the notion of a "domain" is fuzzier than just "source to target." Most of these algorithms are easy to implement and have nice theoretical properties.

Hal Daumé III is an assistant professor in Computer Science at the University of Maryland, College Park. He holds joint appointments in UMIACS and Linguistics. He was previously an assistant professor in the School of Computing at the University of Utah. His primary research interest is in developing new learning algorithms for prototypical problems that arise in the context of language processing and artificial intelligence. This includes topics like structured prediction, domain adaptation and unsupervised learning; as well as multilingual modeling and affect analysis. He associates himself most with conferences like ACL, ICML, NIPS and EMNLP. He earned his PhD at the University of Southern California with a thesis on structured prediction for language (his advisor was Daniel Marcu). He spent the summer of 2003 working with Eric Brill in the machine learning and applied statistics group at Microsoft Research. Prior to that, he studied math (mostly logic) at Carnegie Mellon University. He still likes math and doesn't like to use C (instead he uses O'Caml or Haskell). He doesn't like shoes, but does like activities that are hard on your feet: skiing, badminton, Aikido and rock climbing.

© 2001-2013 Center for Data Analytics and Biomedical Informatics, Temple University