Ted Pedersen is a professor of computer science at the University of Minnesota Duluth who does research in natural language processing. His lab has developed several unsupervised word sense discrimination techniques based on clustering similar contexts without requiring sense-tagged training data. This approach uses local lexical features and measures similarities between contexts to cluster words into senses. Evaluation on Senseval-2 data showed this approach outperformed a majority sense classifier baseline.