Chris Mesterharm's Home Page

Department of Computer Science Rutgers University Piscataway, NJ 08855 Room: 363 Hill Office Hours: mesterha@cs.rutgers.edu

I am a PhD student in the Rutgers Computer Science Department.

My research involves practical extensions to on-line learning algorithms. Here is a link to a short description of my research.

• Experience-Efficient Learning in Associative Bandit Problems. Alexander Strehl, Chris Mesterharm, Michael Littman, and Haym Hirsh. In ICML-2006, pages 889-896, 2006.
• On-line Learning with Delayed Label Feedback. Chris Mesterharm. In Algorithmic Learning Theory, pages 399-123, 2005.
• Tracking Linear-threshold Concepts with Winnow. Chris Mesterharm. In Journal of Machine Learning Research, pages 819-838, 2003.
• Using Linear-threshold Algorithms to Combine Multi-class Sub-experts. Chris Mesterharm. In ICML-2003, pages 544-551, 2003.
• Tracking Linear-threshold Concepts with Winnow. Chris Mesterharm. In Proceedings COLT-2002, pages 138-152, 2002.
• Transforming Linear-threshold Learning Algorithms into Multi-class Linear Learning Algorithms. Chris Mesterharm. Rutgers technical report, dcs-tr-460, 2001.
• A Multi-class Linear Learning Algorithm Related to Winnow. Chris Mesterharm. In Proceedings NIPS-12, pages 519-525, 2000.
• An Apobayesian Relative of Winnow. Nick Littlestone and Chris Mesterharm. In Proceedings NIPS-9, pages 204-210, 1997.

Here is a link to my resume.

• Here is a picture of my nephew/godchild.
• Here is a picture of my other godchild.