Faculty Research Colloquium-Learning to Learn and Reason Better
|Author:||Daniel Lowd University of Oregon|
|Date:||October 19, 2010|
Statistical machine learning has been successfully applied to many domains, such as computational biology, social network analysis, computer vision, robotics, and more. One of the major hurdles in applying statistical machine learning is efficiency: learning complex models can be very slow, and reasoning is typically NP-hard and difficult to approximate. In this talk, I will show how we can use machine learning methods to speed up learning and inference. First, I will show how decision tree learning can help speed up Markov network learning by orders of magnitude. Then I will show how arithmetic circuits can be trained to perform approximate inference in Bayesian networks. I will conclude with a discussion of ongoing work and an overview of some other projects I have been involved in.
Daniel Lowd is an Assistant Professor in the Department of Computer and Information Science at the University of Oregon. His research interests include machine learning, data mining, and artificial intelligence. He received his Ph.D. in 2009 from the University of Washington. He also received graduate research fellowships from Microsoft Research and the National Science Foundation.