Monday, March 28, 2011
11:00 AM
Computer Studies Bldg. Room 209
Brian Kulis
University of California at Berkeley
Learning for Search and Adaptation in Large-scale Data
Machine learning has become ubiquitous, and applications of it are now critical for problems in computer vision, software analysis, finance, biology, and many other areas. One of the most challenging and important research areas in the field of machine learning lies in the development of rich analysis tools that can be applied to massive, complex data. In this talk, I will discuss recent work on developing novel learning methods for data sets where the dimensionality of the data and the number of data points are extremely high. One focus problem is metric learning, which aims to learn a task-specific distance or similarity function given appropriate supervision. For example, how do we compare two program executions? What about images? Another focus problem is that of domain adaptation: how can we take learning models built for one domain and adapt them to new domains? I will develop a unified approach to these and other related problems based on learning linear and non-linear transformations over sets of data points. I will discuss the advances in kernel methods, online learning, and hashing required to make the learning of such transformations practical for very large data sets; furthermore, I will discuss applications of these techniques outside of the two focus problems of the talk. Throughout the talk I will present a variety of results on problems including object recognition, automated software debugging, and image search.

Bio: Brian Kulis is a postdoctoral fellow at UC Berkeley EECS (computer science division) and is also affiliated with the International Computer Science Institute. He obtained his PhD in computer science from the University of Texas in 2008, and his BA degree from Cornell University in computer science and mathematics in 2003. His research focuses on machine learning, data mining, and large-scale optimization. For his research, he has won three best student paper awards at top-tier conferences---two at the International Conference on Machine Learning (in 2005 and 2007) and one at the IEEE Conference on Computer Vision and Pattern Recognition (in 2008). He is also the recipient of an MCD graduate fellowship from the University of Texas (2003-2007) and an Award of Excellence from the College of Natural Sciences at the University of Texas. In the fall of 2007, he was a research fellow at the Institute for Pure and Applied Mathematics at UCLA.

Refreshments will be provided at 10:45AM