AUSTIN, Texas - February 14, 2007
The CAREER Program offers the National Science Foundation's most prestigious awards in support of the early career-development activities of those teacher-scholars who most effectively integrate research and education within the context of the mission of their organization. These CAREER awards, in the amount of $400,000, recognize outstanding scientists and engineers who, early in their careers, show exceptional potential for leadership at the frontiers of knowledge and is the highest honor bestowed by the United States Government on scientists and engineers beginning their independent careers.
"The Computational Complexity of Halfspace-Based Learning." My proposal aims to develop a set of powerful new algorithms in machine learning for halfspace-based classification tasks. Halfspace-based learning is at the core of many of today's most important learning tools such as Support Vector Machines and Boosting. I am also working towards proving lower bounds on the computational complexity of classical halfspace learning problems that have remained open for many years, such as learning the intersection of two halfspaces. Klivans' home page
"Operating System Support For Transactional Memory: Construction and Performance Scalability of Parallel Programs." As CPU manufacturers have shifted their performance improvement strategy from scaling clock frequency to placing multiple, simple processor cores on one chip, there has been a renewed interest in concurrent programming. My proposal addresses the problem of designing operating systems to work with transactional memory, which is a hardware feature that has been proposed to help make concurrent programming easier. The hardware and OS should work together to provide transactional memory, just as they work together to provide virtual memory. The challenges include identifying ways the OS can eliminate work for the hardware, a proper interface for the OS to express hints to hardware, and a programmer model for system calls within a transaction. Witchel's home page
For more information: