Machine learning for ARchitecture and compilation (SMART) Lab (run by Dr. John
Cavazos), University of
Delaware: June 2009 – present
the detection of intrusions on virtual machines using support vector
machines, in particularly the possibility of real-time detection using
methods to optimize CUDA code, including using different modes of
storage (local memory, shared memory, and registers) and loop unrolling
Modeling and Synthesis Laboratory (run by Dr. Chandra
Kambhamettu): September 2006-June 2009
of course projects:
tracking the motion of sea ice from Satellite Imagery, traveled to
APLIS '07 ice camp in the arctic.
stereo/motion algorithms with the goal of tracking cloud motion from a
sequence of satellite imagery, with a focus on taking advantage of the
GPU to allow the results to be processed in real-time.
photon mapping on the GPU using CUDA as a project for an advanced
graphics course entitled 'Rendering and Image
Programmed a CUDA implementation of autostereoscopic image
generation using a light field for a Computational Photography course.