Colloquium: Gus Hart, January 16, 2014

Date: January 16
Speaker: Gus Hart (BYU)
Title: Compressive Sensing as a New Paradigm for Building Models
Abstract: When building physical models (truncated expansions, force-fields, etc.), one often employs the widely-accepted intuition that the physics is determined by a few dominant terms. But this reductionist paradigm is limited because the intuition for identifying the those terms often does not exist or is difficult to develop. Machine learning algorithms (genetic programming, neural networks, Bayesian methods, etc.) attempt to eliminate the a priori need for such intuition but often do so with increased computational burden and human time. Compressive sensing (a new technique in the field of signal processing) provides a simple, general, and efficient solution to this challenge. I will give a tutorial illustrating the basic ideas of CS and then show that CS-generated models are just as robust as those built by current state-of-the-art approaches, but can be constructed at a very small fraction of the computational cost and user effort. I believe that CS could provide a revolutionary advance for some instances of model building.