CBE/ENGR 225 Faculty Seminar: Assistant Professor Alberto Busetto, Department of Electrical and Computer Engineering and Department of Computer Science
Alberto Busetto, Ph.D.
Assistant Professor, Department of Electrical Engineering and Computer Engineering
Department of Computer Science
University of California, Santa Barbara
Tuesday, November 17, 2015
4:00 pm – 5:00 pm
*coffee & cookies served*
Closing the Loop: from Experimental Design to Predictive Modeling
Biological systems are often understood after several iterations of modeling and experimentation. In this talk, I discuss model generation, model selection and experimental design in biology. I introduce a general framework which enables the automation of several aspects of the scientific method. The ultimate goal is two-fold: to make the process of predictive modeling more rigorous, and to improve efficiency by compressing the design-test-learn cycle. The talk starts by introducing methods for the systematic generation of hypothetical models. The models are then evaluated, selected and validated on the basis of new information-theoretic principles to strike the optimal balance between model complexity and statistical robustness. Formal guarantees of near-optimality are discussed for the design of informative experiments, to close the loop between modeling and experimentation. By combining uncertainty quantification, information-theoretic learning and experimental design, we aim at accelerating discovery in biomedicine and bioengineering. I conclude by reporting how such framework helped solve concrete problems in biocomputing, systems biology, and precision medicine.