“Weaving together machine learning, theoretical physics, and neuroscience?
Department of AppliedPhysics
Abstract: Anexciting area of intellectualactivity in this century may well revolve around a synthesis of machinelearning, theoretical physics, and neuroscience. The unification of these fields will likelyenable us to exploit the power of complex systems analysis, developed intheoretical physics and applied mathematics, to elucidate the design principlesgoverning neural systems, both biological and artificial, and deploy theseprinciples to develop better algorithms in machine learning. We willgive several vignettes in this direction, including: (1) determining the best optimization problemto solve in order to perform regression in high dimensions; (2)finding exact solutions to the dynamics of generalization error in deep linearnetworks; (3) developing interpretable machine learning to derive andunderstand state of the art models of the retina; (4) analyzing and explainingthe origins of hexagonal firing patterns in recurrent neural networks trainedto path-integrate; (5) delineating fundamental theoretical limits on theenergy, speed and accuracy with which non-equilibrium sensors can detectsignals.
Bio: SuryaGanguli triple majored in physics,mathematics, and EECS at MIT, completed a PhD in string theory at Berkeley, anda postdoc in theoretical neuroscience at UCSF. He is now an associate professorof Applied physics at Stanford where he leads the Neural Dynamics andComputation Lab. His research spans the fields of neuroscience, machinelearning and physics, focusing on understanding and improving how bothbiological and artificial neural networkslearn striking emergentcomputations. He has been awarded a Swartz-Fellowship in computationalneuroscience, a Burroughs-Wellcome Career Award, a Terman Award, a NeurIPS Outstanding Paper Award, a Sloanfellowship, a James S. McDonnell Foundation scholar award in human cognition, aMcKnight Scholar award in Neuroscience, a Simons Investigator Award in themathematical modeling of living systems, and an NSF career award.