Sam Dillavou

Sam Dillavou is a Visiting Professor in the Department of Physics and Astronomy at the University of Pennsylvania's School of Arts & Sciences. His research focuses on interdisciplinary applications of machine learning to physical systems, granular materials dynamics, and nonlinear physics-driven learning architectures. He explores emergent behaviors in analog circuits and stochastic processes in granular flow clogging, bridging computational methods with experimental physics.

His work spans theoretical and applied domains, including energy-efficient 'smart instruments' and seismic stress analysis. Key areas of investigation include self-training circuits, interpretable machine learning for material science, and the statistical mechanics of granular systems. Dillavou's recent studies emphasize physical learning networks that operate without traditional processors, leveraging analog systems for power-efficient solutions.

Notable contributions include analyzing clog formation in granular hoppers using ML techniques, developing Bellybutton—a customizable image segmentation tool—and studying fault mechanics through seismological stress drop analysis. His research integrates computational modeling with experimental validation, addressing both fundamental physics questions and real-world engineering challenges.

Awards: None explicitly listed.

Advising & Grants: No student advisees listed; active in collaborative projects within the Physics department's experimental and computational research groups. Involved in lab-based studies of granular dynamics and analog computing systems.

Labs/Teams: Affiliated with the Department of Physics and Astronomy's labs focused on nonlinear systems, statistical mechanics, and interdisciplinary computational physics.