Title: Using Random Matrix Theory to Understand Generalization in Graph Neural Networks
Ratul Biswas is an NITMB Fellow. He received his PhD in Mathematics from the University of Minnesota, where he worked under the supervision of Wei-Kuo Chen and Arnab Sen. His research lies in probability theory, with a focus on biologically inspired complex systems and stochastic models that bridge statistics, physics, biology, and computer science. He studies rugged fitness landscapes as models of evolutionary search, examines how network-level similarity and connectivity influence generalization and overparameterization effects in graph-based learning architectures, and analyzes the dynamics of non-reciprocally coupled systems to understand emergent behavior in interacting populations.
Members of the NITMB community are invited to join us for Work-In-Progress meetings, an informal venue an informal venue for members of the NITMB to discuss ongoing and/or planned research.
Tiffany Leighton
Email