Stuart Geman

Stuart Geman is the James Manning Professor of Applied Mathematics at Brown University's Division of Applied Mathematics. He holds a Ph.D. from MIT (1977), focusing on stochastic differential equations. His research spans machine and natural vision, statistical theory, neuroscience, financial modeling, and computational linguistics. He has advised numerous students, including Asohan Amarasingham, Lo-Bin Chang, and Ya Jin. His work explores hierarchical models for visual recognition, neural spike train analysis, market dynamics, and generative image modeling.

Key contributions include the 'compositionality' framework for efficient learning in biological systems, probabilistic image models with hierarchical structure, and statistical methods for neurophysiological data. He co-developed the Gibbs sampling method for image restoration and pioneered nonparametric statistical estimation via sieves. His recent work addresses scale invariance in natural images and transsaccadic neural coding in macaque V1.

Publications span journals like Journal of Neuroscience, Neural Computation, and Proceedings of the National Academy of Sciences. His research integrates computational, statistical, and biological perspectives to address challenges in vision, neuroscience, and financial systems.