Anastasis Kratsios

Anastasis Kratsios is an Assistant Professor in the Department of Mathematics and Statistics at McMaster University. His research focuses on the analytic and statistical foundations of deep learning, particularly in encoding structures from natural and social sciences like stochastic processes, games, and PDEs. He teaches courses including Computational Finance, Mathematical Reasoning, and Calculus for Science.

Research interests span machine learning theory, approximation methods, and their applications in finance and geometry. Notable work includes neural operator design for PDEs, causal deep learning models, and federated learning frameworks. He frequently publishes in top venues like NeurIPS, ICLR, and Journal of Machine Learning Research.

Teaching responsibilities include MFM 713 (Computational Finance II), MATH 1C03 (Introduction to Mathematical Reasoning), and MATH 1A03 (Calculus for Science I). He has developed innovative curricula blending theoretical rigor with practical computational finance tools.

Publications emphasize theoretical guarantees for neural networks, kernel methods, and geometric deep learning. Recent work explores universal approximation under constraints, adversarial learning in market models, and scalable graph representation techniques. His research bridges mathematical foundations with real-world applications in finance, physics, and engineering.