A reading seminar on learning theory, diffusion models, and language models, with shared topics in scaling laws and neural network optimization.
First meeting: 12 October, 18:40 to 20:30 · Siyuan Hall 225.
Variational inference, diffusion, Monte Carlo, Kalman methods, Langevin and Hamiltonian dynamics, rare events, and simulation based inference.
Regression, uncertainty, resampling, model interpretation, and the statistical workflow from assumptions to diagnosis.
Optimization, generalization, kernels, neural networks, generative models, and the mathematical foundations of modern learning.
Linear systems, geometry, eigenvalues, singular values, least squares, and computational interpretations.
Probability, stochastic processes, optimal transport, diffusion, sequential inference, and inference time computation.
Approximation, PDE learning, operator learning, physics informed models, uncertainty quantification, and scaling laws.