3.0

credits

Average Course Rating

Probabilistic machine learning harnesses the power of probability theory to provide models for complex data, as well as the algorithms that enable learning, inference, sampling, and decision-making for these models. The first part of the course will cover classical approaches based on directed and undirected probabilistic graphical models, including latent variable models and temporal models. We develop a toolkit of algorithms used for learning and inference, including message passing algorithms, Markov chain Monte Carlo, and variational inference. Building on these foundational ideas, the second part of the course will cover modern (deep learning) approaches to generative modeling such as variational auto-encoders, generative adversarial networks, normalizing flows, and diffusion models. A background in machine learning is recommended.