/probinf-spring-2023

Probabilistic Inference Course, Department of Computing, Imperial College London, Spring 2023.

Primary LanguageHTML

Probabilistic Inference

Mark van der Wilk

Lecture materials for the Imperial College London course on Probabilistic Inference, in the Department of Computing. Lecture recordings are available on Panopto.

Contributing

I welcome any suggestions and fixes from students. If you want to make a contribution, please fork the repo, and make a pull request. If you make a fix, you can claim a chocolate bar at a lecture.

Plan for the Term

Overview:

  • L1: Course overview
  • L1: Building Probabilistic Models
  • L2: Graphical models

Gaussian Processes and the Behaviour of Bayesian Inference

  • L3: Priors on functions
  • L3: From Linear Models to Gaussian Processes
  • L4: Gaussian Processes
  • L5: Model Selection
  • L6: Marginal Likelihood
  • L6: GP Limitations & Challenges

From Beliefs to Actions

  • L7: Decision Theory
  • L8: Bayesian Optimisation

Approximate Inference

  • L9: Conjugate and Non-Conjugate Models
  • L9: Logistic Regression
  • L9: Monte Carlo
  • L10: Markov Chain Monte Carlo
  • L11: Variational Inference
  • L12: Stochastic & Amortised Variational Inference, VAEs

The Present and the Future

  • L13: Diffusion Models
  • L14: Epilogue (non-examinable)