Index Statistical Machine Learning

Textbook: "Pattern Recognition and Machine Learning" (C.M. Bishop, 2006).

Exam: January 21. Preferably physical, otherwise a "home" exam.

No midterm exam.

Previously, the course followed the textbook chronologically. Currently, the course is converted to be "module-based".

I think the lecturer said the course will involve a little less mathematical details than before, but I'm not sure I got that correctly.

Course Contents:

  • Linear regression (wk. 1-2)
  • Probabilities, max. likelihood, Bayesian regression and model selection (wk.3-5)
  • Classification, logistic regression, and decision theory (wk. 6-7)
  • Neural networks (wk. 8)
  • Kernel methods and Gaussian Processes (wk. 9)
  • Mixture models and EM (wk. 10)
  • Sampling methods (wk. 11)
  • Variational inference (wk. 12)
  • Causal inference (wk. 13, optional)

Other notes:

Linear regression, gradient descent, regularization:

Classification and decision theory:

Discusses:

  • 1.1: Polynomial Curve Fitting
  • 5.2.1 & 5.2.4: Gradient descent
  • 3.1: Linear Basis Function models
  • Appendix E: Lagrange multipliers

Probabilities

Decision theory

Neural Networks

Mixture models

SML-9-Inference_in_mixture_models.md

Gaussian Processes

SML-10-Gaussian_processes.md

See this visual exploration of GPs as well.

Variational inference

SML11-Variational_Inference.md

Markov Chain Monte Carlo

TODO

Exercises checklist

  • [ ] Wk. 1
  • [ ] Wk. 2
  • [X] Wk. 3
  • [ ] Wk. 4
  • [ ] Wk. 5
  • [ ] Wk. 6
  • [ ] Wk. 7
  • [ ] Wk. 8
  • [ ] Wk. 9
  • [ ] Wk. 10