Self-Study Recommendations

The webpage will receive minor updates aligned with the school program.

The following are recommended topics and materials for a voluntary self-study. Some level of familiarity with the listed topics should help you follow and digest the content of ProbAI. We have organized the recommended self-study materials into the following groups:

• Preliminaries: It is essential to have a good understanding of topics 1, 2, and 4.
• Tutorials: To follow tutorials, familiarity with Python and the basics of PyTorch and Pyro (topic 5) are beneficial.
• Supplementary: Topics 3, 6, 7, and 8 may be helpful to follow some of the more advanced lectures.

Topics and materials:

1. Probability theory and Bayesian analysis:
   • [1] chapters 2.1, 2.2, 2.3, 3.1
   • [2] chapters 1.2, 1.3, 1.5.1, 1.5.2
2. Gaussian (normal) distribution:
   • [1] chapters 2.6, 3.2
   • [2] Chapter 2.3
3. The exponential family:
   • [1] Chapter 3.4
   • [2] Chapter 2.4
4. Neural networks, backpropagation:
   • [1] chapters 13.1, 13.2, 13.3.1-13.3.4,
   • [2] chapters 5.1-5.3
   • 3Blue1Brown: Neural networks
   • Additional reading on gradient descent: Sebastian Ruder: An Overview of gradient descent optimization
5. Deep learning frameworks:
   • PyTorch:
     • PyTorch QuickStart
     • PyTorch Tensors
     • PyTorch Automatic Differentiation
     • PyTorch Optimization (optionally)
   • Probabilistic programming with Pyro:
     • Bayesian Regression – Introduction (Part 1)
     • Bayesian Regression – Inference Algorithms (Part 2)
     • Introduction Part I Introduction Part II
     • SVI Part I (additionally)
6. Automatic variational inference (days 3 and 4):
   • Edward: Background on Variational Inference
   • Edward: Background on KL(q|p) minimization (automatic and black-box inference)
   • Follow-up reading: Kucukelbir: Automatic Differentiation Variational Inference
7. Variational autoencoders (day 3):
   • Kingma & Welling: An Introduction to Variational Autoencoders
8. What is ODE (day 5):
   • An introduction to ODEs
   • 3Blue1Brown: Differential equations, a tourist’s guide

[1] Probabilistic Machine Learning: An Introduction (draft), Kevin Patrick Murphy. MIT Press, 2021.
[2] Pattern Recognition and Machine Learning, Christopher Bishop. Springer, 2016.