🗓️ Week 04 - Introduction to Approximate Inference and MCMC

Author

Affiliation

Dr. Simone Rossi

EURECOM

In this week, we will start to analyze the problem of inference in probabilistic models. We will start with an introduction to approximate inference and we will then focus on sampling-based methods. We will cover basic Monte Carlo methods (e.g., rejection sampling) and then we will introduce Markov Chain Monte Carlo (MCMC) methods, which are widely used in practice for sampling from complex distributions. Finally, we will discuss how to use gradient-based methods to improve the efficiency of MCMC sampling, with a focus on Hamiltonian Monte Carlo (HMC).

📑 Lecture Slides

• Introduction to Approximate Inference
• Monte Carlo
• Markov Chain Monte Carlo

Print-friendly version

• Introduction to Approximate Inference
• Monte Carlo
• Markov Chain Monte Carlo

📚 Recommended Reading

• Murphy (2023): Chapter 7.4 (Approximate Inference)
• Murphy (2023): Chapter 11.1, 11.2, 11.3, 11.4 (Monte Carlo methods)
• Murphy (2023): Chapter 12.1, 12.2, 12.5, 12.6 (Markov Chain Monte Carlo)

📙 References

Murphy, Kevin P. 2023. Probabilistic Machine Learning: Advanced Topics. MIT Press. http://probml.github.io/book2.