MCMC convergence
(1.3 hours to learn)
Summary
Markov chain Monte Carlo (MCMC) samplers eventually converge to their stationary distribution, but they may take a long time to do so. The "mixing time" of a chain refers to how long a chain must be run in order for one sample to be independent of another. Diagnosing mixing time is important for judging the reliability of estimates obtained from an MCMC algorithm.
Context
This concept has the prerequisites:
- Markov chain Monte Carlo
- covariance (The accuracy of MCMC depends on the covariance of the function at different steps.)
Core resources (read/watch one of the following)
-Free-
→ Machine learning summer school: Markov chain Monte Carlo (2009)
→ Coursera: Probabilistic Graphical Models (2013)
An online course on probabilistic graphical models.
Location:
Lecture "Using a Markov chain"
Other notes:
- Click on "Preview" to see the videos.
-Paid-
→ Machine Learning: a Probabilistic Perspective
A very comprehensive graudate-level machine learning textbook.
Location:
Section 24.4, pages 856-862
→ Probabilistic Graphical Models: Principles and Techniques
A very comprehensive textbook for a graduate-level course on probabilistic AI.
Location:
Section 12.3.5, pages 518-526
See also
-No Additional Notes-