At STOC 2020, I attend the workshop on New frontiers in approximate counting. Before this I had a very cursory overview of MCMC. While I didn’t understand much of the details of the talks, I could appreciate how MCMC could be such a powerful tool. After the workshop I asked the presenters on Slack for some resources to get into the field. One of the first suggestions I received was this survey by Prof. Dana Randall. I found it to be a perfect overview of the several techniques to bound the mixing time of MCMC samplers, while at the same time being super accessible to a beginner. I was also really appreciative of a table that converted some daunting sounding statistical physics terms into friendlier theoretical CS terms.
Here are the notes I took while I read this. I usually write these because it helps me focus and so they were mostly made for my own reference. However I decided to put them up regardless.