The National Institute of Mathematical and Biological Sciences is having an online seminar on ‘adaptive management’. It should be fun for people who want to understand Markov decision processes—like me!

Adaptive management seeks to determine sound management strategies in the face of uncertainty concerning the behavior of the system being managed. Specifically, it attempts to find strategies for managing dynamic systems while learning the behavior of the system. These webinars review the key concept of a Markov Decision Process (MDP) and demonstrate how quantitative adaptive management strategies can be developed using MDPs. Additional conceptual, computational and application aspects will be discussed, including dynamic programming and Bayesian formalization of learning.

Here are the topics:

Session 1: Introduction to decision problems
Session 2: Introduction to Markov decision processes (MDPs)
Session 3: Solving Markov decision processes (MDPs)
Session 4: Modeling beliefs
Session 5: Conjugacy and discrete model adaptive management (AM)
Session 6: More on AM problems (Dirichlet/multinomial and Gaussian prior/likelihood)
Session 7: Partially observable Markov decision processes (POMDPs)
Session 8: Frontier topics (projection methods, approximate DP, communicating solutions)

I needed to read a lot about these MDPs since they are fundamental for Reinforcement Learning (the one that makes a computer formidable opponent in games). Is there any chance it will be recorded?

You can use Markdown or HTML in your comments. You can also use LaTeX, like this: $latex E = m c^2 $. The word 'latex' comes right after the first dollar sign, with a space after it. Cancel reply

You need the word 'latex' right after the first dollar sign, and it needs a space after it. Double dollar signs don't work, and other limitations apply, some described here. You can't preview comments here, but I'm happy to fix errors.

I needed to read a lot about these MDPs since they are fundamental for Reinforcement Learning (the one that makes a computer formidable opponent in games). Is there any chance it will be recorded?

Please ask the folks at NimBioS who are running this.