New website for sequential decision analytics
Warren Powell
To the RL community,
I am pleased to announce my new website on "sequential decision analytics" which can be found at
https://tinyurl.com/SDAwebsite/
The menu on the right contains:
➤ "Ask Professor Powell" - This is an LLM-based chatbot that will answer questions based on my five books, over 1000 pages of LinkedIn posts, and a large number of educational webpages. Please give this a try – the answers can be a little unpredictable, but over 2,000 people have used the chatbot, sometimes asking fairly elaborate questions.
➤ Sequential decision problems - I define a SDP, discuss the diversity of applications, list the many communities that study this problem (RL is just one). There is a segment on “Optimal Learning” (aka stochastic search, multiarmed bandit problems), followed by a comparison of “reinforcement learning” and “sequential decision analytics”.
➤ Modeling – I span the gap from real problems to models, starting with “What is a decision” and “Framing decision problems” (with no math) to the universal modeling framework and state variables. I end with a section on uncertainty which lists the 12 ways that uncertainty can enter a model.
➤ Policies – I present all four classes of policies. Examples of each of these can be sound in the 2nd edition of Sutton and Barto. The third class is the only one that includes Bellman’s equation. I also discuss hybridsd.
➤ Teaching materials – This includes:
» Courses, covering different ways this material can be taught,
» Books, from my 1100 page “Reinforcement Learning and Stochastic Optimization” to a new book on framing with no mathematics.
» Videos, including a growing list of videos on different topics.
➤ Extra materials – This currently includes a list of motivating applications which are a list of some of the projects that motivated my own work, and a page describing my “7 levels of AI.”
➤ About -This lists the people of CASTLE lab who provided the foundation of this work, a timeline of major developments, awards and project sponsors.
Enjoy!
Warren
