What is wrong with RL

83 views
Skip to first unread message

Warren Powell

unread,
Jan 9, 2023, 10:16:27 AM1/9/23
to Reinforcement Learning Mailing List

Tamar Aviv and noted RL scholar Shie Mannor have published an article “Towards Deployable RL - What’s Broken with RL Research and a Potential Fix” including the graph below.


I have been harping for some time on the problems with RL research.  Here is a formal admission from within the RL community which you can download from https://tinyurl.com/what-is-wrong-with-RL/.  


Told ya!


But they miss the fundamental problem with RL.  If you want to solve important, high impact problems, you need to be able to model them properly, and the RL modeling framework does not allow this.


Chapter 9 of https://tinyurl.com/RLandSO/ describes how to model sequential decision problems…. any sequential decision problem (it can be downloaded from the webpage).  Then see the four classes of policies in chapter 11 for making decisions (also downloadable).   


Now go solve real problems with practical, implementable policies.


Please share this!


image.png

------------------------------
Warren B. Powell
Chief Innovation Officer, Optimal Dynamics
Professor Emeritus, Princeton University
Reply all
Reply to author
Forward
0 new messages