I need create interpretability model for reinforcement learning.
My main goal is to make reinforcement learning technique ( anyone ) explainable / transparent for the user.
We can visualize / present future behaviors of actions of the trained algorithm. That would be simple and a good starting point to do something real / bigger in the next gigs as I have lots of work.
Language: Python, you can use TensorFlow, OpenGym AI etc if you want.
A few papers came up in the search. It's a basic approach but might be a good inspiration: [login to view URL]
This might help you understand what I mean by interpretability model: [login to view URL]
So I need someone who want to work in this project