Peter Thestrup Waade

Finally, I think we might add (a minimal version of) this to the ADTests.JL suite - I'll happily help with that if I can :)

Note @yebai [commenting here](https://github.com/chalk-lab/Mooncake.jl/issues/651) that autodiff over a Turing sample call shouldn't be applied in general, but that gradient-free sampling should work fine

I'll add that FiniteDifferences.jl is able to autodiff over the model (although extremely slowly). Code is pasted below; feel free to ignore. ``` using Turing using LogExpFunctions using ADTypes: AutoFiniteDifferences...

I am likely to be particularly interested in temporal difference models and SARSA models, which I think can be found in the https://github.com/JuliaPOMDP/TabularTDLearning.jl repository :)

Dear @dylan-asmar Thank your for this incredibly useful response! It is very appreciated. I have been able set up two simple example models which fits your example policy, as well...

Just a soft ping on this issue :) I know you must have quite limited time resources, so no pressure form my side. It would be wonderful if I could...