Thank you!
GAIL and DAGGER were on my radar, and both are good papers (I am a DAGGER fan!). I'd read your DQfD paper years ago. I like the idea of combining both expert data and RL. This is something we developed back in 2013 (you cited it as Kim et al. 2013).
www.sologen.net/papers/APID%...

Thank you!
Bagnell's paper is nice intro.
I just found this "Is Behaviour Cloning All You Need" earlier today. Looks very interesting.

Thank you! GAIL was under my radar, and is a good paper.
Thanks for bringing the DICE family to my attention. That's exactly what I wanted to find here.
(Haven't read the Diffusion Policies paper closely.)

I am asking for a course I am designing. I have some papers in mind, but I want to make sure I am not missing a good paper.

This is from Michael Littman's Markov games as a framework for multi-agent reinforcement (1994). Simple, yet powerful!
courses.cs.duke.edu/cps296.3/spr...

The theory of Markov Decision Processes (MDP’s), which underlies much of the recent work on reinforcement learning, assumes that the agent’s environment is stationary and as such contains no other adaptive agents."

cf. Wittgenstein's last statement in Tractatus: "What we cannot speak about we must pass over in silence".

No, it is not virtue signalling! :P

How many bits of "creative" information, injected in the form of nudging prompts, are needed in order to get to a point in the space of ideas that is truely novel?

A related though: Some people may say LLMs can just generate ideas similar to what they have seen in their data. We can argue about that, but even if they do, one can ask a question similar to Aha! Distance:

How many of these questions are needed in order to successfully transition from TO to TN? This is what I call Aha! Distance.

The path from TO to TN is paved by some major questions in the form: "... but TO doesn't predict A" or "what if you do X instead of Y?".
A and X might very well be already known at the time of TO being the prevalant theory, but they are not examined and compared with TO.

We know who are on the faculty hiring committees ... !