Your question is Diffusion and RL for Manipulation. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
Explain how you would apply diffusion policy and reinforcement learning to mobile manipulation, including the advantages and disadvantages of different action space formulations and how you would go beyond data scaling.