1 paper
Mehran Aghabozorgi, Alireza Moazeni, Yanshu Zhang +1
Model-based reinforcement learning promises strong sample efficiency but often underperforms in practice due to compounding model error, unimodal world models that average over mul…