DIME derives a variational lower bound on the maximum entropy RL objective for diffusion policies and shows strong continuous-control benchmark results.
Efficient gradient-free variational inference using policy search
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DIME:Diffusion-Based Maximum Entropy Reinforcement Learning
DIME derives a variational lower bound on the maximum entropy RL objective for diffusion policies and shows strong continuous-control benchmark results.