An offline RL policy can be improved at test time by inferring a latent belief over environment dynamics from past transitions and planning with model-based rollouts averaged over that belief.
Deep reinforcement learning in a handful of trials using probabilistic dynamics models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Reflect-then-Plan: Offline Model-Based Planning through a Doubly Bayesian Lens
An offline RL policy can be improved at test time by inferring a latent belief over environment dynamics from past transitions and planning with model-based rollouts averaged over that belief.