Created: July 19, 2022
Modified: July 19, 2022
Modified: July 19, 2022
imagination rollout
This page is from my personal notes, and has not been specifically reviewed for public consumption. It might be incomplete, wrong, outdated, or stupid. Caveat lector.References:
- Gu et al., Continuous Deep Q-Learning with Model-based Acceleration (2016).
A technique used in model-based rl, where we run the current policy on our model to generate an on-policy fictional trajectory. This trajectory is then fed to a model-free RL algorithm.