Diffusion-based Dynamics Models for Long-Horizon Rollout in Offline Reinforcement Learning.
Hanye ZhaoXiaoshen HanZhengbang ZhuMinghuan LiuYong YuWeinan ZhangPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- real time
- probabilistic model
- experimental data
- model selection
- parameter estimation
- model free
- complex systems
- statistical models
- dynamic model
- transition model
- diffusion models
- reinforcement learning algorithms
- function approximation
- learning tasks
- medical images
- learning process
- multiscale
- learning algorithm
- neural network