DynaMITE-RL: A Dynamic Model for Improved Temporal Meta-Reinforcement Learning.
Anthony LiangGuy TennenholtzChih-Wei HsuYinlam ChowErdem BiyikCraig BoutilierPublished in: CoRR (2024)
Keyphrases
- dynamic model
- reinforcement learning
- state space
- experimental data
- reinforcement learning algorithms
- function approximation
- model free
- optimal policy
- learning algorithm
- rl algorithms
- optimal control
- temporal difference
- markov decision processes
- machine learning
- direct policy search
- multiple models
- learning problems
- transfer learning
- learning classifier systems
- control scheme
- robot manipulators
- dynamic programming
- autonomous learning
- parallel manipulator
- action selection
- action space
- actor critic
- trajectory tracking
- partially observable domains
- image sequences