Residual learning and context encoding for adaptive offline-to-online reinforcement learning.
Mohammadreza NakhaeiAidan ScannellJoni PajarinenPublished in: L4DC (2024)
Keyphrases
- reinforcement learning
- learning algorithm
- online learning
- learning process
- learning capabilities
- contextual information
- inductive inference
- adaptive learning
- prior knowledge
- policy search
- dynamic programming
- learning systems
- mobile learning
- learning problems
- online environment
- eligibility traces
- reinforcement learning methods
- adaptive control
- real time
- passive aggressive
- actor critic
- temporal difference learning
- model free
- function approximation
- optimal policy
- supervised learning
- multi agent
- machine learning