Online Transfer Learning in Reinforcement Learning Domains

Abstract

This paper proposes an online transfer framework to capture the interaction among agents and shows that current transfer learning in reinforcement learning is a special case of online transfer. Furthermore, this paper re-characterizes existing agents-teaching-agents methods as online transfer and analyze one such teaching method in three ways. First, the convergence of Q-learning and Sarsa with tabular representation with a finite budget is proven. Second, the convergence of Q-learning and Sarsa with linear function approximation is established. Third, the we show the asymptotic performance cannot be hurt through teaching. Additionally, all theoretical results are empirically validated.

18 pages, 2 figures

References in corpus (1)

• Regret Bounds for Reinforcement Learning with Policy Advice

Cited by in corpus (8)

• Transfer Learning in Deep Reinforcement Learning: A Survey
• One for Many: Transfer Learning for Building HVAC Control
• Learning to Teach Reinforcement Learning Agents
• Transfer Learning to Learn with Multitask Neural Model Search
• Self-Organizing Maps as a Storage and Transfer Mechanism in Reinforcement Learning
• Online Transfer Learning: Negative Transfer and Effect of Prior Knowledge
• A Bayesian Approach to (Online) Transfer Learning: Theory and Algorithms
• Shared Learning : Enhancing Reinforcement in $Q$-Ensembles