The Effectiveness of World Models for Continual Reinforcement Learning

TytułThe Effectiveness of World Models for Continual Reinforcement Learning
Publication TypeConference Paper
Rok publikacji2023
AutorzyKessler S, Ostaszewski M, Bortkiewicz M, Żarski M, Wołczyk M, Parker-Holder J, Roberts S, Miłoś P
Conference NameConference on Lifelong Learning Agents
Conference LocationMontreal, Canada

World models power some of the most efficient reinforcement learning algorithms. In this work, we showcase that they can be harnessed for continual learning – a situation when the agent faces changing environments. World models typically employ a replay buffer for training, which can be naturally extended to continual learning. We systematically study how different selective experience replay methods affect performance, forgetting, and transfer. We also provide recommendations regarding various modeling options for using world models. The best set of choices is called Continual-Dreamer, it is task-agnostic and utilizes the world model for continual exploration. Continual-Dreamer is sample efficient and outperforms state-of-the-art task-agnostic continual reinforcement learning methods on Minigrid and Minihack benchmarks.


Historia zmian

Data aktualizacji: 20/11/2023 - 15:18; autor zmian: Mateusz Żarski (