Fetching the paper…

Learning Dynamic Attribute-factored World Models for Efficient Multi-object Reinforcement Learning · Around