Sara Magliacane

DAFT-RL enhances learning by focusing on attributes and interactions of objects.

This work addresses adapting causal representations to different settings effectively.

Learning hidden factors from incomplete data in complex systems.