Preprint
Aug 2026
Beyond Instance Slots: Semantically Rich World Models for Physical Interaction Planning
The Semantically Rich World Model is presented, a task-conditioned world model structured around five functional roles: gripper, target, goal, relation, and phase, which transforms object-centric prediction into a semantic interface linking visual dynamics with planning-oriented decision making.
Juntao Cheng, Jingkai Wang, Yi-Jun Shen et al.
· 1 citation