Unified 3D Scene Understanding Through Physical World Modeling
Perceptual 3D Simulation with Physical World Modeling
World Modeling with Probabilistic Structure Integration
Unifying (Machine) Vision via Counterfactual World Modeling
Discovering and using Spelke segments
Taming generative video models for zero-shot optical flow extraction
Understanding Physical Dynamics with Counterfactual World Modeling
Unsupervised 3D Scene Representation Learning via Movable Object Inference