PhysWorld: From Real Videos to World Models of Deformable Objects via Physics-Aware Demonstration Synthesis
This addresses a data scarcity problem for robotics, VR, and AR applications by enabling efficient world models for deformable objects, though it is incremental as it builds on existing simulation and learning methods.
The paper tackles the challenge of learning physics-consistent dynamics models for deformable objects from limited real-world video data by proposing PhysWorld, a framework that synthesizes diverse demonstrations using a simulator to train a lightweight GNN-based world model, achieving competitive performance with inference speeds 47 times faster than the state-of-the-art method.
Interactive world models that simulate object dynamics are crucial for robotics, VR, and AR. However, it remains a significant challenge to learn physics-consistent dynamics models from limited real-world video data, especially for deformable objects with spatially-varying physical properties. To overcome the challenge of data scarcity, we propose PhysWorld, a novel framework that utilizes a simulator to synthesize physically plausible and diverse demonstrations to learn efficient world models. Specifically, we first construct a physics-consistent digital twin within MPM simulator via constitutive model selection and global-to-local optimization of physical properties. Subsequently, we apply part-aware perturbations to the physical properties and generate various motion patterns for the digital twin, synthesizing extensive and diverse demonstrations. Finally, using these demonstrations, we train a lightweight GNN-based world model that is embedded with physical properties. The real video can be used to further refine the physical properties. PhysWorld achieves accurate and fast future predictions for various deformable objects, and also generalizes well to novel interactions. Experiments show that PhysWorld has competitive performance while enabling inference speeds 47 times faster than the recent state-of-the-art method, i.e., PhysTwin.