GS-Agent: Creating 4D Physical Worlds With Generative Simulation
GS-Agent turns language prompts into simulated 4D scenes by coordinating specialist agents around a physics engine.
The paper describes a multi-agent system that handles assets, materials, placement, motion, camera, and lighting through code-driven interaction with simulation tools. Its goal is to make generated worlds more physically plausible and controllable than data-only generation methods. The authors report examples involving liquids, deformable objects, and rigid bodies, with iterative multimodal feedback used to refine the scene. ArXiv · AI/CL/LG's note
The paper describes a multi-agent system that handles assets, materials, placement, motion, camera, and lighting through code-driven interaction with simulation tools. Its goal is to make generated worlds more physically plausible and controllable than data-only generation methods. The authors report examples involving liquids, deformable objects, and rigid bodies, with iterative multimodal feedback used to refine the scene. ArXiv · AI/CL/LG's note
score 5