Siyuan Huang (@siyuanhuang95)

2025-10-19 | โค๏ธ 151 | ๐Ÿ” 21


๐ŸŽ™๏ธ๐ŸŽ™๏ธ๐ŸŽ™๏ธ Excited to share our ICCV 2025 paper: GWM โ€“ Towards Scalable Gaussian World Models for Robotic Manipulation! ๐Ÿค–

GWM learns to predict the future โ€” modeling scene dynamics conditioned on robot actions, and enabling scalable policy training.

With more data, it keeps getting better at understanding the world ๐ŸŒ

๐Ÿ“„ Paper: https://arxiv.org/abs/2508.17600 ๐ŸŒ Project: https://gaussian-world-model.github.io/ ๐Ÿ’ป Code: https://github.com/Gaussian-World-Model/gaussianwm


Auto-generated bookmark

Tags

Robotics AI-ML