arXiv Robotics — research abstracts· Era AI Lab Team, Jianhao Lai, Zhihao Zhan, Zhiquan Liao, Qichang Li, Shubin Chen, Zhiyuan Liu, Hongsen Lei, Jianman Lin, Xiaoxin Lin, Ke Gong, Baiyue Gu, Shiming Chen, Junlin He, Longqi Yang, Junrong Xiao, Haolin Luo, Xiaojuan Su, Jianfeng Li, Yu Luo, Xu Sun, Likui Zhang, Jusheng Zhang, Qinhan Lyu, Tianshui Chen, Liang Lin, Guangrun Wang·2026-10-28 04:00 UTC· 2 hours agoSignalEditorial score85

E0.5: From Reconstructive to Future-Predictive 3D Foundations for Robotic World-Action Modeling

Summary

E0.5 is a geometry-native world-action model built on 3D foundation models, predicting future 3D states and actions. It achieves 89.90% success in clean settings and 50.09% in randomized settings without embodied-data pretraining. Further pretraining on 160k robot trajectories improves real-world manipulation success by 15 and 7 percentage points on AgileX Piper and Zerith H1.

Editorial context

The paper introduces E0.5, a 3D-native world-action model that predicts future geometric states alongside actions, showing improved success rates in real-world manipulation tasks after pretraining on robot trajectories.

Source: arXiv Robotics — research abstracts · Read original article ↗

Loading article text…

Source：arXiv Robotics — research abstracts · arxiv.org

#Research #World models #Sim-to-real #Robot policies #AGIBOT