TwinAligner: Visual-Dynamic Alignment Empowers Physics-aware Real2Sim2Real for Robotic Manipulation

作者：Hongwei Fan, Hang Dai, Jiyao Zhang, Jinzhou Li, Qiyang Yan, Yujie Zhao, Mingju Gao, Jinghang Wu, Hao Tang, Hao Dong · 单位：Science, Peking University, PKU-AgiBot Lab, State Key Laboratory of, Multimedia Information Processing, School of Computer Science, Peking · 会议/期刊：arXiv · 日期：2025-12-22 · 来源：Bottlenecks / Generalization / Sim2Real or Real2Sim2Real Generalization