publications

2026

1. 

   Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success Rate but 65% Fewer Tokens

   Ruiyang Si, Jianxin Bi, Shunyu Yang, Rui Ni, Wenbo Huang, Qiang Wang, Shulong Jiang, Yintianrun Zhang, Feixiang Ruan, Kewei Zhang, Duomin Wang, Xiuyu Li, Haiwen Feng, Zhen Dong, and Daquan Zhou

   Preprint, 2026

   Abs arXiv Code Website

   PyRUA-Lean lets robot agents compose action primitives into Python cells with local checks and retries, reducing repeated model calls and unnecessary visual feedback while improving task success.

2. 

   Dex-X: Learning Visual-Tactile Dexterous Manipulation From Human Videos with Simulated Interaction

   Ruoqu Chen, Feixiang Ruan, Liu Cao, Zihao Wang, Botian Xu, Shiqin Tong, Jiajun Liu, Mingzhi Pei, Chenyu Zhang, Wanli Xing, Kaifeng Zhang, and Mengdi Xu

   Conference on Robot Learning (CoRL), 2026

   RSS 2026 Workshop on Dexterous Manipulation (Spotlight), 2026

   Abs arXiv Website

   A framework that learns visual-tactile dexterous manipulation skills from human videos by leveraging simulated interaction, bridging the gap between passive human demonstrations and executable robot policies.

3. 

   R2S-Eval: Robot Evaluation with Real-to-Sim Calibration via Vision-Language Models

   Yidi Wang^*, Feixiang Ruan^*, Ruoqu Chen, Jie Yin, Yang Yu, Mengdi Xu, and Kaifeng Zhang

   Preprint, 2026

   Abs arXiv Website

   R2S-Eval combines real-to-sim calibration with vision-language model preference evaluation to produce reliable, stable, and quality-aware rankings of robot policies.