|
I am Bofang Jia, an incoming research-based master's(MEng) student at Nanyang Technological University, advised by Prof. Ziwei Wang. I received my Bachelor of Engineering degree from Southwest University in June 2025. During my undergraduate studies, I was also a visiting student at Westlake University, advised by Prof. Donglin Wang. Email / Google Scholar / GitHub / Twitter / LinkedIn |
|
|
|
My research interests are primarily focused on Embodied Artificial Intelligence and Computer Vision. My research interests focus on how to achieve broad generalization capabilities for open-world manipulation tasks, the application of Vision-Language-Action models to robots, and inference acceleration. |
|
Can Cui, Pengxiang Ding*, Wenxuan Song, Shuanghao Bai, Xinyang Tong, Zirui Ge, Runze Suo, Wanqi Zhou, Yang Liu, Bofang Jia, Hangyu Liu, Mingyang Sun, Han Zhao, Siteng Huang, Donglin Wang arXiv, 2025 [Project Page] [Paper] [Code] OpenHelix is a dual-system framework for robotics(VLA Model), which leverages the reasoning capabilities of high-level multimodal large language models (MLLMs) to enhance low-level policies with minimal training cost. This work first summarizes and compares the structural designs of existing dual-system architectures, and then conducts systematic empirical evaluations on their core design elements. Finally, we present a low-cost open-source model to support further research and applications. |
|
Bofang Jia, Pengxiang Ding, Can Cui, Mingyang Sun, Pengfang Qian, Siteng Huang, Zhaoxin Fan, Donglin Wang arXiv, 2024 [Project Page] [Paper] [Code] SDM Policy distills diffusion-based visuomotor policies into one-step generators via a two-stage process combining score and distribution matching. It uses a dual-teacher setup to enhance robustness and achieves 6 times faster inference. |
|
Guochuan Xian, Siyuan Du, Xi Tang, Yuan Shi, Bofang Jia, Banghao Tang, Zhefu Leng, Li Li Neural Computing and Applications (NCAA), 2024 [Project Page] [Paper] [Code] DLEE is a high-quality Chinese dataset for document-level legal event extraction, enabling training and evaluation of NLP models in complex legal reasoning contexts. |
|
|
I have some experience in designing and implementing robotic systems. Here are some of the robotic systems I have helped develop: [Link] |
|
|