|
Hongji Chen 「陈鸿基」 I am an undergraduate student at Fudan University (Class of 2029). I have been a research intern at Fudan NLP Lab since October 2025. I am currently in NLP-LI Lab, advised by Assoc. Prof. Tao Gui and Prof. Qi Zhang. I am also a research intern at Shanghai AI Lab (October 2026 – present). I received a Ph.D. offer from the Shanghai Innovation Institute in Summer 2026, becoming the youngest SII Ph.D. offer recipient up to that time. My research spans Digital Agents and Physical Agents, with a primary focus on Computer Use Agents and Agentic RL. I am also interested in Vision-Language-Action (VLA) models and Embodied Agents. Please feel free to contact me if you are interested in collaboration. I do not pursue research in Multimodal Generation, World Models, or Unified Multimodal Models. In my view, multimodal generation is unrelated to intelligence. Mail / Google Scholar / Github / Twitter / LinkedIn |
|
Publications |
|
AI Can Learn Scientific Taste
Jingqi Tong, Mingzhe Li, Hangcheng Li, Yongzhuo Yang, Yurong Mou, Weijie Ma, Zhiheng Xi, Hongji Chen, Xiaoran Liu, Qinyuan Cheng, Ming Zhang, Qiguang Chen, Weifeng Ge, Qipeng Guo, Tianlei Ying, Tianxiang Sun, Yining Zheng, Xinchi Chen, Jun Zhao, Ning Ding, Xuanjing Huang, Yugang Jiang, Xipeng Qiu arXiv, 2026 project page / arXiv / github / models & datasets / demo We propose Reinforcement Learning from Community Feedback (RLCF), training Scientific Judge to model research taste from large-scale community signals and using it to improve Scientific Thinker for high-potential scientific ideation. |
Selected Projects |
|
Shanghai Innovation Institute Research Cockpit
Major participant, Shanghai Innovation Institute Participated in building a full-process agentic research operating system for organized research. The system upgrades AI from stage-specific assistance to automated literature search, idea generation and evaluation, research-plan generation from goals, execution tracking, iterative verification, and lifecycle management. InspireSkill is a concrete outcome of this project, providing reusable skills and tool interfaces for agents to operate the Inspire compute platform through reproducible workflows. |
Research Experiences |
|
|
|
Shanghai AI Lab Oct. 2026 - Present Research Intern |
|
Shanghai Innovation Institute Apr. 2026 - Aug. 2026 Research Intern, advised by Prof. Xipeng Qiu |
|
|