Jingkun An | 安璟坤

I am an upcoming Ph.D. student in Computer Science at the University of California, Davis (UC Davis), supervised by Prof. Junshan Zhang and Prof. Lifu Huang. Previously, I received my Bachelor's degree from Beihang University in 2024. During my two internships at Peking University, I was honored to be advised by Dr. Yemin Shi and Prof. Shanghang Zhang, respectively.

Email / Google Scholar / GitHub / CV

News

[2026/08]🎉 SafeSteer gets accepted to EMNLP 2026 (Main)! See you in Budapest!

[2026/06]🎉 RoboTracer gets accepted to ECCV 2026! See you in Malmö!

[2026/06]🎉 Released SafeSteer, a lightweight framework for safety alignment via localized on-policy distillation — 100 harmful samples, minial alignment tax! Code available at GitHub.

[2025/09]🎉 RefSpatial-Bench is used by Qwen3-VL and Gemini-Robotics-1.5 to evaluate spatial reasoning in complex scenes!

[2025/09]🎉 RoboRefer gets accepted to NeurIPS 2025! See you in San Diego!

[2024/12]🎉 AGFSync gets accepted to AAAI 2025! See you in Philadelphia!

[2024/12]🎉 Medical MLLM is Vulnerable gets accepted to AAAI 2025! See you in Philadelphia!

[2023/10]🎉 Led the team 'TrustMe' to win the Grand Prize (Top 1) in the 2023 'Challenge Cup' National Competition!

[2023/05]🎉 M3Fair Win the third prize (Top 5) in 2023 NIH Bias Detection!

Publications ( *, †, ‡ indicates the equal contributions, corresponding author, project leader, respectively.)

Currently, my interest lies in Embodied Agents, which are at the intersection of Multimodal Large Language Models and Embodied AI, with particular interests in high-level planning and low-level control with spatio-temporal intelligence, working towards an generalist agent in a complex real-world environment. Representative works are highlighted.

	SafeSteer: Localized On-Policy Distillation for Efficient Safety Alignment Hao Li , Jingkun An , Zijun Song , Pengyu Zhu, Rui Li, Hao Wang, Wendi Feng, Yesheng Liu, Lijun Li, Jin-Ge Yao, Lei Sha^† Paper / Project / Code / Copy BibTeX Copy Success! TL;DR: Localized Distillation, Minimal Alignment Tax!* EMNLP 2026 (Main)
	RoboTracer: Mastering Spatial Trace with Reasoning in Vision-Language Models for Robotics Enshen Zhou , Yibo Li , Jingkun An , Jiayuan Zhang, Shanyu Rong, Mengzhen Liu, Yi Han, Yuheng Ji, Huajie Tan, Jiawei He, Pengwei Wang, Zhongyuan Wang, Cheng Chi^†, Lu Sheng^†, Shanghang Zhang^† Paper / Project / Code / Copy BibTeX Copy Success! TL;DR: From what you say to where it moves using RoboTracer!* ECCV 2026
	TIGeR: Tool-Integrated Geometric Reasoning in Vision-Language Models for Robotics Yi Han , Cheng Chi ^‡, Enshen Zhou , Shanyu Rong, Jingkun An, Pengwei Wang, Zhongyuan Wang, Lu Sheng^†, Shanghang Zhang^† Paper / Project / Code / Copy BibTeX Copy Success! TL;DR: Equipping VLMs to perform accurate geometric reasoning for robotics!* ICRA 2026
	RoboRefer: Towards Spatial Referring with Reasoning in Vision-Language Models for Robotics Enshen Zhou , Jingkun An , Cheng Chi ^‡, Yi Han, Shanyu Rong, Chi Zhang, Pengwei Wang, Zhongyuan Wang, Tiejun Huang, Lu Sheng^†, Shanghang Zhang^† Paper / Project / Code / Copy BibTeX Copy Success! TL;DR: From words to exactly where you mean using RoboRefer!* NeurIPS 2025
	Medical MLLM is Vulnerable: Cross-Modality Jailbreak and Mismatched Attacks on Medical Multimodal Large Language Models Xijie Huang , Xinyuan Wang , Haotao Zhang , Yinghao Zhu , Jiawen Xi, Jingkun An, Hao Wang, Hao Liang, Chengwei Pan^† Paper / Project / Code / Copy BibTeX Copy Success! TL;DR: Medical MLLM is Vulnerable! AAAI 2025
	AGFSync: Leveraging AI-Generated Feedback for Preference Optimization in Text-to-Image Generation Jingkun An , Yinghao Zhu , Zongjian Li, Enshen Zhou, Haoran Feng, Xijie Huang, Bohua Chen, Yemin Shi, Chengwei Pan^† Paper / Project / Code / Copy BibTeX Copy Success! TL;DR: Train T2I Diffusion model with AI-Generated Feedback for DPO!* AAAI 2025
	M3Fair: Mitigating Bias in Healthcare Data through Multi-Level and Multi-Sensitive Attribute Reweighting Method Yinghao Zhu , Jingkun An , Enshen Zhou, Hao Li, Haoran Feng, Paper / Project / Code / Copy BibTeX Copy Success! TL;DR: BeFair (A Bias Detection and Mitigation Tool)! 2023 NIH Bias Detection Third Prize (Top 5)

Experience

	School of Computer Science at Peking University 2025.02 - 2025.11 Research Intern Research Advisor: Dr. Lu Sheng, Prof. Shanghang Zhang
	School of Electronics Engineering and Computer Science at Peking University 2023.12 - 2024.08 Research Intern Research Advisor: Dr. Yemin Shi, Prof. Li Yuan

Education

Beihang University (BUAA)
2020.09 - 2024.06
Bachelor of Software Engineering
GPA ranking: 27/187

Selected Awards and Honors

2024: Outstanding Graduate of Beihang University.

2023: Grand Prize (Top 1) in "Challenge Cup" Competition of Science Achievement in China.