Selected Publications

	Video-R1: Reinforcing Video Reasoning in MLLMs arXiv 2025 Kaituo Feng, Kaixiong Gong, Bohao Li, Zonghao Guo, Yibing Wang, Tianshuo Peng, Junfei Wu, Xiaoying Zhang, Benyou Wang, Xiangyu Yue Explore the R1 paradigm for eliciting video reasoning within MLLMs. Paper Code
	Reinforcing Spatial Reasoning in Vision-Language Models with Interwoven Thinking and Visual Drawing arXiv 2025 Junfei Wu, Jian Guan, Kaituo Feng, Qiang Liu, Shu Wu, Liang Wang, Wei Wu, Tieniu Tan Achieveing o3-like thinking for spatial reasoning across images and videos. Paper Code
	SophiaVL-R1: Reinforcing MLLMs Reasoning with Thinking Reward arXiv 2025 Kaixuan Fan*, Kaituo Feng*, Haoming Lyu, Dongzhan Zhou, Xiangyu Yue (*equal contribution) Intergrating thinking-level reward to address the phenomenon of "wrong thinking, correct answer". Paper Code
	Critique-GRPO: Advancing LLM Reasoning with Natural Language and Numerical Feedback arXiv 2025 Xiaoying Zhang, Hao Sun, Yipeng Zhang, Kaituo Feng, Chaochao Lu, Chao Yang, Helen Meng Using external critiques as language feedback for improving reasoning Paper Code
	AV-Odyssey: Can Your Multimodal LLMs Really Understand Audio-Visual Information? arXiv 2024 Kaixiong Gong*, Kaituo Feng*, Bohao Li*, Yibing Wang, Mofan Cheng, Shijia Yang, Jiaming Han, Benyou Wang, Yutong Bai, Zhuoran Yang, Xiangyu Yue (*equal contribution) We propose a comprehensive benchmark for evaluating audio-visual understanding abilities of MLLMs. Paper Code
	On the Road to Portability: Compressing End-to-End Motion Planner for Autonomous Driving CVPR 2024 Kaituo Feng, Changsheng Li, Dongchun Ren, Ye Yuan, Guoren Wang We constitute the first attempt to explore a knowledge distillation method to compress end-to-end autonomous driving planners. Paper Code
	Keypoint-based Progressive Chain-of-Thought Distillation for LLMs ICML 2024 Kaituo Feng, Changsheng Li, Xiaolu Zhang, Jun Zhou, Ye Yuan, Guoren Wang We propose a new compression method to progressively distill the emergent reasoning capabilities of LLMs into smaller models, as well as encouraging the precise mimicry of significant tokens. Paper
	Towards Open Temporal Graph Neural Networks ICLR 2023, Oral, 90/4922 Kaituo Feng, Changsheng Li, Xiaolu Zhang, Jun Zhou We propose the first class-incremental learning for temporal GNNs, allowing temporal graphs to evolve in the real-world scenarios with an open class set Paper Code
	Shared Growth of Graph Neural Networks via Prompted Free-Direction Knowledge Distillation IEEE TPAMI, KDD 2022 Kaituo Feng, Yikun Miao, Changsheng Li, Ye Yuan, Guoren Wang We utilize reinforcement learning to exchange beneficial knowledge between two GNNs Paper

Selected Honors and Awards

• National Scholarship, Ministry of Education of China (TOP 2%), 2024.

• National Scholarship, Ministry of Education of China (TOP 2%), 2023.

• Outstanding Undergraduate Student of Beijing Institute of Technology, 2022.

• Silver Medal of 45th ACM-ICPC Asia Regional Contest, 2020.

• First Prize (top 1%) of China Undergraduate Mathematical Contest in Modeling (CUMCM), 2020.

• Gold Medal of Group Programming Ladder Tournament China Finals, 2020.