I am Zongji Yu. I received my B.Eng. in Robotics Engineering from the School of Control Science and Engineering, Shandong University (SDU), where my graduation project was recognized as an Outstanding Undergraduate Graduation Project of Shandong University.

I am currently a Ph.D. student at the School of Automation and Perception, Shanghai Jiao Tong University (SJTU), advised by Professor Chen Gong in the LEAP Group. My research focuses on Continual Learning and Evolving AI.

I am also a research intern at Shanghai AI Laboratory, under the mentorship of Young Scientist Chaochao Lu. My work there focuses on Intrinsic AI Safety: developing AI systems with safety capabilities that arise from within their learning and reasoning processes.

My broader research goal is to build learning systems that adapt continuously, evolve responsibly, and remain safe in open-ended environments.

🔥 News

  • 2025.12:  🎉 I began a research internship at Shanghai AI Laboratory, under the mentorship of Young Scientist Chaochao Lu, working on Intrinsic AI Safety.
  • 2025.11:  🎉 I became a candidate for the President’s Scholarship of Shandong University (the highest honor of SDU)!
  • 2025.10:  🎉 I received the BYD Scholarship of 10,000 CNY (the only one in the entire college)!
  • 2025.09:  🎉 I successfully obtained a recommendation for admission to the School of Automation and Perception, Shanghai Jiao Tong University to pursue Ph.D, and got 191(ranked 2/81) among all directly admitted doctoral students.
  • 2025.09:  🎉 I have been invited to serve as a reviewer for TCSVT (Trans on Circuits and Systems for Video Technology, IF=11.015).
  • 2025.07:  🎉 A paper about Mamba and underwater tasks was accepted by ACM MM 2025!
  • 2025.06:  🎉 I have been invited to serve as a reviewer for the IEEE Journal of Oceanic Engineering (IF=3.883).
  • 2025.05:  🎉 A paper about MPC and optimistic was accepted by ICITS 2025!

📝 Publications

TPAMI 2026
UIS-Mamba++ framework

UIS-Mamba++: Iterative Graph Scan and Hidden State Modulation for Robust Underwater Instance Segmentation

Zongji Yu, Hao Fang, Zhiyang Chen, Sam Kwong, Runmin Cong

  • Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). Under Review
  • Extends UIS-Mamba with Iterative Graph Scan for stage-wise dynamic graph updates and graph-driven Hidden State Modulation for fine-grained modeling of instance interiors, boundaries, and backgrounds.
arXiv 2026
MCPO framework

Harmony in Diversity: Multi-domain Contrastive Policy Optimization for Large Reasoning Models

Zongji Yu, Wenshui Luo, Yiliu Sun, Hao Fang, Runmin Cong, Chaochao Lu, Chen Gong

  • arXiv preprint, 2026. Preprint
  • Introduces Multi-domain Contrastive Policy Optimization (MCPO) to promote cross-domain knowledge sharing and in-domain knowledge consolidation for large reasoning models.
TKDE 2026
UniGFM-RAG overall framework

UniGFM-RAG: Unify Graph Foundation Models and Large Language Models for End-to-End Retrieval and Reasoning

Zicheng Zhao, Linhao Luo, Zongji Yu, Junnan Dong, Yanchao Tan, Shirui Pan, Chen Gong

  • Submitted to IEEE Transactions on Knowledge and Data Engineering (TKDE). Under Review
  • Unifies graph retrieval and language generation through a fully differentiable GraphRAG framework, coupling a Graph Foundation Model retriever with a Large Language Model generator.
ACM MM 2025
sym

UIS-Mamba: Exploring Mamba for Underwater Instance Segmentation via Dynamic Tree Scan and Hidden State Weaken

Runmin Cong, Zongji Yu, Hao Fang, Haoyan Sun, Sam Kwong

  • Accepted at the 33rd ACM International Conference on Multimedia Poster
  • The first underwater instance segmentation model based on visual Mamba, designed with dynamic tree scan and background hidden state weaken modules.

🎖 Honors and Awards

  • 2025.11:  💰 the President’s Scholarship of Shandong University - Candidate
  • 2025.10:  💰 BYD Scholarship – 10,000 CNY
  • 2025.05:  🏆 2025 Mathematical Contest in Modeling (MCM) – International First Prize (M Award)
  • 2024.11:  💰 Inspur Scholarship – 5,000 CNY
  • 2024.04:  🏆 2024 Mathematical Contest in Modeling (MCM) – International First Prize (M Award)
  • 2023.12:  🏆 2023 Asia and Pacific Mathematical Contest in Modeling (APMCM) – International First Prize
  • 2023.11:  💰 Shandong Provincial Government Scholarship – 7,000 CNY

📖 Education

  • 2022.09 – 2026.06:  🇨🇳 School of Control Science and Engineering, Shandong University (SDU)

    - Bachelor of Engineering in Robotics Engineering; Outstanding Undergraduate Graduation Project of Shandong University

  • 2025.08 – Present:  🇨🇳 School of Automation and Perception, Shanghai Jiao Tong University (SJTU)

    - Ph.D. in Control Science and Engineering; Advisor: Professor Chen Gong; LEAP Group

💻 Professional Services

  • 2025, Reviewer for the Trans on Circuits and Systems for Video Technology(TCSVT, IF=11.015).
  • 2025, Reviewer for the IEEE Journal of Oceanic Engineering(JOE, IF=3.883).

🌍 Visitor Map