About me
Mengqi Zhang 张孟奇
Assistant Professor · M.Sc. Supervisor · General Artificial Intelligence Laboratory
School Computer Science and Technology, Shandong University
Ph.D. CASIA, 2023
I am Mengqi Zhang (张孟奇), an Assistant Professor and M.Sc. Supervisor in the School of Computer Science and Technology, Shandong University, and a member of the General Artificial Intelligence Laboratory. I received my Ph.D. degree in 2023 from the Institute of Automation, Chinese Academy of Sciences (CASIA), where I was affiliated with the State Key Laboratory of Multimodal Artificial Intelligence Systems, under the supervision of Prof. Liang Wang (NSFC Distinguished Young Scholar, IEEE Fellow), and was co-advised by Prof. Shu Wu and Prof. Qiang Liu.
Research Interests
My current research centers on trustworthy and controllable large language models (LLMs), with particular interests in LLM agents, knowledge editing, agentic retrieval-augmented generation (RAG), interpretability, and model safety. Previously, I worked extensively on temporal knowledge graph reasoning and recommender systems.
I am always looking for self-motivated students interested in research on LLMs. Feel free to reach out if you are interested in pursuing a Master's degree, a Ph.D., an undergraduate research assistant opportunity, or potential collaborations.
欢迎对大模型研究感兴趣的同学（包括保研、考研同学和本科生科研助理）通过邮件与我联系，也欢迎任何形式的学术探讨与合作！
Preprints
On the Hidden Costs of Counterfactual Knowledge Training in LLM Unlearning.
Uncovering Entity Identity Confusion in Multimodal Knowledge Editing.
Uncovering Context Reliance in Unstructured Knowledge Editing.
Open Problems and a Hypothetical Path Forward in LLM Knowledge Paradigms.
Selected Publications Full list: Google Scholar or Publications
* equal contribution # corresponding author
Disentangling Knowledge Representations for Large Language Model Editing.
Spectral Characterization and Mitigation of Sequential Knowledge Editing Collapse.
LLM Unlearning Should Be Form-Independent.
Uncovering Overfitting in Large Language Model Editing.
KELE: Residual Knowledge Erasure for Enhanced Multi-hop Reasoning in Knowledge Editing.
UIPE: Enhancing LLM Unlearning by Removing Knowledge Related to Forgetting Targets.
ExcluIR: Exclusionary Neural Information Retrieval.
Knowledge Graph Enhanced Large Language Model Editing.
Generative Retrieval as Multi-Vector Dense Retrieval.
Learning Long- and Short-term Representations for Temporal Knowledge Graph Reasoning.
Learning Latent Relations for Temporal Knowledge Graph Reasoning.
Dynamic Graph Neural Networks for Sequential Recommendation.
Deep Contrastive Multiview Network Embedding.
Personalized Graph Neural Networks with Attention Mechanism for Session-Aware Recommendation.
Dynamic Graph Collaborative Filtering.
Honors and Awards
Shandong Provincial Young Science and Technology Talent Support Program 山东省青年科技人才托举工程（A类）
SIGIR Best Paper Honorable Mention 最佳论文提名奖
CAS Presidential Scholarship 中国科学院院长奖
Beijing Outstanding Graduate Student Award 北京市优秀毕业生
University of Chinese Academy of Sciences Outstanding Graduate Award 中国科学院大学优秀毕业生
Experience
School of Computer Science and Technology, Shandong University
2023.07 -- Present Assistant Professor
School of Artificial Intelligence, University of Chinese Academy of Sciences
Institute of Automation, Chinese Academy of Sciences
2019.09 -- 2023.06 Ph.D. in Computer Application Technology
Advisor: Professor Liang Wang
School of Mathematics and Systems Science, Beihang University
2016.09 -- 2019.01 M.Sc. in Mathematics
Advisor: Professor Xin Jiang
School of Mathematics and Computational Science, Xiangtan University
2012.09 -- 2016.06 B.Sc. in Mathematics
Advisor: Professor Huayi Wei
Professional Services
ACL Rolling Review
NeurIPS, ICLR, ICML, KDD, SIGIR, AAAI, ACL, EMNLP, etc.
IEEE TKDE, TOIS, Pattern Recognition, IEEE TBD, Neural Networks, etc.