Zifeng Ding

Zifeng Ding is a Research Fellow at the Department of Computer Science and Technology, University of Cambridge, affiliated with the School of Technology. His research focuses on artificial intelligence, machine learning, knowledge graphs, and their applications in natural language processing and computer vision. Key interests include temporal knowledge graph reasoning, multimodal learning, and the reliability of large language models (LLMs).

He has contributed to benchmark development (e.g., TCP, FOReCAst), dataset creation (AVerImaTeC), and methods addressing LLM limitations (e.g., hallucination mitigation, bias analysis). His work bridges theoretical advancements in graph neural networks with practical applications in real-world systems like supply chain analysis and evidence-based fact verification. Current projects explore parameter-efficient model tuning (Perft) and cross-lingual knowledge editing via in-context learning.

Publications span top venues, emphasizing temporal dynamics, causal reasoning, and trustworthiness in AI systems. He collaborates widely on topics ranging from dynamic graph modeling (DyGMamba) to future outcome prediction (FOReCAst benchmark).