Cheng Han

Cheng Han is a tenure-track Assistant Professor in the School of Science and Engineering at the University of Missouri -- Kansas City (UMKC), where he conducts research in adaptable and sustainable intelligence, focusing on efficient AI systems and parameter-efficient fine-tuning methods for large-scale models.

• Ph.D., Rochester Institute of Technology (RIT)
• M.S., Pennsylvania State University (PSU)
• B.S., Tianjin University (TJU)

His research interests center on creating energy-wise AI systems that empower communities and address environmental and social challenges. He focuses on multimodal and visual prompt tuning, transfer learning, and robust AI. His work bridges theoretical innovation with real-world deployment, particularly in efficient adaptation of vision and language models.

His recent publications span top venues like NeurIPS, ICCV, CVPR, ICLR, EMNLP, and IEEE TPAMI. The research trends highlight a strong focus on parameter efficiency, prompt engineering, model robustness, and multimodal understanding. He investigates when and why prompt tuning outperforms full fine-tuning and develops novel frameworks like E^2VPT and M^2PT for efficient adaptation.

Cheng Han actively contributes to the academic community as a reviewer and committee member.

• Program Committee, AAAI (2023–present)
• Program Committee, SIAM SDM (2024)
• Reviewer for NeurIPS, ICLR, CVPR, ICML, ICCV, TPAMI, TMLR, and others

He advises Ph.D. students and teaches courses such as Deep Learning (COMP-SCI 5567). He has given invited talks at ICLR, ICCV, and seminars at NSF and Naval Research Laboratory. His research is supported by academic collaborations and likely grant funding, given his active publication and service profile. He leads a research group focused on sustainable and efficient AI, with code available on GitHub.