Wen Zhang

Huazhong University of Science and Technology

Papers

Total Citations

331

H-Index

About

Wen Zhang is a prominent researcher specializing in **transfer learning**, machine learning, and knowledge adaptation, with a particular focus on understanding and mitigating the challenges that arise when transferring knowledge across domains. His most influential contribution, "A Survey on Negative Transfer" (2022), has rapidly accumulated over 331 citations, underscoring its significance to the research community. This work addresses one of the most critical yet underexplored problems in transfer learning — the phenomenon of *negative transfer*, where leveraging source domain knowledge actually *hurts* rather than helps performance in the target domain. By systematically surveying this issue, Zhang provided researchers with a foundational framework for diagnosing, understanding, and overcoming this subtle but consequential failure mode. His research is especially relevant in real-world scenarios where labeled data is scarce due to cost, privacy, or practical constraints — settings increasingly common in healthcare, robotics, and natural language processing. Zhang's work has shaped how the community approaches the reliability and safety of transfer learning systems, making him a key voice in building more robust and trustworthy machine learning pipelines.