My research mainly focuses on the safety of large vision-language models, specifically multi-backdoor attacks on the image modality.
Fenglin Zhu is a graduate student at the SHEIN-MING WU School of Intelligent Engineering, South China University of Technology. He is getting ready to start his third year of graduate school and is now pursuing further research in the safety of large vision-language models, specifically multi-backdoor attacks on the image modality.