[Paper Summary] "BadPre: Task-agnostic Backdoor Attacks to Pre-trained NLP Foundation Models"
top of page
Home: Welcome
Jaechul (Harry) Roh
[CV] | [Google Scholar] | [Github]
I am a Computer Science Ph.D. student at the University of Massachusetts Amherst under the supervision of Prof. Amir Houmansadr. I have recently graduated B.Eng in Computer Engineering at Hong Kong University of Science and Technology (HKUST).
My research interests rely on Trustworthy AI and Adversarial ML. Especially, I am fascinated by both adversarial attack as well as adversarial training that aids to overcome the vulnerability of various machine learning models in wide range of domains. I am also interested in exploring other fields of study such as the relationship between adversarial attack and federated learning, backdoor attacks/defense, and robust optimization.
Publications
-
Robust Smart Home Face Recognition under Starving Federated Data
Jaechul Roh, Yajun Fang
2022 6th International Conference on Universal Village (IEEE UV2022)
2022/09/26
[paper] [code] [slides] [video]
Jaechul Roh, Minhao Cheng, Yajun Fang
2022 6th International Conference on Universal Village (IEEE UV2022)
2022/09/26
[paper] [code] [slides] [video]
-
Impact of Adversarial Training on the Robustness of Deep Neural Networks
Jaechul Roh
2022 IEEE 5th International Conference on Information Systems and Computer
Aided Education (ICISCAE)
2022/04/15
[paper] [code]
bottom of page