Yuxuan Chen (Shandong University)
Yuxuan Chen is now an associate professor at Department of cyber security, Shandong University, China. He obtained his PhD from Florida Tech where he was advised by Shengzhi Zhang. His research is focused on adversarial machine learning, especially audio adversarial attacks and defenses.
Short Abstract: Although deep-learning based Automatic Speech Recognition (ASR) systems have achieved remarkable success, they are vulnerable to audio adversarial examples that are generated by adding small perturbations on benign audio samples. The existence of audio adversarial examples raises security concerns in many real-world applications. Thus, it has raised increasing attention in recent years. Compared with the adversarial attack on image models, attacking ASR models is more challenging, due to the complexity of audio signals and ASR models. In this talk, I will share the recent research progress of adversarial attack on ASR models, including white-box audio adversarial attacks and transferable audio adversarial attacks towards black-box commercial ASR models.