Tom Goldstein – Evasion and poisoning attacks on neural networks: theoretical and practical perspectives
Abstract: In this talk I will give an overview of adversarial attacks and dataset poisoning attacks on neural networks. Using empirical studies, I will show examples where these attacks can pose a real threat to real-world systems, such as copyright detection system, financial markets, and Google’s AutoML API. Then, I’ll dive into the theory of adversarial attacks, and present situations where such attacks cannot be avoided.
Bio: Thomas Goldstein obtained his PhD in Mathematics at UCLA, and was a research scientist at Rice University and Stanford University. He has been the recipient of several awards, including SIAM’s DiPrima Prize, a DARPA Young Faculty Award, and a Sloan Fellowship. His research lies at the intersection of machine learning and optimization, and targets applications in computer vision and signal processing. Dr, Goldstein works at the boundary between theory and practice, leveraging mathematical foundations, complex models, and efficient hardware to build practical, high-performance systems. He designs optimization methods for a wide range of platforms ranging from powerful cluster/cloud computing environments to resource limited integrated circuits and FPGAs.
Please reach out to Meg Tully ([email protected]) for zoom meeting information