Vasisht Duddu
Ph.D. Student, Computer Science
Secure Systems Group
Cheriton School of Computer Science
CrySP Lab, University of Waterloo
vasisht.duddu at uwaterloo.ca
I am pursuing my Ph.D. in Computer Science at the University of Waterloo advised by N. Asokan. My research is supported by David R. Cheriton Graduate Scholarship (2024).
Research: I study different risks to security, privacy, fairness, and transparency in machine learning models. I design attacks to exploit these risks, defenses to counter them, and study the interplay between risks and defenses. Additionally, I work on ensuring accountability in machine learning pipelines to meet regulatory requirements.
Past: I completed my masters from UWaterloo advised by N. Asokan. My thesis was on the effectiveness of Shapley Values to quantify membership privacy risk. Before UWaterloo, I worked with Antoine Boutet (INRIA's Privatics Lab) on quantifying data privacy risks in machine learning. I completed my undergraduate from IIIT-Delhi, India where I received Dean's Award for Innovation, Research and Development for Fault Tolerant Neural Networks in benign and adversarial settings (advised by Valentina Balas). I have worked with Reza Shokri (National University of Singapore) on confidential machine learning, and Debasis Samanta (IIT Kharagpur) on timing side channels in neural networks.