Hossein Yalame
<<Invited Talk>>
〇 Date: 2025.5.23 (Fri.)
〇 Speaker: Hossein Yalame (Robert Bosch GmbH)
〇 Title: MPC-based PPML: Advancing Privacy-Preserving Machine Learning
〇 Abstract
This talk will explore how secure Multi-Party Computation (MPC) can enable privacy-preserving machine learning (PPML) workflows. We will begin with a brief introduction to MPC, followed by the transition from traditional MPC to its application in PPML. We will then delve into secure inference and training using MPC, highlighting current challenges and approaches. The talk will conclude with insights into integrating these methods with federated learning. The content follows a practical roadmap, starting from foundational principles and moving through real-world applications and emerging research directions in privacy-enhancing technologies for AI.
