Privacy-Preserving Decentralized Memory for Federated Learning - Secure MPC

Learn about privacy-preserving decentralized memory systems for federated learning through this 31-minute conference talk by Jared Saia from the University of New Mexico, presented at the Simons Institute. Explore the intersection of secure multi-party computation (MPC) and federated learning, focusing on how decentralized memory architectures can maintain privacy while enabling collaborative machine learning across distributed networks. Discover the technical challenges and solutions involved in managing memory systems that protect sensitive data during federated learning processes, while understanding the cryptographic foundations that make secure computation possible in distributed environments.