Results 961 - 970 of 24441
Simons Institute Director Venkat Guruswami interviews Dan Spielman (Yale) and Nikhil Srivastava (Simons Institute), two of the organizers of First Proof, a project that aims to measure the capabilities of AI systems in the context of research mathematics.
While differential privacy is the gold standard in centralized privacy-preserving machine learning, it is not well suited to decentralized learning where participants collaboratively train a model via peer-to-peer messages. We present a relaxation of differential privacy that captures a relevant trust setting for decentralized learning and show how decentralization can lead to privacy amplification mechanisms. We illustrate these findings on graphs from the Fediverse (alternative social media networks).