Peter RichtarikThe First Optimal Parallel SGD (in the Presence of Data, Compute and Communication Heterogeneity)
Watch keynotePDFFrancesco OrabonaUnderstanding Adam and AdamW through Proximal Updates, Scale-Freeness, and Relaxed Smoothness
Watch keynotePDFMartin TakacElevating Federated Learning: A Shift Towards Second-Order Optimization Methods
Watch keynotePDFAlexander BeznosikovNavigating Byzantine Attacks in Distributed Learning: a Trial Function MethodologyWatch keynotePDFAlexey NaumovStochastic approximation: old and new
Watch keynotePDFAndrey KuznetsovMultimodal architectures and context compression methods (online)
Watch keynotePDFBoris MordukhovichConvergence of Descent Methods under Polyak-Kurdyka-Lojasiewicz Properties (online)
Watch keynotePDFVladimir SpokoinyInference for nonlinear inverse problems
Watch keynotePDFGesualdo ScutariBringing Statistical Thinking in Decentralized Optimization. Vignettes from High-dimensional Statistics over Networks (online)
Watch keynote