TheoreticallyBetterandNumericallyFaster DistributedOptimizationwith Smoothness-AwareQuantizationTechniques
–Neural Information Processing Systems
This issue is further exacerbated by the fact that modern highly performing models are typically overparameterized[Brownetal.,2020,Narayananetal.,2021].
Neural Information Processing Systems
Feb-8-2026, 13:05:26 GMT