2025
Convergence of Clipped SGD on Convex $(L_0,L_1)$-Smooth Functions
NeurIPS 2025poster
We study stochastic gradient descent (SGD) with gradient clipping on convex functions under a generalized smoothness assumption called $(L_0,L_1)$-smoothness. Using gradient clipping, we establish a high probability convergence rate that matches the SGD rate in the $L$ smooth case up to polylogarith…