Accelerated stochastic first-order method for convex optimization under heavy-tailed noise
By: Chuan He, Zhaosong Lu
Potential Business Impact:
Makes computer learning faster with messy data.
We study convex composite optimization problems, where the objective function is given by the sum of a prox-friendly function and a convex function whose subgradients are estimated under heavy-tailed noise. Existing work often employs gradient clipping or normalization techniques in stochastic first-order methods to address heavy-tailed noise. In this paper, we demonstrate that a vanilla stochastic algorithm -- without additional modifications such as clipping or normalization -- can achieve optimal complexity for these problems. In particular, we establish that an accelerated stochastic proximal subgradient method achieves a first-order oracle complexity that is universally optimal for smooth, weakly smooth, and nonsmooth convex optimization, as well as for stochastic convex optimization under heavy-tailed noise. Numerical experiments are further provided to validate our theoretical results.
Similar Papers
Can SGD Handle Heavy-Tailed Noise?
Optimization and Control
Makes computers learn better with messy data.
Complexity of normalized stochastic first-order methods with momentum under heavy-tailed noise
Optimization and Control
Makes computer learning faster and better.
Minimisation of Submodular Functions Using Gaussian Zeroth-Order Random Oracles
Optimization and Control
Helps computers find the best choices faster.