AI RESEARCH

Bregman meets L\'evy: Stochastic mirror descent with heavy-tailed noise in continuous and discrete time

arXiv CS.LG

ArXi:2606.03769v1 Announce Type: cross We study the robustness of stochastic mirror descent (SMD) under heavy-tailed noise, focusing on whether the method retains its convergence guarantees when run with infinite-variance stochastic gradient input. To address this question in a principled manner, we begin by