{"id":"arxiv/2610.11298","url":"https://arxiv.org/abs/2610.11298","tags":["math.OC"],"score":0,"title":"Nonminimizing Attractors of Stochastic Subgradient Methods on Semialgebraic Functions","domain":"math.OC","source":"arxiv","authors":["Jiajin Li"],"reviews":0,"subject":"math.OC","summary":"We disprove a conjecture of Davis, Drusvyatskiy, and Jiang (2026) that subdifferential regularity can be removed from the generic guarantee of convergence to local minimizers for stochastic subgradient methods. We construct a globally Lipschitz, coercive, semialgebraic function in two dimensions for which the method converges to a nonminimizing Clarke critical point. For suitable power-law stepsizes, this convergence holds for open sets of initial points and linear tilts of the objective, uniformly over all perturbation sequences within a prescribed bound.","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-08","reactions":{}}