{"id":"arxiv/2601.01853","url":"https://arxiv.org/abs/2601.01853","tags":["math.OC"],"score":0,"title":"Long-Run Dynamics of AdaGrad-Norm: Trajectory Stability and Asymptotic Stationarity","domain":"math.OC","source":"arxiv","authors":["Ruinan Jin","Xiaoyu Wang"],"reviews":0,"subject":"math.OC","summary":"AdaGrad-Norm is widely studied, yet its long-run behavior in smooth nonconvex optimization remains delicate. For the standard current-denominator recursion, the same stochastic gradient determines both the update and its normalizer, rendering the effective stepsize nonpredictable. At the square-root normalization, the quadratic smoothness budget admits only logarithmic control rather than a uniform summability bound. We establish trajectory stability and full-sequence asymptotic stationarity directly for this unmodified recursion. Under global smoothness, non-flatness at infinity,…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-01-05","reactions":{}}