{"id":"arxiv/2606.23637","url":"https://arxiv.org/abs/2606.23637","tags":["cs.LG","math.OC"],"score":0,"title":"Muown Implicitly Performs Angular Step-size Decay","domain":"cs.LG","source":"arxiv","authors":["Florian Hübler","Kai Lion","Antonio Orvieto","Niao He"],"reviews":0,"subject":"cs.LG","summary":"Matrix-aware optimizers such as Muon and Muown have recently shown strong empirical performance for pre-training Transformers. In particular, Muown separates each weight matrix into row magnitudes and an un-normalized direction variable, updating the former with Adam and the latter with Muon. We show that the directional update of Muown is equivalent to a Riemannian step on the normalized directions, while the magnitude of the un-normalized parameterization only modulates the angular step size. This explains the step-size stability of Muown and suggests making the angular step size explicit.…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-06-22","reactions":{}}