{"id":"arxiv/2604.17067","url":"https://arxiv.org/abs/2604.17067","tags":["math.OC","cs.LG","math.ST"],"score":0,"title":"Trajectory-Restricted Optimization Conditions and Geometry-Aware Linear Convergence","domain":"math.OC","source":"arxiv","authors":["Faris Chaudhry","Anthea Monod","Keisuke Yano"],"reviews":0,"subject":"math.OC","summary":"Linear convergence of first-order methods is typically characterized by global optimization conditions whose constants reflect worst-case geometry of the ambient space. In high-dimensional or structured problems, these global constants can be arbitrarily conservative and fail to capture the geometry actually encountered by optimization trajectories. In this paper, we develop a trajectory-restricted framework for linear convergence based on localized geometric regularity. We introduce restricted variants of the Polyak--Łojasiewicz inequality, error bound, and quadratic growth conditions that…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-04-18","reactions":{}}