{"id":"arxiv/2611.07406","url":"https://arxiv.org/abs/2611.07406","tags":["math.OC"],"score":0,"title":"Random Reshuffling for Smooth Convex Optimization: Dominates Stochastic Gradient Descent","domain":"math.OC","source":"arxiv","authors":["Zijian Liu"],"reviews":0,"subject":"math.OC","summary":"Stochastic Gradient Descent (SGD) is one of the most classical optimization algorithms with favorable theoretical guarantees, yet its practical implementation differs subtly from its well-known form and is often referred to as Shuffling Stochastic Gradient Descent (Shuffling SGD). A particularly popular strategy in Shuffling SGD is Random Reshuffling (RR), which has achieved great empirical success. Despite its strong performance, RR has long been considered a heuristic due to a lack of theoretical support. Over the last decade, people have finally established provable convergence rates for…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-11-05","reactions":{}}