{"id":"arxiv/2601.19811","url":"https://arxiv.org/abs/2601.19811","tags":["stat.ML","cs.AI","cs.LG","math.ST","stat.ME"],"score":0,"title":"Revisiting Incremental Stochastic Majorization-Minimization Algorithms with Applications to Mixture of Experts","domain":"stat.ML","source":"arxiv","authors":["TrungKhang Tran","TrungTin Nguyen","Gersende Fort","Tung Doan","Hien Duy Nguyen","Binh T. Nguyen","Florence Forbes","Christopher Drovandi"],"reviews":0,"subject":"stat.ML","summary":"Processing high-volume, streaming data is increasingly common in modern statistics and machine learning, where batch-mode algorithms are often impractical because they require repeated passes over the full dataset. This has motivated incremental stochastic estimation methods, including the incremental stochastic Expectation-Maximization (EM) algorithm formulated via stochastic approximation. In this work, we revisit and analyze an incremental stochastic variant of the Majorization-Minimization (MM) algorithm, which generalizes incremental stochastic EM as a special case. Our approach relaxes…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-01-27","reactions":{}}