{"id":"arxiv/2609.02418","url":"https://arxiv.org/abs/2609.02418","tags":["math.OC"],"score":0,"title":"On the Duality between Feature and Sample Screening","domain":"math.OC","source":"arxiv","authors":["Thu-Le Tran"],"reviews":0,"subject":"math.OC","summary":"Feature and sample screening reduce the cost of machine learning by eliminating irrelevant features and noninformative samples, respectively. Although recognized as primal-dual counterparts, their relationship remains informal and model-dependent. Viewing screening and duality as transformations of objective functions, we introduce Fenchel-Rockafellar (FR) representations, a class of convex problems encompassing the Lasso and SVM that is closed under both transformations. We then prove that feature and sample screening form an equivariant pair: dualization followed by feature screening is…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-02","reactions":{}}