{"id":"arxiv/2607.23836","url":"https://arxiv.org/abs/2607.23836","tags":["math.CO","cs.IT","math.PR"],"score":0,"title":"Elementary Symmetric Polynomial Inequalities for Centered Vectors and Matrices","domain":"math.CO","source":"arxiv","authors":["Yanjun Han","Jonathan Niles-Weed"],"reviews":0,"subject":"math.CO","summary":"We prove new inequalities for elementary symmetric polynomials (ESPs) for vectors that sum to zero, and for square matrices with zero row and column sums. We apply these results to obtain a unified upper bound on the mean-field approximation guarantee for permutation mixtures, as well as a sharp $χ^2$ version of the de Finetti theorem for finite sequences over a small alphabet. The main proof ideas were developed by the GPT-5.5 Pro model.","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-26","reactions":{}}