{"id":"arxiv/2606.17426","url":"https://arxiv.org/abs/2606.17426","tags":["stat.ML","cs.LG","math.PR"],"score":0,"title":"Bounded Difference Concentration for Infinitely Exchangeable Sequences with Applications to AI Benchmark Uncertainty","domain":"stat.ML","source":"arxiv","authors":["Fangyuan Lin","Spencer Frei","Victor H. de la Pena"],"reviews":0,"subject":"stat.ML","summary":"We consider the concentration properties of functions of infinitely exchangeable random variables. By conditioning on the de Finetti directing measure, we show that the deviation of any function with bounded-difference constants $c_1, \\dots, c_n$ decomposes into a conditional sampling fluctuation and a latent mixture fluctuation. When this latent mixture is $σ_{\\mathrm{mix}}^2$-subgaussian, we establish a concentration inequality with an effective variance proxy of $\\frac{1}{4}\\sum_i c_i^2 + σ_{\\mathrm{mix}}^2$. Crucially, we demonstrate that for zero-sum linear contrasts, such as the…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-06-16","reactions":{}}