On McDiarmid's inequality for Hamming distance
Fan · math.PR · 2016-11-12 · 原文
We improve the rate function of McDiarmid's inequality for Hamming distance. In particular, applying our result to the separately Lipschitz functions of independent random variables, we also refine the convergence rate function of McDiarmid's inequality around a median. Moreover, a non-uniform bound for the distance between the medians and the mean is also given. We also give some extensions of McDiarmid's inequalities to the case of nonnegative functionals of dependent random variables.
讲义
讲义·推断 依据「原文」自动生成的结构化摘要（推断），非原文表述；以原文为准。
1. 人话版
We improve the rate function of McDiarmid's inequality for Hamming distance.
In particular, applying our result to the separately Lipschitz functions of independent random variables, we also refine the convergence rate function of McDiarmid's inequality around a median.
2. 领域脉络
本文类目：math.PR，属于其所在研究脉络的最新进展。
3. 机制拆解
Moreover, a non-uniform bound for the distance between the medians and the mean is also given.
We also give some extensions of McDiarmid's inequalities to the case of nonnegative functionals of dependent random variables.
4. 证据与数字
摘要未给出量化结果——留意原文的实验与数据。
5. 反例与边界
摘要未声明局限与反例——这是需要警惕的信号，精读时先问边界。
6. 跨领域连接与意外收获
思考本文机制能否迁移到你正在跟进的问题。
7. 可复用方法
把本文机制与你手头项目对照，找一个两周内能验证的最小实验。
8. 术语表
精读时把不熟的术语记入此处，作为下次回忆的锚点。