Unified Unbiased Variance Estimation for Maximum Mean Discrepancy: Robust Finite-Sample Performance with Imbalanced Data and Exact Acceleration under Null and Alternative Hypotheses
Shijie Zhong, Yikun Yang, Da Gong, Jiangfeng Fu · 2026
Read Unified Unbiased Variance Estimation for Maximum Mean Discrepancy: Robust Finite-Sample Performance with Imbalanced Data and Exact Acceleration under Null and Alternative Hypotheses by Shijie Zhong, Yikun Yang, Da Gong, Jiangfeng Fu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-13874.txt.
Description
Shijie Zhong, Yikun Yang, Da Gong, and Jiangfeng Fu refine a key statistic for two-sample testing in 2026. Maximum mean discrepancy, or MMD, compares two samples via kernels. Its inferential accuracy hinges on how well its variance is known. Past estimators differ under null and alternative hypotheses, and for balanced or imbalanced data.
The team gives a unified unbiased variance estimation for MMD. They target robust finite-sample performance with imbalanced data. They also give exact acceleration under both hypotheses. A single sound estimator could make two-sample tests simpler and more trusted.
Catalog record
| CREATOR | Shijie Zhong, Yikun Yang, Da Gong, Jiangfeng Fu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-13874