Finite-Sample Wasserstein Error Bounds and Concentration Inequalities for Nonlinear Stochastic Approximation
Seo Taek Kong, R. Srikant · 2026
Read Finite-Sample Wasserstein Error Bounds and Concentration Inequalities for Nonlinear Stochastic Approximation by Seo Taek Kong, R. Srikant (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-2602-02445.txt.
Description
Seo Taek Kong and 1 co-author prove finite-sample error bounds for stochastic approximation, the family of iterative methods behind gradient descent and time-based-difference learning, in a 2026 arXiv paper. Such methods return either the last iterate or a running average. The question is how close those outputs sit to a Gaussian limit after finitely many steps.
Their system combines moment bounds for normalized iterates with quantitative Gaussian approximation of the weighted driving noise. The resulting Wasserstein distance bounds yield convergence of moments and high-probability guarantees whose leading terms are Gaussian. The heavier dependence on confidence level hides in a fix that vanishes as samples grow. The authors verify the conditions for nonlinear methods with independent or Markovian data, and for linear methods with Markovian data. Applications include stochastic gradient descent and time-based-difference learning in reinforcement learning.
Catalog record
| CREATOR | Seo Taek Kong, R. Srikant |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 134 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-02445