Functional Central Limit Theorem for Stochastic Gradient Descent
Kessang Flamand, Victor-Emmanuel Brunel · 2026
Read Functional Central Limit Theorem for Stochastic Gradient Descent by Kessang Flamand, Victor-Emmanuel Brunel (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-2602-15538.txt.
Description
Flamand and Brunel trace the full path of stochastic gradient descent, rather than just its final point or its average. For a convex objective under mild regularity rules, they prove a functional central limit theorem for the suitably rescaled trajectory. The result gives a diffusion limit that captures how the iterates wobble around the minimiser over time, with temporal structure intact.
That full-path view sets the work apart from classic limit theorems for the last iterate or for Polyak-Ruppert averages, which flatten time away. It also reaches non-smooth cases such as robust location estimation, including the geometric median. Readers who tune or analyse stochastic optimisation will find a precise picture of long-run fluctuation, not just of end-point error.
Catalog record
| CREATOR | Kessang Flamand, Victor-Emmanuel Brunel |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 66 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-15538