First-Exit Time Analysis for Truncated Heavy-Tailed Dynamical Systems
Xingyu Wang, Chang-Han Rhee · 2026
Read First-Exit Time Analysis for Truncated Heavy-Tailed Dynamical Systems by Xingyu Wang, Chang-Han Rhee (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-2602-07968.txt.
Description
Stochastic systems with heavy-tailed noise can jump suddenly, and truncating those jumps, as in some tuning algorithms, changes when the system first escapes a region. Xingyu Wang and Chang-Han Rhee analyze first-exit times for a stochastic gap equation and its truncated variant, where each step is capped at a threshold.
By connecting large-deviation theory with metastability, they characterize the joint distribution of exit times and exit locations precisely. As the truncation threshold varies, a discrete hierarchy of phase shifts in exit times appears, illustrating the catastrophe principle in which rare large jumps dominate. Probabilists and machine-learning theorists studying heavy-tailed stochastic gradient descent gain exact asymptotics for a practically motivated model.
Catalog record
| CREATOR | Xingyu Wang, Chang-Han Rhee |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 178 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07968