Probabilistic Floating-Point Round-Off Analysis via Concentration Inequalities
Yichen Tao, Hongfei Fu, Jiawei Chen, Jean-Baptiste Jeannin · 2026
Read Probabilistic Floating-Point Round-Off Analysis via Concentration Inequalities by Yichen Tao, Hongfei Fu, Jiawei Chen, Jean-Baptiste Jeannin (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-2605-04232.txt.
Description
Yichen Tao and three co-authors work on bounding the small rounding errors that build up when computers use floating-point numbers. Worst-case bounds are safe but often far too large to be useful, since the biggest errors are also rare.
Their method gives a probabilistic bound instead: a threshold that the error stays below with a stated level of confidence. They build it by applying concentration results to an existing Taylor-based analysis, after handling a technical obstacle caused by absolute values in that expansion. The aim is bounds that real programs for science and optimization can actually use.
Catalog record
| CREATOR | Yichen Tao, Hongfei Fu, Jiawei Chen, Jean-Baptiste Jeannin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 126 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.04232