High-Dimensional Sequential Change Detection
Robert Malinas, Dogyoon Song, Benjamin D. Robinson, Alfred O. Hero III · 2025
Read High-Dimensional Sequential Change Detection by Robert Malinas, Dogyoon Song, Benjamin D. Robinson, Alfred O. Hero III (2025) 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-2502-05377.txt.
Description
Robert Malinas, Dogyoon Song, Benjamin Robinson, and Alfred Hero tackle fast change detection in high-dimensional data streams, in this 2025 paper. A sensor network may watch hundreds of variables at once, and the task is to spot, as quickly as possible, the moment their pattern shifts. In high dimensions the new pattern must be guessed from a short window of past data.
The team extends quickest change detection theory to the case where the number of variables grows with the window size. They define a normalized divergence measure and show it governs the delay before detection. With a perfect guess of the new pattern, delay scales inversely with that measure alone. They also name estimators that shrink the guess error fastest. Their method is guaranteed to beat approaches built on old fixed-dimension theory.
Catalog record
| CREATOR | Robert Malinas, Dogyoon Song, Benjamin D. Robinson, Alfred O. Hero III |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-05377