Sequential anomaly identification with observation control under generalized error metrics

Aristomenis Tsopelakos, Georgios Fellouris · 2024

Read Sequential anomaly identification with observation control under generalized error metrics by Aristomenis Tsopelakos, Georgios Fellouris (2024) 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-2412-04693.txt.

Description

Aristomenis Tsopelakos and Georgios Fellouris study how to spot odd sources in a stream of data while choosing what to watch, in a math paper from 2024. Many sources run at once, but only a few can be sampled at each step. The goal is to name the odd ones fast, with few errors.

The authors set up two forms of the task. One allows a set total number of errors of any kind, and the other sets apart caps for false alarms and missed cases. For each form, they prove the best possible average time needed as errors shrink toward zero. They then give a plan that reaches that bound. The plan mixes rules for when to stop and decide with a chance-based rule for which sources to sample.

Catalog record

CREATORAristomenis Tsopelakos, Georgios Fellouris
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE131 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-04693

Browse the full library · About the Tower of Babel