Bandits attack function optimization

Philippe Preux, Rémi Munos, Michal Valko · 2026

Read Bandits attack function optimization by Philippe Preux, Rémi Munos, Michal Valko (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-2605-03496.txt.

Description

Philippe Preux, Rémi Munos, and Michal Valko treat function search as a sequential decision problem. A strict budget limits how many times the objective can be evaluated. The view comes from bandit theory. Each test costs, so the method must balance broad search against local refinement near good peaks.

Their Simultaneous Optimistic Optimization algorithm splits the domain in a fixed way. It expands the most promising cells first. The approach carries guarantees on the returned solution, and runs efficiently in practice. Tests on the CEC 2014 suite show how the search balance performs.

Catalog record

CREATORPhilippe Preux, Rémi Munos, Michal Valko
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE32 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.03496

Browse the full library · About the Tower of Babel