cs.LGApr 27, 2026

Stochastic simultaneous optimistic optimization

Authors: Michal ValkoAlexandra CarpentierRémi Munos

Organizations: INRIA Lille - Nord Europe, SequeL team, 40 avenue Halley 59650, Villeneuve d’Ascq, France · Statistical Laboratory, CMS, Wilberforce Road, CB3 0WB, University of Cambridge, United Kingdom

Abstract

We study the problem of global maximization of a function f given a finite number of evaluations perturbed by noise. We consider a very weak assumption on the function, namely that it is locally smooth (in some precise sense) with respect to some semi-metric, around one of its global maxima. Compared to previous works on bandits in general spaces (Kleinberg et al., 2008; Bubeck et al., 2011a) our algorithm does not require the knowledge of this semi-metric. Our algorithm, StoSOO, follows an optimistic strategy to iteratively construct upper confidence bounds over the hierarchical partitions of the function domain to decide which point to sample next. A finite-time analysis of StoSOO shows that it performs almost as well as the best specifically-tuned algorithms even though the local smoothness of the function is not known.

Explore similar work

CardsList
  1. Bandits attack function optimization

    May 5, 2026Philippe Preux, Rémi Munos, Michal ValkoBanditsObjective Functions