Non-Stationary Lipschitz Bandits

Nicolas Nguyen, Solenne Gaucher, Claire Vernade · 2025

Read Non-Stationary Lipschitz Bandits by Nicolas Nguyen, Solenne Gaucher, Claire Vernade (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-2505-18871.txt.

Introduction

Nicolas Nguyen, Solenne Gaucher, and Claire Vernade studied online choice under change in a 2025 paper on bandit problems. In a bandit task, a learner picks an act, gains a reward, and learns as it goes. Here the set of acts is endless, close acts pay much the same, and the pay map can shift at any time in any way.

The authors built a method that watches for big shifts in total reward and adapts on its own. It splits the act space into a tree of finer parts to track those shifts. With no prior word of when shifts come, the method hits the best loss bound known for this setting. The proof gives the first such sharp pledge for this class of problems.

Catalog record

CREATORNicolas Nguyen, Solenne Gaucher, Claire Vernade
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE174 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-18871

Browse the full library · About the Tower of Babel