Best of both worlds: Stochastic & adversarial best-arm identification

stat.ML, cs.LG

Submitted: 2026-04-16

Updated: 2026-10-19

Comments: 44 pages. Corrected version of the COLT 2018 paper, with Lean formalizations of the principal theoretical results: https://github.com/victorgabillon/best-of-both-worlds-bandits

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Related papers

• Behavior of prediction performance metrics with rare events
• Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
• A Posterior-Dynamics Framework for Imaging Inverse Problems with Pretrained Diffusion Priors
• Efficient quadratic entropy with distance sketches
• Embedding-Bias in Conditional Independence Testing
• From Log-Odds to Shapley Values: An Explanatory Geometry for the Weighted Naive Bayes Classifier