AI Paper Atlas

Research map / Bandits and online learning / Online convex optimization and games

Robust Lipschitz Bandits to Adversarial Corruptions

Yue Kang, Cho-Jui Hsieh, Thomas Chun Man Lee

NeurIPS 2023 · Bandits and online learning › Online convex optimization and games · approach: Best-of-both-worlds bandit algorithms · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use Robust Zooming with a confidence radius containing a corruption-budget termZooming · Lipschitz bandits, strong adversary, known Cregret upper bound: O\left( \ln{(T)}^{\frac{1}{d_{z}+2}} T^{\frac{d_{z}+1}{d_{z}+2}} + C^{\frac{1}{d_{z}+1}} T^{\frac{d_{z}}{d_{z}+1}} \righImproved
Cap the corruption part of Robust Zooming’s confidence radius at 1 (ablation)Robust Zooming · Strong adversary, known Cregret upper boundNo clear effect
Perform RMEL elimination after each round instead of after each epoch (ablation)RMEL · Weak adversary, unknown Celimination speedImproved
Use a larger tolerance ratio B between adjacent RMEL layers (ablation)RMEL · Weak adversary, unknown Cregret upper boundHurt
Use RMEL or BoB Robust Zooming rather than classic Zooming under Oracle and Garcelon attacksLipschitz bandits · Simulated strong-adversary Oracle and Garcelon attackscumulative regretImproved
Use RMEL rather than BoB Robust Zooming under the tested attack settingsLipschitz bandits · Simulated Oracle and Garcelon attackscumulative regretImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy