atlas

Research map / Bandits and online learning / Multi-armed bandits

Doubly Adversarial Federated Bandits

ICML 2023 · Bandits and online learning › Multi-armed bandits · approach: Multi-agent bandits · 0 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 FedExp3 (federated Exp3 with gossip averaging of loss estimators) instead of GossipUCBfederated multi-armed bandit on MovieLens genre recommendation · MovieLens-Latest-full, complete graph, grid, RGG(0.5)average cumulative regret: significant margin (no numbers in text)Helped
Replace stationary-setting GossipUCB with the adversarial FedExp3 under non-stationary preferencesfederated bandit on MovieLens · MovieLens non-stationary environmentcumulative regret: significant margin (no numbers in text)Helped
Gossip matrix spectral gap: regret scales with (1-sigma_2(W))^{-1/3} as network connectivity changes (ablation)FedExp3 on synthetic Bernoulli federated bandit (N=36, K=20, T=3000) · synthetic, RGG(d) d in {0.3..0.9}, grid, complete graphcumulative regretNo clear effect
Sublinear regret growth with time for FedExp3 under doubly adversarial lossesFedExp3 · synthetic federated banditcumulative regret vs timeHelped

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