AI Paper Atlas

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

Maximum Average Randomly Sampled: A Scale Free and Non-parametric Algorithm for Stochastic Bandits

Masoud Moravej Khorasani, Erik Weyer

NeurIPS 2023 · Bandits and online learning › Multi-armed bandits · approach: Bandit formulations for ML problems · 0 citations (OpenAlex)

Read the paper

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 MARS (random sub-sampled mean UCB) instead of Vanilla UCB when the assumed variance is wrong (set to 2)MARS · Gaussian bandit, K=5, misspecified variancecumulative regretImproved
Use MARS instead of Bootstrapped UCB and GIRO under misspecified varianceMARS · Gaussian bandit, K=5, misspecified variancecumulative regretImproved
Use MARS instead of Vanilla UCB when variances are correctly knownMARS · Gaussian bandit, K=5, true variancecumulative regretHurt
Use MARS instead of Normal-TS when the Gaussian model is correctly specifiedMARS · Gaussian bandit, K=5, true variancecumulative regretHurt
Use MARS instead of Vanilla UCB, Thompson sampling, Bootstrapped UCB and GIRO on uniform rewardsMARS · Uniform[-1,1] bandit, K=5cumulative regretImproved
Raise PHE perturbation parameter a from 2.1 to 5.1PHE · Gaussian bandit, K=5, true variancecumulative regretImproved
Raise PHE parameter a from 2.1 to 5.1PHE · Uniform[-1,1] bandit, K=5cumulative regretHurt
Use MARS 95% bound instead of naive bootstrap, bootstrapped threshold, sub-Gaussian and Hoeffding boundsMARS UCB · 95% UCB of sample mean, Gaussian and uniform dataupper confidence bound tightnessImproved
Compare MARS bound with naive bootstrap at small sample sizesMARS UCB · 95% UCB of sample mean, small samplesupper confidence bound tightnessHurt

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