AI Paper Atlas

Research map / Bandits and online learning / Contextual bandits and Thompson sampling

A Contextual Bandit Bake-off

Alberto Bietti, Alekh Agarwal, John Langford

JMLR 2021 · Bandits and online learning › Contextual bandits and Thompson sampling · approach: Oracle-efficient contextual bandits

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 IWR reduction instead of IPS (ablation)Bag · 516 OpenML datasets, optimized hyperparameterssignificant win-loss difference: +55Improved
use IWR reduction instead of DR (ablation)Bag · 516 OpenML datasetssignificant win-loss difference: +25Improved
use IWR reduction instead of IPS (ablation)epsilon-greedy · 516 OpenML datasetssignificant win-loss difference: +128Improved
use IWR reduction instead of DR (ablation)epsilon-greedy · 516 OpenML datasetssignificant win-loss difference: +150Improved
use -1/0 loss encoding instead of 0/1 (all datasets) (ablation)Greedy · multiclass/multilabel datasetssignificant wins/losses of -1/0 vs 0/1: 132 / 42Improved
use -1/0 loss encoding instead of 0/1 (all datasets) (ablation)epsilon-greedy · multiclass/multilabel datasetssignificant wins/losses: 94 / 27Improved
use -1/0 instead of 0/1 on datasets with at least 10,000 examples (ablation)RegCB-optimistic · 64 large datasetssignificant wins/losses: 10 / 18Hurt
use -1/0 instead of 0/1 on datasets with at least 10,000 examples (ablation)Cover-NU · 64 large datasetssignificant wins/losses: 14 / 20Hurt
shift RCV1 costs from [0,1] to [-1,0] (ablation)all methods · RCV1progressive validation lossNo clear effect
shift MSLR/Yahoo costs with c=-1 (ablation)Greedy · MSLRprogressive validation loss: 0.798 -> 0.791Improved
shift Yahoo costs with c=-1 (ablation)Greedy · Yahooprogressive validation loss: 0.593 -> 0.589Improved
optimize hyperparameters per dataset instead of fixing them (ablation)Cover-NU vs RegCB-optimistic · all datasets (521)significant win-loss difference: RO beats C-nu by 28 -> C-nu beats RO by 33Mixed

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