AI Paper Atlas

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

Neural Exploitation and Exploration of Contextual Bandits

Yikun Ban, Yuchen Yan, Arindam Banerjee, Jingrui He

JMLR 2026 · 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
learned exploration network f2 instead of UCB bound (EE-Net vs NeuralUCB)neural contextual bandit · MNISTcumulative regret: 943.5 -> 842.3Improved
EE-Net instead of NeuralTSneural contextual bandit · Disincumulative regret: 523.2 -> 476.4Improved
EE-Net instead of Neural-epsilonneural contextual bandit · Yelpcumulative regret: 5276.1 -> 4403.1Improved
EE-Net instead of NeuralTSneural contextual bandit · Movielenscumulative regret: 1583.1 -> 1472.4Improved
neural networks instead of linear UCBcontextual bandit · MNISTcumulative regret: 7863.2 -> 842.3Improved
decay epsilon over time (Neural-epsilon+) (ablation)Neural-epsilon · Movielenscumulative regret: 1573.4 -> 1578.4No clear effect
decay epsilon over time (Neural-epsilon+) (ablation)Neural-epsilon · Yelpcumulative regret: 5276.1 -> 5162.4Improved
raise exploration network input dimension from 10 to 200 (ablation)EE-Net · Movielenscumulative regret: 1472.3 -> 1452.1Improved
raise exploration network input dimension from 200 to 500 (ablation)EE-Net · Movielenscumulative regret: 1452.1 -> 1467.6Hurt
use label y1 = r - f1 instead of abs or ReLU variants for f2 (ablation)EE-Net · Movielens and MNISTcumulative regretImproved
add exploration network f2 (extra training cost)EE-Net vs NeuralUCB/TS · training timecomputation: approximately 38-46% more computationHurt
replace matrix-inverse UCB with single forward pass of f2EE-Net vs NeuralUCB/TS · inference timedecision time: 32-59% faster inferenceImproved

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