AI Paper Atlas

Research map / Bandits and online learning / RL theory and MDPs

Tightening Exploration in Upper Confidence Reinforcement Learning

Hippolyte Bourel, Odalric Maillard, Mohammad Sadegh Talebi

ICML 2020 · Bandits and online learning › RL theory and MDPs · approach: Average-reward and latent MDPs · 7 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 UCRLnew instead of UCRLAverage-reward RL · 6-, 12-, 20-, 40-, 70-, and 100-state RiverSwimnormalized regret bound: 124.9 -> 30.0; 589.3 -> 59.5; 1736.3 -> 94.9; 7243.9 -> 176.9; 22576 -> 293.6; 46394 (UCRL value; UCRLnew 100-state valuImproved
Use UCRLnew instead of SCAL+Average-reward RL · 6-, 12-, 20-, 40-, and 70-state RiverSwimnormalized regret bound: 69.1 -> 30.0; 235.5 -> 59.5; 542.2 -> 94.9; 1609.6 -> 176.9; 3802.4 -> 293.6Improved
Use UCRLnew instead of UCRLBAverage-reward RL · 6-, 12-, 20-, 40-, and 70-state RiverSwimnormalized regret bound: 38.6 -> 30.0; 85.8 -> 59.5; 148.5 -> 94.9; 305.3 -> 176.9; 540.0 -> 293.6Improved
Replace UCRL with UCRLnewUCRL · 6-state RiverSwimregretImproved
Replace UCRL with UCRLnewUCRL · 25-state RiverSwimregretImproved
Replace UCRL with UCRLnewUCRL · 2-room and 4-room grid worldsregretImproved
Replace UCRL with UCRLnewUCRL · 100-state randomly generated MDPregretImproved
Add tight confidence sets and EVI-NOSS planning (ablation)UCRLnew · Numerical environmentsregretImproved
Use tight confidence sets instead of UCRLB-style confidence intervals (ablation)UCRLnew · Numerical environmentsregretImproved
Use EVI-NOSS instead of EVI (ablation)UCRLnew · Numerical environmentsregretImproved

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