AI Paper Atlas

Research map / Reinforcement learning / MDPs and planning

DOPE: Doubly Optimistic and Pessimistic Exploration for Safe Reinforcement Learning

Archana Bura, Aria Hasanzadezonuzy, Dileep Kalathil, Srinivas Shakkottai, Jean-Francois Chamberland

NeurIPS 2022 · Reinforcement learning › MDPs and planning · approach: Offline RL sample complexity · 9 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
combine model optimism and reward optimism with pessimistic constraintsDOPE · Finite-horizon CMDPobjective regret: reduces the objective regret bound over OptPessLP by a factor of \sqrt{|\mathcal{S}|}Improved
use DOPE instead of OptPess-LPFactored CMDP · Factored CMDPobjective regretImproved
use DOPE instead of OptCMDPFactored CMDP · Factored CMDPconstraint regretImproved
use DOPE instead of OptPess-LPMedia Streaming CMDP · Media Streaming CMDPobjective regretImproved
use DOPE instead of OptCMDPMedia Streaming CMDP · Media Streaming CMDPconstraint regretImproved
use a less conservative safe baseline policy (ablation)DOPE · Factored CMDP and Media Streaming CMDPobjective regretImproved
use AlwaysSafe π_α instead of AlwaysSafe π_T (ablation)AlwaysSafe · Factored CMDPobjective regretImproved

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