AI Paper Atlas

Research map / Reinforcement learning / Offline RL

Why So Pessimistic? Estimating Uncertainties for Offline RL through Ensembles, and Why Their Independence Matters

Seyed Kamyar Seyed Ghasemipour, Shixiang Shane Gu, Ofir Nachum

NeurIPS 2022 · Reinforcement learning › Offline RL · approach: Pessimism and uncertainty in offline RL · 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 independently trained Q-functions with separate targets instead of shared pessimistic targets (ablation)MSG · D4RL antmazetask performanceImproved
use Shared-LCB targets instead of independent targets (ablation)MSG · D4RL Gymtask performanceHurt
use Shared-Min targets instead of independent targets (ablation)MSG · D4RL Gymtask performanceMixed
increase ensemble size from 4 to 64 (ablation)MSG · D4RL Gymtask performanceNo clear effect
increase ensemble size (ablation)MSG · D4RL antmaze-largetask performanceImproved
use a small ensemble (N=4) rather than larger ensembles (ablation)MSG · D4RL antmazetask performanceMixed
use efficient ensemble approximations instead of separate-network deep ensembles (ablation)MSG · D4RL antmazetask performanceHurt
use Batch Ensembles among the efficient ensemble variants (ablation)MSG · D4RL Gymtask performanceImproved
increase the CQL-style support regularizer weight (ablation)MSG · narrow-distribution or chaotic-dynamics datasetstask performanceImproved
use shared pessimistic targets, which can make the LCB penalty positiveensemble Q-value evaluation · offline RL toy MDP constructionspessimistic value estimation: 221 of 1000Hurt
use MSG with deep ensembles rather than reported prior methodsD4RL antmaze · antmaze-medium and antmaze-largetask performanceImproved
use smaller network architectures than the baselinesMSG · RL Unpluggedtask performanceMixed

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