AI Paper Atlas

Research map / Reinforcement learning / Offline RL

When to Trust Your Simulator: Dynamics-Aware Hybrid Offline-and-Online Reinforcement Learning

Haoyi Niu, Shubham Sharma, Yiwen Qiu, Ming Li, Guyue Zhou, Jianming HU, Xianyuan Zhan

NeurIPS 2022 · Reinforcement learning › Offline RL · approach: Offline-to-online fine-tuning · 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
replace adaptive value regularization weight with uniform weight (add adaptive omega) (ablation)H2O · HalfCheetah Medium Replay - Gravityaverage return: 6675±179 -> 6813±289Improved
remove dynamics-ratio importance weight in Bellman error (add it) (ablation)H2O · HalfCheetah Medium Replay - Gravityaverage return: 4721±196 -> 6813±289Improved
add dynamics ratio to uniform-weight variant (ablation)H2O · HalfCheetah Medium Replay - Gravityaverage return: 5223±198 -> 6675±179Improved
remove value regularization entirely (add it) (ablation)H2O · HalfCheetah Medium Replay - Gravityaverage return: 6501±147 -> 6813±289Improved
add dynamics ratio to variant without value regularization (ablation)H2O · HalfCheetah Medium Replay - Gravityaverage return: 5290±356 -> 6501±147Improved
add value regularization when dynamics ratio is absent (ablation)H2O · HalfCheetah Medium Replay - Gravityaverage return: 5290±356 -> 4721±196Hurt
naively add offline real data to policy evaluation and improvement of DARC (DARC+)DARC · HalfCheetah Medium Expert - Gravityaverage return: 4759±353 -> 72±109Hurt
use H2O instead of CQL trained only on real offline dataCQL · HalfCheetah Medium - Gravityaverage return: 6066±73 -> 7085±416Improved
use H2O instead of SAC trained in imperfect simulatorSAC(sim) · HalfCheetah Medium - Frictionaverage return: 2684±2646 -> 6848±445Improved
DARC versus H2O on Medium Expert FrictionH2O · HalfCheetah Medium Expert - Frictionaverage return: 9038±1480 vs 6745±562Hurt

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