AI Paper Atlas

Research map / Reinforcement learning / Policy gradient and actor-critic

OGPO: Sample Efficient Full-Finetuning of Generative Control Policies

Sarvesh Patil, Mitsuhiko Nakamoto, Manan Agarwal, Shashwat Saxena, Jesse Zhang, Giri Anantharaman, Cleah Winston, Chaoyi Pan, Douglas Chen, Nai-Chieh Huang, Zeynep Temel, Oliver Kroemer and 5 more

ICML 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Model-based RL with guarantees

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
add success-buffer behavior cloning regularization (ablation)OGPO · Robomimic state-based taskssuccess rateImproved
add Best-of-N inference with critic (ablation)OGPO+ · Robomimic Toolhangsuccess rateNo clear effect
use conservative advantages (all ensemble critics agree on sign) (ablation)OGPO+ · Robomimic image-based taskssuccess rateImproved
use conservative advantages (ablation)OGPO+ · offline-to-online adaptationperformance dipImproved
add Q-variance reduction of TD targets (ablation)OGPO · pixel-based Robomimiccritic accuracyMixed
use PPO policy extraction instead of AWR (flow-matching loss weighting) (ablation)OGPO · Robomimic taskssuccess rateImproved
use PPO likelihood ratio instead of exponentiated advantage weights (ablation)OGPO · Robomimic taskssuccess rateImproved
backpropagate through denoising chain (BPTT) instead of zero-order PPO (ablation)OGPO · Robomimic taskssuccess rateHurt
remove negative-advantage gradients (ablation)OGPO · Robomimic Transportsuccess rateHurt
normalize advantages by group standard deviation (GRPO-style) (ablation)OGPO · Robomimicsuccess rateHurt
add CalQL+WSRL offline-to-online recipe (ablation)OGPO · Robomimicsuccess rateNo clear effect
use pixel-based critic with vanilla OGPO (ablation)OGPO · Robomimic Squaresuccess rateHurt

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