AI Paper Atlas

Research map / Reinforcement learning / Offline RL

MOBODY: Model-Based Off-Dynamics Offline Reinforcement Learning

Yihong Guo, Yu Yang, Pan Xu, Anqi Liu

ICLR 2026 · Reinforcement learning › Offline RL · approach: Offline RL sample complexity theory · 6 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
learn target dynamics with separate action encoders and roll out the model, instead of reward regularization / filteringoff-dynamics offline RL · MuJoCo gravity/friction sumnormalized score: 986.13 -> 1427.26Improved
use dynamics trained on target data only instead of MOBODY dynamics (remove shared/separate-encoder model) (ablation)MOBODY · Walker2d-Gravity 5.0normalized score: 46.05 -> 3.56Hurt
use CQL instead of target-Q-weighted BC policy learning (ablation)MOBODY · Walker2d-Gravity 5.0normalized score: 46.05 -> 11.37Hurt
add cycle transition loss (VAE-style) (ablation)MOBODY dynamics model · Walker2d-Gravity 0.1normalized score: 35.34 -> 65.85Improved
add target-Q weighting to behavior cloning (vs vanilla BC) (ablation)MOBODY · Walker2d-Gravity 0.1normalized score: 19.53 -> 65.85Improved
target-Q weighting vs vanilla BC at gravity 2.0 (ablation)MOBODY · Walker2d-Gravity 2.0normalized score: 47.13 -> 44.32No clear effect
learn dynamics with separate action encoders vs combined source+target data (ablation)MOBODY dynamics · Walker2d-friction-0.5normalized score: 41.38 -> 76.96Improved
separate action encoders vs combined data for dynamics (ablation)MOBODY dynamics · Walker2d-friction-0.5rollout MSE: 1.96 -> 1.25Improved
pretrain-finetune dynamics vs MOBODY dynamics (ablation)MOBODY dynamics · Ant-friction-0.5normalized score: 51.09 -> 62.41Improved
MOBODY vs best baseline at larger dynamics shiftMOBODY · Walker2d-Gravity 5.0normalized score: 5.65 -> 46.05Improved
MOBODY vs SRPO at mild shiftMOBODY · Hopper-Gravity 0.5normalized score: 67.06 -> 33.57Hurt
MOBODY vs DARA on Hopper morph-foot MediumMOBODY · Hopper morph-foot Mediumnormalized score: 40.61 -> 13.05Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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