AI Paper Atlas

Research map / Reinforcement learning / Offline RL

Learning Massively Multitask World Models for Continuous Control

Nicklas Hansen, Hao Su, Xiaolong Wang

ICLR 2026 · Reinforcement learning › Offline RL · approach: Offline RL sample complexity theory · 11 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
Condition the agent on language instructions (ablation)Newt · MMBench 200 tasksnormalized score: 0.371 -> 0.438Improved
Pretrain on demonstrations then finetune online instead of training from scratchNewt · MMBench transfer tasks, zero-shotnormalized score: 0.013 -> 0.192Improved
Pretrain on demonstrations then finetune online instead of training from scratchNewt · MMBench transfer tasks, 100k stepsnormalized score: 0.480 -> 0.868Improved
Add RGB image observations to state agent (ablation)Newt · MMBench all domainsnormalized score: 0.438 -> 0.442No clear effect
Add RGB image observations (ablation)Newt · MMBench RoboDesknormalized score: +0.125 (to 0.500)Improved
Add RGB image observations (ablation)Newt · MMBench DMControlnormalized score: -0.029 (to 0.471)Hurt
Replace unseen object noun with 'cube' in instruction (ablation)Newt · Unseen push taskssuccess rate: 0.3 -> 21.0Improved
Replace unseen object noun with 'cube' in instruction (ablation)Newt · Unseen pick-and-place taskssuccess rate: 10.5 -> 0Hurt
Use any single way of using demonstrations (pretraining, oversampling, BC loss) (ablation)Newt · MMBenchnormalized scoreImproved
Scale model and batch size (ablation)Newt · MMBench 200 tasksnormalized scoreImproved
Use Newt instead of PPO, FastTD3 and TD-MPC2Newt · MMBench state obsnormalized scoreImproved
Use Newt on MuJoCo, Box2D and Atari domains vs BCNewt · MMBench MuJoCo/Box2D/Atarinormalized scoreNo clear effect

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