AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

Test-Time Mixture of World Models for Embodied Agents in Dynamic Environments

Jinwoo Jang, Minjong Yoo, Sihyung Yoon, Honguk Woo

ICLR 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Extrapolation and single-sample adaptation · 2 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 multi-granular (object plus scene) routing instead of object-only (ablation)TMoW · VirtualHome (unseen, zero-shot)success rate (%): 65.25 -> 80.74Improved
use multi-granular routing instead of scene-only (ablation)TMoW · VirtualHome (unseen, zero-shot)success rate (%): 8.74 -> 80.74Improved
add test-time prototype refinement (ablation)TMoW · VirtualHome (unseen, zero-shot)success rate (%): 73.30 -> 80.74Improved
add test-time prototype refinement (ablation)TMoW · VirtualHome (unseen, zero-shot)pending steps: 14.85 -> 13.12Improved
distilled mixture augmentation instead of training new model from scratch (ablation)TMoW · VirtualHome (unseen, 1-shot)success rate (%): 59.84 -> 81.56Improved
use top-3 routing instead of top-1 (ablation)TMoW · VirtualHome (unseen, zero-shot)success rate (%): 65.43 -> 80.16Improved
increase top-K from 3 to 5 (ablation)TMoW · VirtualHome (unseen, zero-shot)success rate (%): 80.16 -> 77.52Hurt
increase top-K from 3 to 7 (ablation)TMoW · VirtualHome (unseen, zero-shot)success rate (%): 80.16 -> 66.01Hurt
TMoW vs SayCanPay in zero-shot unseen domainsTMoW · VirtualHome (unseen)success rate (%): 49.53 -> 80.16Improved
TMoW vs SayCanPay on RLBench unseen domainsTMoW · RLBench (unseen)success rate (%): 38.54 -> 62.75Improved
TMoW vs SayCanPay in 5-shot expansionTMoW · VirtualHome (unseen, 5-shot)success rate (%): 58.88 -> 83.61Improved

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