AI Paper Atlas

Research map / Video / Video generation

Infinite-World: Scaling Interactive World Models to 1000-Frame Horizons via Pose-Free Hierarchical Memory

Ruiqi Wu, Xuanhua He, Meng Cheng, Tianyu Yang, Yong Zhang, Zhuoliang Kang, Xunliang Cai, Xiaoming Wei, Chun-Le Guo, Chongyi Li, Ming-Ming Cheng

ICML 2026 · Video › Video generation · approach: World-model video generation · 0 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
add revisit-dense dataset finetuning (ablation)Infinite-World baseline · Five-chunk horizonMemory consistency rank: 2.40 -> 1.83Improved
add revisit-dense dataset finetuning (ablation)Infinite-World baseline · Five-chunk horizonAction responsiveness rank: 2.95 -> 1.61Improved
add uncertainty-aware action labeling (ablation)Infinite-World baseline · Five-chunk horizonAction responsiveness rank: 2.95 -> 2.17Improved
add uncertainty-aware action labeling (ablation)Infinite-World baseline · Five-chunk horizonVisual fidelity rank: 2.10 -> 1.78Improved
add UAL on top of RDD finetuning (ablation)Infinite-World · Five-chunk horizonAction responsiveness rank: 1.61 -> 1.38Improved
add UAL and RDD together on top of baseline (ablation)Infinite-World · Five-chunk horizonMemory consistency rank: 2.40 -> 1.69Improved
use hierarchical compression instead of no compression or direct compression (ablation)HPMC · GPU memory on 80GB H800GPU memory vs context length: plateau at approximately 45GB; no compression OOM at over 180 framesImproved
scale pilot training data from 100 to 1,000 sequences (ablation)toy DiT generator with memory · pilot study on synthetic 3D scenesloop-closure qualityNo clear effect
infer on 6 chunks when trained with max 4-chunk context (ablation)toy DiT generator with memory · pilot study on synthetic 3D scenesmemory stabilityHurt
Infinite-World versus Yume 1.5 on VBench averageInfinite-World · Long-horizon interactive benchmarkVBench average score: 0.8141 vs 0.8119No clear effect

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