Latent Video Prediction Learns Better World Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Alrasheed, Ali J, Parast, Aryan Yazdan, Azam, Basim, Bailey, James, Akhtar, Naveed
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910223523905536
author Alrasheed, Ali J
Parast, Aryan Yazdan
Azam, Basim
Bailey, James
Akhtar, Naveed
author_facet Alrasheed, Ali J
Parast, Aryan Yazdan
Azam, Basim
Bailey, James
Akhtar, Naveed
contents Self-supervised video models are increasingly framed as world models, yet their evaluation remains largely confined to a single top-1 accuracy score on clean benchmarks. This leaves a major gap in comprehending their potential as world models. We present the first systematic study addressing this gap, analyzing four matched-capacity frontier video foundation models, V-JEPA 2.1, V-JEPA 2, VideoPrism, and VideoMAEv2, across five robustness axes relevant to their deployment as video world models: feature discriminability, corruption robustness, fine-grained discrimination, occlusion robustness, and sensitivity to temporal direction. Our evaluations establish that across all five axes, latent-prediction models form a distinct and consistent profile. They degrade more gracefully under pixel corruption, preserve usable class structure rather than mere geometric stability under occlusion, capture fine-grained physical contact cues without reconstructing pixels, and uniquely encode the arrow of time. These advantages can even survive task adaptation: a frozen V-JEPA 2 backbone with a lightweight attentive probe outperforms a fully fine-tuned VideoMAE and a supervised TimeSformer on corruption and occlusion robustness. Our extensive results offer concrete new evidence in favor of latent prediction for robust world modeling.
format Preprint
id arxiv_https___arxiv_org_abs_2605_15618
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Latent Video Prediction Learns Better World Models
Alrasheed, Ali J
Parast, Aryan Yazdan
Azam, Basim
Bailey, James
Akhtar, Naveed
Computer Vision and Pattern Recognition
Artificial Intelligence
Self-supervised video models are increasingly framed as world models, yet their evaluation remains largely confined to a single top-1 accuracy score on clean benchmarks. This leaves a major gap in comprehending their potential as world models. We present the first systematic study addressing this gap, analyzing four matched-capacity frontier video foundation models, V-JEPA 2.1, V-JEPA 2, VideoPrism, and VideoMAEv2, across five robustness axes relevant to their deployment as video world models: feature discriminability, corruption robustness, fine-grained discrimination, occlusion robustness, and sensitivity to temporal direction. Our evaluations establish that across all five axes, latent-prediction models form a distinct and consistent profile. They degrade more gracefully under pixel corruption, preserve usable class structure rather than mere geometric stability under occlusion, capture fine-grained physical contact cues without reconstructing pixels, and uniquely encode the arrow of time. These advantages can even survive task adaptation: a frozen V-JEPA 2 backbone with a lightweight attentive probe outperforms a fully fine-tuned VideoMAE and a supervised TimeSformer on corruption and occlusion robustness. Our extensive results offer concrete new evidence in favor of latent prediction for robust world modeling.
title Latent Video Prediction Learns Better World Models
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2605.15618