AI TrackerFrontier desk · AEST
Back to news
normalaitech

GeoState: Learning View-Decoupled Geometry States for 3D World Modeling

arXiv cs.AI · 16 Oct 2026, 03:56 AEST

GeoState introduces a view-decoupled geometry representation for predictive 3D world modeling. This approach separates scene-state evolution from camera-dependent observation, enhancing prediction accuracy.

Key facts

Read aloud

GeoState introduces a view-decoupled geometry representation for predictive 3D world modeling. This approach separates scene-state evolution from camera-dependent observation, enhancing prediction accuracy. Recent 3D world models increasingly use features from geometry foundation models GFMs as predictive world states, benefiting from their rich geometric priors. We propose GeoState, a view-decoupled geometry representation for predictive 3D world modeling that separates scene-state evolution from camera-dependent observation. Experiments show that GeoState outperforms prior 3D world models on Waymo, supports geometrically consistent cross-view querying, and generalizes zero-shot to KITTI and nuScenes.

Sources in this cluster