Training Object Permanence in World Models

Research official + media 2 src. ~1 min

Introduces WROP: 150 cognitive-science-inspired tasks (Blender-generated, 1.5M samples) testing whether video world models grasp object permanence. Trains PWM-WROP, a 16B world model on AWS Trainium2, which ranks first among continuation models and third overall in a blind Elo study across 14 video models. Data, exam and weights are released.

Why it matters

Top Hugging Face Daily paper for Sep 25 with 177 upvotes; first systematic attempt to train a cognitive prior (object permanence) into video world models at scale.

Importance: 4/5

Top HF Daily paper (177 upvotes, 100+ bump), large-scale training study with open release

world-models video-generation cognitive-science benchmark

Sources

official Training Object Permanence in World Models (arXiv)

secondary Hugging Face Daily Papers entry