AI 视频模型不懂「物体还在那」，但能教
人类婴儿天生知道：东西被挡住，它还在。现在的视频生成模型没有这个直觉——你让它续写一个球滚到墙后的画面，它可能直接让球消失。研究者造了一套 150 个认知科学启发的任务，用 Blender 生成 150 万段视频，专门训练模型「物体恒存」:球被挡住后，该从哪边出来、速度怎么变。训练后的 16B 模型在盲测里打败了所有同类续写模型。这不是你明天能用的功能，但它指向一个关键趋势：AI 对物理世界的理解，正在从「画得像」走向「想得对」。
📄 原文摘要(英文)
Object permanence and solidity are hallmarks of human cognitive priors. Recent studies show that video generation models, a paradigmatic class of current world models, have begun to show emerged reasoning abilities, making them ideal candidates for building human-like physical intelligence. Do video models have emerged object permanence in them? If not, could we train them with a core-cognition inspired dataset? We introduce WROP (World Reasoning with Object Permanence), a data infrastructure of 150 hand-designed cognitive science inspired tasks, divided into six cognitive categories. We build Blender generators that randomize speed, lighting, camera angle, and other nuisance parameters while preserving each task's cognitive structure, yielding 10,000+ samples per task. We release a 1.5M-sample training corpus and a 300-question exam. On this exam we evaluate 14 video models: 3 reference-to-video, 7 edit, and 4 continuation, among which PWM-WROP, our 16B world model. In a blind pairwise Elo study, PWM-WROP ranks first among continuation models and third overall, behind only a statistical tie between two reference-to-video models. We release the data, exam, model answers, scores, weights, and PWM, our native-PyTorch training stack on AWS Trainium2.