Newest papers

PhysiCalWorld: In-Context Physics Calibration for World Models

Yunhao Luo, He Wentao, Yinpei Dai, Hyelin Nam, Hongze Fu, Joyce Yue Chai, Nima Fazeli, Jeong Joon Park · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.30125 · cited by 0 works (OpenAlex) · open-access PDF

Russian

Существующие модели мира обычно воспринимают сцену через текущую визуальную рамку или узкое историческое окно. Однако эти наблюдения могут не раскрывать физические свойства, которые определяют динамику перехода, такие как распределение массы объекта, амортизация сустава или поверхностное трение. Без калибровки этих физических факторов для мировых моделей, прогнозы, как правило, регрессируют к эмпирическому предшественнику и вызывают сильное дрейф в новейших сценах тестирования. Для этого мы предлагаем PhysiCalWorld, который калибрует модели мира, обусловленные действием, к конкретной динамике примера через контекстное пробуждение, без тонкого настройки или аналитического моделирования физики. Наш метод кодирует исследовательский клип взаимодействия, собранный

Machine translation (Meta M2M100); the English original follows.

Abstract

Existing world models typically perceive a scene through current visual frame or a narrow history window. However, these observations may not reveal the physics properties that determine transition dynamics, such as an object's mass distribution, joint damping, or surface friction. Without calibrating these physics factors for world models, predictions tend to regress toward an empirical prior and induce severe drift in novel test scenes. To this end, we propose PhysiCalWorld, which calibrates action-conditioned world models to the instance-specific dynamics via in-context prompting, without fine-tuning or analytical physics modeling. Our method encodes an exploratory interaction clip collected within the test scene into a compact physics descriptor token, which grounds subsequent predictions in the scene's actual physical behavior. PhysiCalWorld is trained end-to-end with the latent prediction loss alone, without physics labels. At inference time, the descriptor token is extracted once and can be reused across rollouts with negligible compute overhead. We evaluate PhysiCalWorld across four environments covering diverse physics variations and show that it substantially improves rollout prediction accuracy and visual planning success rate. The learned descriptors also form a physically structured latent space and generalize beyond the physics ranges seen during training.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Inference, Calibration, Physics beyond the Standard Model, Physics, Computer science, Space (punctuation)

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.30125, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange