Newest papers

PhysTrace: Interpretable Temporal Evaluation of Physical Consistency in Video Generative Models

Taekoan Yoo, Wonkyung Jung, Kyunghun Kim, Kyeongbo Kong · 2026 · arXiv

DOI: 10.48550/arXiv.2610.31429 · open-access PDF

Korean

세계 행동 모델은 예측, 계획 및 행동 생성을위한 물리적 역학을 캡처하기 위해 미리 훈련된 비디오 모델에 점점 더 의존합니다.현재의 평가는 확실성을 평가하고, 탐사 표현을 측정하거나, 예측적 놀라움을 측정하지만, 역사의 조건으로 인한 물리적 기대가 관찰과 충돌하는 것을 거의 식별하지 않습니다.우리는 MLLM 생성, 역사의 조건으로 인한 기대를 관찰 된 비디오 창과 조정하는 훈련이없는 PhysTrace 프레임워크를 도입하여 관찰 된 물리적 위반을 일시적으로 지역화하고, 작업별 탐사 또는 분류없이 해석 가능한 설명을 제공합니다.

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

Abstract

World action models increasingly rely on pretrained video models assumed to capture physical dynamics for prediction, planning, and action generation. Existing evaluations assess plausibility, probe representations, or measure predictive surprise, but rarely identify which history-conditioned physical expectations conflict with observations. We introduce PhysTrace, a training-free framework aligning MLLM-generated, history-conditioned expectations with observed video windows to temporally localize physical violations and provide interpretable descriptions without task-specific probes or classi

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

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=2610.31429, 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