Newest papers

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

Taekoan Yoo, Wonkyung Jung, Kyunghun Kim, Kyeongbo Kong · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31429 · cited by 0 works (OpenAlex) · open-access PDF

German

Welt-Aktionsmodelle verlassen sich zunehmend auf vorbereitete Video-Modelle, die angenommen werden, um die physikalische Dynamik für Vorhersage, Planung und Aktionsgenerierung zu erfassen. Bestehende Bewertungen bewerten die Plausibilität, Sonderepräsentationen oder messen prädiktive Überraschungen, aber identifizieren selten, welche historisch bedingten physikalischen Erwartungen mit Beobachtungen kollidieren. Wir führen PhysTrace ein, ein Training-freies Framework, das MLLM-generierte, historisch bedingte Erwartungen mit beobachteten Videofenstern ausgleicht, um physikalische Verletzungen zeitlich zu lokalisieren und interpretierbare Beschreibungen ohne Aufgaben-spezifische Proben oder Klassifizierungen zu liefern. Wir

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 classifiers. We also introduce IntPhys2-Timing, adding violation intervals to IntPhys2. PhysTrace achieves a Hit Rate of up to 70.7% across physical principles on IntPhys2-Timing and outperforms the compared methods in overall possible-impossible discrimination accuracy on IntPhys2. Applied to reconstructed videos along the trajectories of pretrained diffusion/flow-based video models, PhysTrace reveals attribute- and noise-dependent patterns in physical violation localization, enabling interpretable physical consistency analysis across timesteps.

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

Topics: Consistency (knowledge bases), Computer science, Measure (data warehouse), Artificial intelligence, Action recognition, Action (physics)

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