Paper page - Principia: Relational Physics Tests for Video Models

Source: https://huggingface.co/papers/2609.04200

Abstract

Principia evaluates video generators on Newtonian physics via calibration-independent relational consistency across paired objects, revealing major physical reasoning gaps.

Evaluating physical reasoning in video models is difficult because absolute motion measurements depend on frame rate, object scale, and camera calibration, all of which are often ambiguous or unavailable in generated video. We propose a different approach. When two objects in the same scene obey the same physical law, their motions must satisfy predictable relationships, and these relationships hold independent of calibration. We introducePrincipia, a benchmark that evaluatesNewtonian physicsthroughrelational consistencybetween paired objects.Principiaspans eight phenomena - gravity, restitution, friction, rotational inertia, projectile motion, momentum, pendulum, and mass-spring oscillation - across translational, rotational, collisional, and oscillatory dynamics, using real-world scenes recorded under controlled protocols. We also introduce acalibration-independent consistency scorethat quantifies physical violation directly in image space. Across thousands of generations from six state-of-the-artvideo generators, no model exceeds 0.42 onPrincipiadespite all scoring around 0.8 on VBench.Vision-language modelsare evaluated on their ability to detect relational physics violations, with the best model achieving only 67% accuracy and most performing near chance level.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2609\.04200

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.04200 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.04200 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.04200 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.