The Invisible Hand of Physics: When Video Diffusion Models Know More Than They Show

• URL: http://arxiv.org/abs/2606.05328v1
• Date: Wed, 03 Jun 2026 18:11:51 GMT
• Title: The Invisible Hand of Physics: When Video Diffusion Models Know More Than They Show
• Abstract summary: We show that physical plausibility is linearly decodable from diffusion transformer states across IntPhys and InfLevel.<n>Surprisingly, this signal is absent from the VAE latent input and emerges inside the denoising transformer itself.
• Score: 19.515022194108955
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Modern video diffusion models generate increasingly realistic and temporally coherent videos, motivating their use as candidate world simulators. Yet it remains unclear whether these models internally encode physical structure, or merely reproduce motion patterns seen during training. We study this question by probing video diffusion models along latent trajectories corresponding to real videos with known physical plausibility. To obtain such trajectories, we approximately invert the deterministic sampling process by integrating the learned velocity field backward from a clean video latent to noise, giving access to the model's intermediate states and attention maps. Using these recovered trajectories, we show that physical plausibility is linearly decodable from diffusion transformer states across IntPhys and InfLevel, reaching around 81.27% average accuracy and outperforming dedicated representation-learning baselines such as V-JEPA and VideoMAE. Surprisingly, this signal is absent from the VAE latent input and emerges inside the denoising transformer itself, despite the model not being trained with a self-supervised predictive objective. These findings suggest that physically meaningful representations can arise as a byproduct of generative denoising.

Related papers

• SIFT: Self-Imagination Fine-Tuning for Physically Plausible Motion in Video Diffusion Models [53.82339966194578]
  Motion entanglement is the unintended coupling of independent motion sources, such as camera movement and object motion.<n>We propose a Self-Imagination Fine-Tuning (SIFT) paradigm, which enables the model to learn from its own generated videos.
  arXiv  Detail & Related papers  (2026-06-26T05:39:40Z)
• Tempered Self-Similarity Alignment for Physically Plausible Video Generation [83.40337664171939]
  Self-similarity loss transforms into probabilistic correspondence and trains the video generative model to align its correspondence with those of the visual foundation model on dynamically changing regions.<n>Our method demonstrates substantial improvements in physical plausibility across diverse interaction scenarios, validating the effectiveness of transferring relational knowledge for video generation.
  arXiv  Detail & Related papers  (2026-05-24T09:28:05Z)
• Video Prediction of Dynamic Physical Simulations With Pixel-Space Spatiotemporal Transformers [3.951575888190684]
  This study investigates a transformer adaptation for video prediction with a simple end-to-end approach, comparing various selftemporal-attention layouts.<n>We introduce a simple yet effective transformer for autoregressive video prediction, utilizing continuous pixel-space representations for video prediction horizon.
  arXiv  Detail & Related papers  (2025-10-23T17:58:45Z)
• LikePhys: Evaluating Intuitive Physics Understanding in Video Diffusion Models via Likelihood Preference [57.086932851733145]
  We introduce LikePhys, a training-free method that evaluates intuitive physics in video diffusion models.<n>We benchmark intuitive physics understanding in current video diffusion models.<n> Empirical results show that, despite current models struggling with complex and chaotic dynamics, there is a clear trend of improvement in physics understanding as model capacity and inference settings scale.
  arXiv  Detail & Related papers  (2025-10-13T15:19:07Z)
• TRAVL: A Recipe for Making Video-Language Models Better Judges of Physics Implausibility [70.24211591214528]
  Video generative models produce sequences that violate intuitive physical laws, such as objects floating, teleporting, or morphing.<n>Existing Video-Language Models (VLMs) struggle to identify physics violations, exposing fundamental limitations in their temporal and causal reasoning.<n>We introduce TRAVL, a fine-tuning recipe that combines a balanced training dataset with a trajectory-aware attention module to improve motion encoding.<n>We propose ImplausiBench, a benchmark of 300 videos (150 real, 150 generated) that removes linguistic biases and isolates visual-temporal understanding.
  arXiv  Detail & Related papers  (2025-10-08T21:03:46Z)
• Physics-Grounded Motion Forecasting via Equation Discovery for Trajectory-Guided Image-to-Video Generation [54.42523027597904]
  We introduce a novel framework that integrates symbolic regression and trajectory-guided image-to-video (I2V) models for physics-grounded video forecasting.<n>Our approach extracts motion trajectories from input videos, uses a retrieval-based pre-training mechanism to enhance symbolic regression, and discovers equations of motion to forecast physically accurate future trajectories.
  arXiv  Detail & Related papers  (2025-07-09T13:28:42Z)
• FlowMo: Variance-Based Flow Guidance for Coherent Motion in Video Generation [51.110607281391154]
  FlowMo is a training-free guidance method for enhancing motion coherence in text-to-video models.<n>It estimates motion coherence by measuring the patch-wise variance across the temporal dimension and guides the model to reduce this variance dynamically during sampling.
  arXiv  Detail & Related papers  (2025-06-01T19:55:33Z)
• Think Before You Diffuse: Infusing Physical Rules into Video Diffusion [55.046699347579455]
  The complexity of real-world motions, interactions, and dynamics introduce great difficulties when learning physics from data.<n>We propose DiffPhy, a generic framework that enables physically-correct and photo-realistic video generation by fine-tuning a pre-trained video diffusion model.
  arXiv  Detail & Related papers  (2025-05-27T18:26:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.