# Physics-Grounded Fluid Video Generation with a Simulation Dataset and Dual-Stream Optical-Flow Supervision

arXiv: https://arxiv.org/abs/2607.25321 · HTML: https://mlsanity.com/paper/2607.25321/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Video generation, one of 14 directions in Diffusion models.
- Builds on 20 of its 22 references that are in the map.
- First posted on arXiv 28 Jul 2026.

## Placed in

- Diffusion models / [Video generation](https://mlsanity.com/diffusion/video-generation/)

## Builds on (most cited first, 20 of 20)

- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [RAFT: Recurrent All-Pairs Field Transforms for Optical Flow](https://mlsanity.com/paper/2003.12039/) (2020)
- [Video Diffusion Models](https://mlsanity.com/paper/2204.03458/) (2022)
- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [Make-A-Video: Text-to-Video Generation without Text-Video Data](https://mlsanity.com/paper/2209.14792/) (2022)
- [FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space](https://mlsanity.com/paper/2506.15742/) (2025)
- [PhysGaussian: Physics-Integrated 3D Gaussians for Generative Dynamics](https://mlsanity.com/paper/2311.12198/) (2023)
- [PhysDreamer: Physics-Based Interaction with 3D Objects via Video Generation](https://mlsanity.com/paper/2404.13026/) (2024)
- [PhysGen: Rigid-Body Physics-Grounded Image-to-Video Generation](https://mlsanity.com/paper/2409.18964/) (2024)
- [VideoPhy-2: A Challenging Action-Centric Physical Commonsense Evaluation in Video Generation](https://mlsanity.com/paper/2503.06800/) (2025)
- [DreamPhysics: Learning Physics-Based 3D Dynamics with Video Diffusion Priors](https://mlsanity.com/paper/2406.01476/) (2024)
- [Gaussian Splashing: Unified Particles for Versatile Motion Synthesis and Rendering](https://mlsanity.com/paper/2401.15318/) (2024)
- [Phys4DGen: Physics-Compliant 4D Generation with Multi-Material Composition Perception](https://mlsanity.com/paper/2411.16800/) (2024)
- [Integrating Reinforcement Learning with Visual Generative Models: Foundations and Advances](https://mlsanity.com/paper/2508.10316/) (2025)
- [Seeing What Matters: Visual Preference Policy Optimization for Visual Generation](https://mlsanity.com/paper/2511.18719/) (2025)
- [TeleBoost: A Systematic Alignment Framework for High-Fidelity, Controllable, and Robust Video Generation](https://mlsanity.com/paper/2602.07595/) (2026)
- [Learning What to Trust: Bayesian Prior-Guided Optimization for Visual Generation](https://mlsanity.com/paper/2511.18919/) (2025)
- [Reward-Aware Trajectory Shaping for Few-step Visual Generation](https://mlsanity.com/paper/2604.14910/) (2026)
- [Learning to Credit the Right Steps: Objective-aware Process Optimization for Visual Generation](https://mlsanity.com/paper/2604.19234/) (2026)
- [From World Action Models to Embodied Brains: A Roadmap for Open-World Physical Intelligence](https://mlsanity.com/paper/2607.11689/) (2026)

## Built on by (most cited first, 1 of 1)

- [Phase-aware video generation for physics-grounded dynamics and interactions](https://mlsanity.com/paper/2610.11791/) (2026)

## Abstract (arXiv)

Video diffusion models generate visually compelling content but routinely violate elementary physics when the subject involves fluids: liquid columns break apart in mid-air, container water levels fail to rise as liquid is poured in, and splashes disperse without regard to momentum or gravity. We attribute this gap to the fact that large-scale video-text corpora contain almost no explicit motion supervision, so models learn to imitate fluid appearance rather than dynamics. We address this with two contributions. First, we build a physics-simulation fluid dataset combining 1,638 MPM-simulated pouring/sloshing videos with 2,320 keyword-filtered real pouring videos mined from stock footage, plus two held-out test sets: a 1,515-video real-video benchmark and an 18-prompt text-to-first-frame generalization benchmark. Second, we introduce a dual-stream image-to-video architecture built on a pretrained diffusion-transformer video generator. It augments the standard RGB decoder with a lightweight Optical-Flow Decoder branch trained with explicit end-point-error and smoothness losses, fused into the RGB stream via zero-initialized convolutions so the pretrained backbone starts undisturbed. Only the two decoders are updated; the encoder, temporal transformer, and text encoder remain frozen. Across two model scales (1.3B and 14B) and two test sets, our method improves VideoPhy-2 Physical-Commonsense and Video-Quality scores over the frozen backbone by up to 8.75 and 4.65 points, outperforms a leading open competitor, and is preferred by human raters in a blind study. A direct optical-flow read-out evaluation further shows an end-point error as low as 0.54 pixels in-distribution, confirming the model has internalized a coherent motion prior rather than merely improving surface appearance.