Awesome Physical AIGC Lists

This repository provides a comprehensive investigation of advanced physical aware AIGC works.

Grounding Creativity in Physics: A Brief Survey of Physical Priors in AIGC

Siwei Meng ¹, Yawei Luo ², Ping Liu¹

¹Department of Computer Science, University of Nevada, Reno
²School of Software Technology, Zhejiang University

If you believe there are additional works that should be included in our list, please do not hesitate to send us an email (siweim@unr.edu, yaweiluo@zju.edu.cn, pino.pingliu@gmail.com) or raise an issue. Your suggestions and comments are invaluable to ensuring the completeness and accuracy of our resource.

News

• May 2025: Our work "MAGIC: Motion-Aware Generative Inference via Confidence-Guided LLM" [arXiv] is posted online.
• Apr 2025: Our work "Grounding Creativity in Physics: A Brief Survey of Physical Priors in AIGC" [arXiv] has been accepted by IJCAI 2025 Survey Track. The final version will be released soon.

Content

---

• Relevant Surveys
  • AIGC Surveys
  • Physical Aware AIGC Surveys
• Datasets
  • Synthetic Datasets
  • Real-world Datasets
• Benchmark
• Static 3D Generation
  • Single Image-based 3D Generation
  • Text-based 3D Generation
• Dynamic 3D Generation
  • Vision-based Dynamic 3D Generation
  • NeRF-based Dynamic 3D Generation
  • GS-based Dynamic 3D Generation
• 4D Generation
• Video Generation
  • T2V
  • I2V
  • Sim2Real
• Reasoning
• Others
  • Fluids

---

Relevant Surveys

AIGC Surveys

1. [arXiv 2023] A comprehensive survey of ai-generated content (AIGC): A history of generative ai from GAN to ChatGPT Paper
2. [arxiv 2024] Advances in 3D generation: A survey Paper
3. [arxiv 2024] A comprehensive survey on 3D content generation Paper
4. [arxiv 2025] Advances in 4D Generation: A Survey Paper

Physical Aware AIGC Surveys

1. [arXiv 2025] Exploring the Evolution of Physics Cognition in Video Generation: A Survey Paper
2. [ACM COMPUT SURV 2025] Physics-informed computer vision: A review and perspectives Paper
3. [arXiv 2025] Generative physical AI in vision: A survey Paper
4. [IJCAI 2025] Grounding Creativity in Physics: A Brief Survey of Physical Priors in AIGC Paper
5. [arXiv 2025] Exploring the Evolution of Physics Cognition in Video Generation: A Survey Paper
6. [arXiv 2025] Aligning Perception, Reasoning, Modeling and Interaction: A Survey on Physical AI Paper

---

Datasets

Synthetic Datasets

1. [arXiv 2025] Synthetic Video Enhances Physical Fidelity in Video Synthesis Paper Project
2. [PAC-NeRF] PAC-NeRF: Physics Augmented Continuum Neural Radiance Fields for Geometry-Agnostic System Identification Paper Project
3. [Spring-Gaus] Reconstruction and Simulation of Elastic Objects with Spring-Mass 3D Gaussians Paper Project

Real-world Datasets

1. [VIDEOPHY] VIDEOPHY: Evaluating Physical Commonsense for Video Generation Paper Project
2. [Physics 101] Physics 101: Learning Physical Object Properties from Unlabeled Videos Paper Project
3. [WISA-80K] WISA: World Simulator Assistant for Physics-Aware Text-to-Video Generation Paper Project

---

Benchmark

1. [PISA | ICML 2025] PISA Experiments: Exploring Physics Post-Training for Video Diffusion Models by Watching Stuff Drop Paper Project
2. [PhyGenBench | arXiv 2025] Towards World Simulator: Crafting Physical Commonsense-Based Benchmark For Video Generation Paper Project
3. [T2VPhysBench | arXiv 2025] T2VPhysBench: A First-Principles Benchmark for Physical Consistency in Text-to-Video Generation Paper
4. [Physics IQ Benchmark | ICCVW 2025] Physics IQ Benchmark: Do generative video models understand physical principles? Paper Project
5. [PhysUniBench | arXiv 2025] PhysUniBench: An Undergraduate-Level Physics Reasoning Benchmark for Multimodal Models Paper Project
6. [UGPhysics | ICML 2025] UGPhysics: A Comprehensive Benchmark for Undergraduate Physics Reasoning with Large Language Models Paper Code
7. [PhyBench | arXiv 2024] PhyBench: A Physical Commonsense Benchmark for Evaluating Text-to-Image Models Paper
8. [PhysBench | ICLR 2025] PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding Paper
9. [PhyWorldBench | arXiv 2025] PhyWorldBench: A Comprehensive Evaluation of Physical Realism in Text-to-Video Models Paper
10. [PhyDetEx | arXiv 2025] PhyDetEx: Detecting and Explaining the Physical Plausibility of T2V Models Paper Code Dataset
11. [ViRectify | arXiv 2025] ViRectify: A Challenging Benchmark for Video Reasoning Correction with Multimodal Large Language Models Paper
12. [PAI-Bench | arXiv 2025] PAI-Bench: A Comprehensive Benchmark For Physical AI Paper
13. [VideoPhy | arXiv 2024] VideoPhy: Evaluating Physical Commonsense for Video Generation Paper Code
14. [VideoPhy-2 | arXiv 2025] VideoPhy-2: A Challenging Action-Centric Physical Commonsense Evaluation in Video Generation Paper Code
15. [VideoScience-Bench | arXiv 2025] Benchmarking Scientific Understanding and Reasoning for Video Generation using VideoScience-Bench Paper
16. [IntPhys 2 | arXiv 2025] IntPhys 2: Benchmarking Intuitive Physics Understanding In Complex Synthetic Environments Paper
17. [PAI-Bench | arXiv 2025] PAI-Bench: A Comprehensive Benchmark For Physical AI Paper

---

Static 3D Generation

Single Image-based 3D Generation

1. [PhysComp | NeurIPS 2024] Physically Compatible 3D Object Modeling from a Single Image Paper
2. [NeurIPS 2024] PhyCAGE: Physically Plausible Compositional 3D Asset Generation from a Single Image Paper
3. [arXiv 2025] DSO: Aligning 3D Generators with Simulation Feedback for Physical Soundness Paper
4. [arXiv 2025] LLM-to-Phy3D: Physically Conform Online 3D Object Generation with LLMs Paper

Text-based 3D Generation

1. [arXiv 2024] Atlas3D: Physically Constrained Self-Supporting Text-to-3D for Simulation and Fabrication Paper
2. [arXiv 2025] PhiP-G: Physics-Guided Text-to-3D Compositional Scene Generation Paper
3. [arXiv 2025] LAYOUTDREAMER: Physics-guided Layout for Text-to-3D Compositional Scene Generation Paper
4. [ICCV 2025 Best Paper] Generating Physically Stable and Buildable Brick Structures from Text Paper Project

---

Dynamic 3D Generation

Vision-based Dynamic 3D Generation

1. [CVPR 2024] Generative Image Dynamics Paper Project
2. [NeurIPS 2024] MotionCraft: Physics-based Zero-Shot Video Generation Paper Project
3. [RAL 2023] Physics 101: Learning Physical Object Properties from Unlabeled Videos Paper Project
4. [ECCV 2024] PhysGen: Rigid-Body Physics-Grounded Image-to-Video Generation Paper Project
5. [CVPR 2025] PhyT2V: LLM-Guided Iterative Self-Refinement for Physics-Grounded Text-to-Video Generation Paper Project

NeRF-based Dynamic 3D Generation

1. [WACV 2024] ParticleNeRF: A Particle-Based Encoding for Online Neural Radiance Fields Paper Project
2. [ICLR 2023] PAC-NeRF: Physics Augmented Continuum Neural Radiance Fields for Geometry-Agnostic System Identification Paper Project
3. [CVPR 2024] Improving Physics-Augmented Continuum Neural Radiance Field-Based Geometry-Agnostic System Identification with Lagrangian Particle Optimization Paper
4. [CVPR 2024] PIE-NeRF: Physics-based Interactive Elastodynamics with NeRF Paper Project

GS-based Dynamic 3D Generation

1. [ECCV 2024] Reconstruction and Simulation of Elastic Objects with Spring-Mass 3D Gaussians Paper Project
2. [ECCV 2024] PhysDreamer: Physics-Based Interaction with 3D Objects via Video Generation Paper Project
3. [arXiv 2024] Physics3D: Learning Physical Properties of 3D Gaussians via Video Diffusion Paper Project
4. [arXiv 2024] PhysMotion: Physics-Grounded Dynamics From a Single Image Paper Project
5. [arXiv 2024] Gaussian Splashing: Unified Particles for Versatile Motion Synthesis and Rendering Paper
6. [arXiv 2024] Gaussian Splashing: Direct Volumetric Rendering Underwater Paper
7. [AAAI 2025] DreamPhysics: Learning Physics-Based 3D Dynamics with Video Diffusion Priors Paper Project
8. [CVPR 2024] PhysGaussian: Physics-Integrated 3D Gaussians for Generative Dynamics Paper Project
9. [ICLR 2025] OmniPhysGS: 3D Constitutive Gaussians for General Physics-Based Dynamics Generation Paper Project
10. [NeurIPS 2024] NeuMA: Neural Material Adaptor for Visual Grounding of Intrinsic Dynamics Paper Project
11. [NeurIPS 2024] Gaussian-Informed Continuum for Physical Property Identification and Simulation Paper Project
12. [arXiv 2025] MAGIC: Motion-Aware Generative Inference via Confidence-Guided LLM Paper
13. [arXiv 2025] ParticleGS: Particle-Based Dynamics Modeling of 3D Gaussians for Prior-free Motion Extrapolation Paper Project 14 [arXiv 2025] PIG: Physically-based Multi-Material Interaction with 3D Gaussians Paper

---

4D Generation

1. [arXiv 2024] Phy124: Fast Physics-Driven 4D Content Generation from a Single Image Paper
2. [arXiv 2024] GASP: Gaussian Splatting for Physic-Based Simulations Paper Project
3. [arXiv 2024] Phys4DGen: A Physics-Driven Framework for Controllable and Efficient 4D Content Generation from a Single Image Paper Project
4. [arXiv 2024] Trans4D: Realistic Geometry-Aware Transition for Compositional Text-to-4D Synthesis Paper
5. [arXiv 2024] Unleashing the Potential of Multi-modal Foundation Models and Video Diffusion for 4D Dynamic Physical Scene Simulation Paper
6. [ICLR 2025] Compositional 4D Dynamic Scenes Understanding with Physics Priors for Video Question Answering Paper
7. [arXiv 2025] Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding Paper

---

Video Generation

T2V

1. [arXiv 2025] ProPhy: Progressive Physical Alignment for Dynamic World Simulation Paper
2. [arXiv 2025] Think Before You Diffuse: LLMs-Guided Physics-Aware Video Generation Paper
3. [CVPR 2025] PhyT2V: LLM-Guided Iterative Self-Refinement for Physics-Grounded Text-to-Video Generation Paper
4. [arXiv 2025] RDPO: Real Data Preference Optimization for Physics Consistency Video Generation Paper
5. [arXiv 2025] VideoREPA: Learning Physics for Video Generation through Relational Alignment with Foundation Models Paper
6. [arXiv 2025] Is Sora a World Simulator? A Comprehensive Survey on General World Models and Beyond Paper
7. [arXiv 2025] Reasoning Physical Video Generation with Diffusion Timestep Tokens via Reinforcement Paper
8. [arXiv 2025] Enhancing Physical Plausibility in Video Generation by Reasoning the Implausibility Paper
9. [PhysHPO | arXiv 2025] Hierarchical Fine-grained Preference Optimization for Physically Plausible Video Generation Paper
10. [PhysCorr | arXiv 2025] PhysCorr: Dual-Reward DPO for Physics-Constrained Text-to-Video Generation with Automated Preference Selection Paper

I2V

1. [arXiv 2025] Physics-Grounded Motion Forecasting via Equation Discovery for Trajectory-Guided Image-to-Video Generation Paper
2. [arXiv 2025] MIND-V: Hierarchical Video Generation for Long-Horizon Robotic Manipulation with RL-based Physical Alignment Paper
3. [NeurIPS 2025] PhysCtrl: Generative Physics for Controllable and Physics-Grounded Video Generation Paper
4. [NeurIPS 2025] PhysMaster: Mastering Physical Representation for Video Generation via Reinforcement Learning Paper

Sim2Real

1. [arXiv 2025] Force Prompting: Video Generation Models Can Learn and Generalize Physics-based Control Signals Paper Project
2. [arXiv 2025] WISA: World Simulator Assistant for Physics-Aware Text-to-Video Generation Paper
3. [arXiv 2025] Synthetic Video Enhances Physical Fidelity in Video Synthesis Paper Project

---

Physical Reasoning in VLM/VDM

1. [arXiv 2025] Bridge Thinking and Acting: Unleashing Physical Potential of VLM with Generalizable Action Expert Paper
2. [arXiv 2025] LikePhys: Evaluating Intuitive Physics Understanding in Video Diffusion Models via Likelihood Preference Paper
3. [arXiv 2025] FineQuest: Adaptive Knowledge-Assisted Sports Video Understanding via Agent-of-Thoughts Reasoning Paper
4. [arXiv 2025] IPR-1: Interactive Physical Reasoner Paper
5. [arXiv 2025] PhyDetEx: Detecting and Explaining the Physical Plausibility of T2V Models Paper
6. [arXiv 2025] PhysVLM-AVR: Active Visual Reasoning for Multimodal Large Language Models in Physical Environments Paper
7. [arXiv 2025] PhyVLLM: Physics-Guided Video Language Model with Motion-Appearance Disentanglement Paper

---

Others

Fluids

1. [arXiv 2025] RainyGS: Efficient Rain Synthesis with Physically-Based Gaussian Splatting Paper
2. [arXiv 2025] Vid2Fluid: 3D Dynamic Fluid Assets from Single-View Videos with Generative Gaussian Splatting Paper
3. [CVPR 2025] FluidNexus: 3D Fluid Reconstruction and Prediction from a Single Video Paper

Audio Generation

1. [arXiv 2025] PAVAS: Physics-Aware Video-to-Audio Synthesis Paper