Research map / Video / Video generation
PhysVid: Physics Aware Local Conditioning for Generative Video Models
Saurabh Pathak, Elahe Arani, Mykola Pechenizkiy, Bahram Zonooz
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add chunk-aware local physics conditioning to Wan-1.3B (ablation) | Wan2.1-1.3B · VideoPhy | physical commonsense (PC): 0.2401 -> 0.3169 | Improved |
| add chunk-aware local physics conditioning over plain finetuning (ablation) | Wan2.1-1.3B · VideoPhy | physical commonsense (PC): 0.2866 -> 0.3169 | Improved |
| add chunk-aware local physics conditioning over plain finetuning (ablation) | Wan2.1-1.3B · VideoPhy2 | physical commonsense (PC): 0.6261 -> 0.6411 | Improved |
| add local counterfactual (negative physics) guidance at inference (ablation) | PhysVid · VideoPhy | physical commonsense (PC): 0.2924 -> 0.3169 | Improved |
| add local counterfactual guidance at inference (ablation) | PhysVid · VideoPhy2 | physical commonsense (PC): 0.6334 -> 0.6411 | Improved |
| PhysVid local conditioning vs Wan-1.3B | Wan2.1-1.3B · VideoPhy | semantic alignment (SA): 0.46 -> 0.43 | Hurt |
| PhysVid 1.7B vs 14B model | Wan2.1 · VideoPhy2 | physical commonsense (PC): 0.59 -> 0.64 | Improved |
| plain finetuning on WISA data (ablation) | Wan2.1-1.3B · VideoPhy | physical commonsense (PC): 0.2401 -> 0.2866 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 2 on most shared results.
- WISA — WISA: World simulator assistant for physics-aware text-to-video generation (NeurIPS 2025) · better on 0 of 1 shared result
- VideoREPA-5B — VideoREPA: Learning Physics for Video Generation through Relational Alignment with Foundation Models (NeurIPS 2025) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PhysCtrl: Generative Physics for Controllable and Physics-Grounded Video Generation (NeurIPS 2025 · similarity 0.93)
- PhysPO: Physics-Aware Local Preference Optimization for Physically Consistent Video Diffusion (ECCV 2026 · similarity 0.93)
- PhyT2V: LLM-Guided Iterative Self-Refinement for Physics-Grounded Text-to-Video Generation (CVPR 2025 · similarity 0.90)
- Physics-Informed Video Generation via Mixture-of-Experts Latent Alignment (NeurIPS 2026 · similarity 0.90)
- Hierarchical Fine-grained Preference Optimization for Physically Plausible Video Generation (NeurIPS 2025 · similarity 0.89)
- VideoREPA: Learning Physics for Video Generation through Relational Alignment with Foundation Models (NeurIPS 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
