AI Paper Atlas

Research map / Video / Video generation

PhysRVG: Physics-Aware Unified Reinforcement Learning for Video Generative Models

Qiyuan Zhang, Biao Gong, Shuai Tan, Zheng Zhang, Xing Zhu, Yujun Shen, Yuyuan Li, kelu Yao, Chunhua Shen, Changqing Zou

ECCV 2026 · Video › Video generation · approach: World-model video generation · 21 citations (OpenAlex)

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 changedWhereMeasuredEffect
add MDcycle RL post-training with physics-grounded rewardWan2.2 5B · PhysRVGBenchIoU / TO: IoU 0.15 -> 0.64; TO 162.78 -> 15.03Improved
add plain RL after full fine-tuning (ablation)FT model · PhysRVGBenchTO: 46.27 -> 17.25Improved
replace plain RL by MDcycle (add flow-matching mimicry branch) (ablation)FT+RL · PhysRVGBenchTO: 17.25 -> 15.03Improved
use full-parameter fine-tuning vs LoRA only (ablation)Baseline · PhysRVGBenchIoU: LoRA 0.41 vs FT 0.38No clear effect
add collision-aware weights (1,2,3) vs uniform weights (ablation)reward · PhysRVGBenchTO: 15.41 -> 15.03Improved
raise collision weight to 5 (ablation)reward · PhysRVGBenchTO: 15.03 -> 16.17Hurt
restrict SDE sampling to high-noise window (75-100%) with 2 steps (ablation)RL sampling · PhysRVGBenchTO: 27.24 -> 15.03Improved
set mimicry Threshold to 8 vs 4 or 12 (ablation)MDcycle · PhysRVGBenchTO: 22.39 (T=4), 16.97 (T=12) -> 15.03Improved
use 16k vs 6k FT steps before MDcycle (ablation)MDcycle · PhysRVGBenchTO: 37.23 -> 15.03Improved
use PEFT instead of full-parameter training inside RL (ablation)MDcycle · training stabilityreward curveImproved
use PhysRVG vs Kling2.5 on VBench total scorevideo generator · VBenchTotal Score: 78.77 -> 78.89No clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy