arXiv AI

From Scores to Samples: Elastic Forcing for Autoregressive Video Generation

arXiv Computer Vision
Oct 9

Connected Self Forcing: Beyond Local Learning in Video Autoregression

The paper introduces Connected Self Forcing, a training framework that reconnects gradient paths across autoregressive video chunks to improve long‑horizon visual quality and temporal consistency. It allows feedback from later predictions to influence earlier context generation, using shortcut gradient replay to keep memory usage manageable. The method is combined with distribution‑matching distillation and shows experimental gains without altering the inference procedure.

By Dongbin Zhang, Chaoda Zheng, Kangjie Chen, Xiangyu Li, Shijia Chen, Jinhao Deng, Yuqi Zhang, Guangfeng Jiang, Hongbin Lin, Choo Sin Wai, Minqi Wang, Puyi Wang, Jingye Zhang, Yu Zhang, Xianming Liu, Boyang Wang
arXiv Computer Vision
Sep 25

ViRDM: Taming Representation Distribution Matching for Few-Step Causal Video Generation

ViRDM is a new post‑training method for few‑step causal video generation that eliminates the need for a large teacher model and an online critic. By applying representation distribution matching (RDM) with a precomputed target distribution, a lightweight VAE decoder, and staged vector–Jacobian products, ViRDM overcomes memory, optimization, and temporal dynamics challenges. The approach reduces GPU memory usage and training time, achieving state‑of‑the‑art VBench performance with only 20 generator updates and 16 A100 GPU‑hours.

By Zichong Meng, Chongjian Ge, Chun-Hao P. Huang, Yang Zhou, Huaizu Jiang
arXiv Computer Vision
Nov 9

Trajectory Forcing: Exploiting Diffusion Trajectories for Autoregressive Long Video Generation

The paper introduces Trajectory Forcing (TraF), an autoregressive diffusion method that uses the denoising trajectories of previously generated video chunks as context for generating longer videos. TraF is trained in two stages: first with ground‑truth videos using pseudo‑trajectories, then with a teacher model that supplies real trajectories from the student’s own rollout. On the 5‑second VBench benchmark, TraF outperforms Self Forcing and Causal Forcing, and it achieves the highest overall quality for 30‑second video generation, extending the training horizon sixfold.

By Bohan Wang, Shuo Chen, Yunzhi Li, Zhaozheng Chen, Junzhe Zhang, Hanwang Zhang