AI RESEARCH
Pathwise Test-Time Correction for Autoregressive Long Video Generation
arXiv CS.CV
•
ArXi:2602.05871v2 Announce Type: replace Distilled autoregressive diffusion models facilitate real-time short video synthesis but suffer from severe error accumulation during long-sequence generation. While existing Test-Time Optimization (TTO) methods prove effective for images or short clips, we identify that they fail to mitigate drift in extended sequences due to unstable reward landscapes and the hypersensitivity of distilled parameters. To overcome these limitations, we