AI RESEARCH
LongLive-RAG: A General Retrieval-Augmented Framework for Long Video Generation
arXiv CS.CV
•
ArXi:2606.02553v1 Announce Type: new Autoregressive (AR) video diffusion enables variable-length synthesis, but long-horizon generation often suffers from accumulated errors and identity drift. For efficiency, existing methods commonly adopt sliding-window attention during generation. This creates an irreversible generation trajectory: once the active window accumulates appearance errors, subsequent generations can only condition on this degraded trajectory and drift further away.