ATAS：以世界状态为中心的智能体视频生成框架

AI 摘要

研究团队提出 Any Time, Any Space（ATAS）框架，用于长视频的智能体生成。该框架将剧本编译为带有明确位置、相机和事件时间引用的镜头请求，并借助场景记忆（静态几何）与事件记忆（按场景、事件、时间索引的动态观测）在生成前后检索与注册证据，从而保证非连续请求间的空间与时间一致性。在四种视频生成器上，ATAS 提升了跨镜头一致性指标并取得最高综合基准分，在空间重访、事件重捕获和人类评分上也优于对比的智能体与 3D 引导方法。

正文节选

Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generation Abstract Long-form agentic video generation must turn a screenplay into shots that remain spatially and temporally consistent across nonconsecutive requests. Passing prompts or recent frames between shots does not provide a shared spatial frame for revisiting a scene or a time-indexed record for replaying an earlier event. We present Any Time, Any Space (ATAS), which compiles a screenplay into shot requests with ex

发布时间：—

抓取时间：2026-10-15 11:16

来源机构：Hugging Face