视频生成遇上智能体执行框架：智能体化视频生成、世界建模与递归自我改进
原标题：Paper page - Video Generation Meets Agent Harness: Agentic Video Generation, Agentic World Modeling and Recursive Self-Improvement
AI 摘要
该论文提出以 agent harness（智能体执行运行时）为视角，研究视频生成与世界模型的交汇。报告从三个方向展开：智能体化视频生成（协调脚本、分镜、素材、制作与剪辑）、智能体化世界建模（由 harness 中介观测、控制与状态，游戏引擎或智能体编写的程序维护世界状态、视频模型渲染观测），以及视频模型开发中的递归自我改进（区分输出修订、harness 状态更新、生成器参数与评估器变更）。论文还分析了视听感知、长程规划、记忆、代码交互、反馈引导控制、技能复用与多智能体协作等关键能力，并指出自主视频生产、终身记忆、可编程混合控制、自主模型开发、溯源、成本核算与评估等开放挑战。
正文节选
Video Generation Meets Agent Harness: Agentic Video Generation, Agentic World Modeling and Recursive Self-Improvement Abstract Agent harnesses integrate video generation and world models to enable planning, tool use, and recursive self-improvement for autonomous multi-shot video production. Video generation is shifting from end-to-end synthesis toward agentic generation, where planning, tool use, and feedback enable coherent, controllable multi-shot videos. In parallel, video world models provid