研究人员开发了WorldWeaver (W^2)，一种新颖的流式多智能体视频扩散模型，旨在提高多智能体交互世界建模的一致性。与使用观察历史作为条件上下文的先前管道不同，W^2 采用了跨智能体世界状态寄存器。这些寄存器存储共享的世界信息并跟踪个体智能体的状态，在每个生成的块之后动态更新。在双智能体Minecraft视频生成中的实验表明，这种显式世界状态建模增强了逻辑一致性和整体生成质量。 AI
影响 这项研究可能带来在多智能体环境中更连贯、逻辑上更一致的AI生成视频。
排序理由 该集群描述了一篇详细介绍新颖模型架构及其实验验证的新研究论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
Multi-agent interactive world models should not only generate consistent observations, but also maintain world states that persist across agents and evolve across views. Existing autoregressive video diffusion pipelines carry forward observation history as conditioning context, w…
Multi-agent interactive world models should not only generate consistent observations, but also maintain world states that persist across agents and evolve across views. Existing autoregressive video diffusion pipelines carry forward observation history as conditioning context, w…
arXiv:2607.21594v1 Announce Type: new Abstract: Multi-agent interactive world models should not only generate consistent observations, but also maintain world states that persist across agents and evolve across views. Existing autoregressive video diffusion pipelines carry forwar…