PulseAugur
实时 09:47:33
English(EN) VibeWorlding: Can Multimodal Agents Construct 3D Open Worlds End-to-End?

新框架VibeWorlding使AI能够从文本构建3D世界

研究人员开发了VibeWorlding,这是一个用于训练和评估多模态智能体的框架,这些智能体能够根据文本提示构建3D开放世界。该系统包括一个基准数据集(VWE-BENCH)和一个强化学习框架(VibeWorlding-Gym)。实验表明,包括GPT-5.5和Qwen3.8-Max在内的当前大型语言模型在此任务上面临挑战,成功率低于60%。然而,强化学习训练显著改进了开源模型,其中VibeWorlder-30B-A3B的表现甚至优于前沿模型。 AI

影响 这项研究推进了3D内容创作中的智能体能力,可能对游戏开发和虚拟现实等领域产生影响。

排序理由 该集群描述了一篇研究论文,该论文介绍了一个用于3D世界生成中多模态智能体的新框架和基准。

在 Hugging Face Daily Papers 阅读 →

AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →

新框架VibeWorlding使AI能够从文本构建3D世界

报道来源 [4]

  1. arXiv cs.AI TIER_1 English(EN) · Yansong Ning, Jingwen Ye, Zhongkai Wu, Yang Sun, Yiqin Zhu, Xingyi Li, Weidong Zhang, Hao Liu ·

    VibeWorlding:多模态代理能否端到端构建三维开放世界?

    arXiv:2608.15265v1 Announce Type: new Abstract: Constructing an interactive 3D open world from a user query is important. However, existing methods are primarily evaluated on idealized, simple queries, making it difficult to systematically analyze and compare how multimodal agent…

  2. Hugging Face Daily Papers TIER_1 English(EN) ·

    VibeWorlding:多模态智能体能否端到端构建三维开放世界?

    A unified framework benchmarks and trains multimodal agents that infer intent, plan 3D scenes, invoke tools, and reflect on feedback, revealing that reinforcement learning improves open-source models beyond closed-source frontiers.

  3. arXiv cs.CV TIER_1 English(EN) · Shengyuan Wang, Zhiheng Zheng, Yu Shang, Lixuan He, Yangcheng Yu, Fan Hangyu, Jie Feng, Qingmin Liao, Yong Li ·

    UrbanWorld2.0:一个用于城市规模现实对齐3D世界生成的多模态代理框架

    arXiv:2511.18005v2 Announce Type: replace Abstract: The automated generation of high-fidelity, city-scale 3D environments remains a formidable challenge with profound academic and industrial implications. However, existing methods struggle to achieve the necessary quality, fideli…

  4. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    多模态代理正在从头开始构建完整的 3D 世界。VibeWorlding 将文本提示端到端地转化为开放世界——在 Hugging Face 上获得 20 票。该

    Multimodal agents are building entire 3D worlds from scratch. VibeWorlding turned text prompts into open worlds—end-to-end, with 20 upvotes on Hugging Face. The paper: https:// huggingface.co/papers/2608.152 65 # AI # MachineLearning # Research