WORLDWIDE_SanGuai / Research bulletin

[ worldwide.arakne.de ]

世界很大,讀一小點。

SanGuai 的三日研究佈告板。每天更新,讀完就走。

EST. 202672H上架 72 小時
人工智慧 / 2611.16217

WorldDraft:若程式碼構築世界,影片模型能否只負責渲染?

WorldDraft: If Code Builds the World, Can Video Models Just Render It?

作者: Zhiyu Hou, Wen-Jie Shu, Harold Haodong Chen, Weilin Ruan, Chenfei Liao, Liuzhou Zhang, Disen Lan, Xiaogang Xu et al.論文發布: 2026-11-12離架倒數: 62 小時

研究速讀

這篇論文提出 WorldDraft 基準,將 4D 世界建構與視覺生成分工,把影片模型當成生成式渲染器,並以 432 組配對測試案例檢驗世界保存與視覺實現兩項能力。作者稱 12 個模型的評測顯示兩目標存在張力,且提出 WorldDraft-Primer 提示策略改善表現。

值得留意

  1. 論文主張以多模態程式代理建構 4D 世界,讓影片模型專注渲染,並提出世界保存與視覺實現兩項評估準則。
  2. WorldDraft 含 432 組嚴格配對測試案例,涵蓋 48 個可執行 4D 世界家族與六類動態挑戰。
  3. 作者稱 12 個代表性模型的評測顯示世界保存與視覺實現互相拉扯,並提出 WorldDraft-Primer 視覺提示策略。

閱讀限制

僅依標題與摘要整理,未讀全文,無法核實基準設計、實驗數字與模型選擇是否如摘要所述。

原始摘要 [EN]

Multimodal coding agents decouple 4D world construction from visual generation, repurposing video models from world generators toward generative renderers. This division of labor raises a fundamental question: if code builds the world, can video models just render it? A reliable generative renderer hinges on two complementary criteria: (i) world preservation, retaining prescribed 4D kinematics and camera paths; and (ii) visual realization, persistently binding text-specified target attributes over time. To evaluate this capability, we introduce WorldDraft, a benchmark of 432 strictly paired test cases across 48 executable 4D world families spanning six dynamic challenge categories. Under a 3 times 3 factorial design, WorldDraft orthogonally crosses draft granularity (abstract proxy to previs) and appearance alignment (neutral to conflicting) under identical dynamics. Extensive evaluations across 12 representative models reveal that: (I) world preservation and visual realization remain in stark operational tension; (II) richer draft scaffolding anchors structure but paradoxically impairs appearance fidelity; and (III) conflicting source cues trigger severe source leakage (up to 64%), exposing fragile cross-modal disentanglement. To mitigate these bottlenecks, we propose WorldDraft-Primer, a plug-and-play visual prompting strategy that consistently improves both objectives across frozen video renderers. Together, WorldDraft and WorldDraft-Primer lay a foundation for bridging programmatic 4D simulation with video generation, advancing physically grounded generative rendering.

AI 摘要僅根據公開標題與摘要整理;研究結論來自作者,請以原文為準。