Newest papers

State World Models: Execute the Rules, Generate the Pixels

Chan Yan Ki, Yanran Zhang, Wenzhao Zheng, Xiaofeng Wang, Zheng Zhu, Lei Chen, Jie Zhou, Jiwen Lu · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26811 · cited by 0 works (OpenAlex) · open-access PDF

Korean

현재의 세계 모델은 단일 이미지에서 액션 조건화된 비디오를 생성할 수 있습니다. 많은 사용자가 생성된 장면을 탐색할 수 있도록 카메라 움직임에 초점을 맞추고 있습니다. 그러나 행동이 현재 프레임에서 보이지 않는 상태에 따라 지속적인 규칙을 따르는 것을 보장하지는 않습니다. 우리는 Stateful Harness, 카메라 움직임을 지원하기 위해 비디오 렌더링에서 세계 진화를 분리하는 프레임을 제시합니다. 그것은 지정된 규칙에 따라 진화하는 세계 상태를 유지하고, 동일한 얼어 붙은 비디오 모델로 각각의 결과를 지정하고, 생성 된 클립을 사용하여 다음 회전을위한 세계 상태를 업데이트합니다. 각 회전에서, Harness는 사용자가 선택한 행동을 해결하고 결과

Machine translation (Meta M2M100); the English original follows.

Abstract

Existing world models can generate action-conditioned video from a single image. Many focus on camera movement, allowing users to explore the generated scene. However, navigation alone does not ensure that an action follows persistent rules when its outcome depends on state that is invisible in the current frame. We present Stateful Harness, a framework that separates world evolution from video rendering to support stateful and rule-governed interaction. It maintains an evolving world state under specified rules, renders each resulting event with a frozen video model, and uses the generated clip to update the world state for the next turn. At each turn, the harness resolves the action selected by the user and compiles the resulting event into an instruction, while a preview clip from an editable 3D scene provides spatial and camera guidance. To evaluate rule compliance, we build WorldState-Bench, a benchmark of 50 counterfactual twin pairs that share an initial frame and action but differ in hidden state and required outcome. Under rule-only prompting, the best of the 11 evaluated video backends produces the required endings for both cases in 32% of pairs. For MiniMax-H3, specifying the required endings raises this rate from 32% to 64%, while the full harness reaches 92% with the same frozen model. These results support separating rule execution from video rendering for world interactions with hidden state.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Rendering (computer graphics), Stateful firewall, Counterfactual thinking, Artificial intelligence, Event (particle physics)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange