EvoWorld, 세계 상태 추론으로 동영상 생성 모델의 인과 관계 개선
EvoWorld: Towards Out-of-Sight World Evolution in Generative Video Models
- 매체
- Hugging Face
- 게시
- 중요도
- 2/5
- 화제 지수
- 11
한국어 요약
생성 동영상 모델이 시각적 사실성만 추구하고 세계 상태 변화를 추론하지 못하는 문제를 다룬다. 비전-언어 모델이 '세계 뇌' 역할을 하며 생성된 장면을 관찰하고 상태 변화를 추론하면, 동영상 렌더러가 이를 시각화하는 협력 프레임워크 EvoWorld를 제시했다. 새로운 평가 과제 Out-of-Sight Causal Evolution을 제안하고, EvoWorld-Data와 EvoWorld-Bench 데이터셋·벤치마크를 소개했다. 사건 발생 인과 관계와 일관성에서 기존 모델을 크게 개선했다.
원문 발췌 · English
Generative video models are increasingly used as visual world simulators, yet their primary strength lies in rendering plausible pixels rather than reasoning about how a world should evolve. Coherent world evolution requires more than visual fidelity: when an event occurs, a model must infer its consequences for the entities in the world and then render the resulting state. For example, after a wild beast appears, the model should not only depict the wild beast, but also infer whether a nearby person should flee, take cover, or fight back based on the realized scene. Existing diffusion-based video models lack an explicit mechanism for maintaining world state and reasoning over such event-conditioned consequences, often producing visually plausible but behaviorally static or causally inconsistent outputs. To make this reasoning gap concrete and measurable, we propose a new problem, Out-of-Sight Causal Evolution, which requires a model to render an event, reason about its consequences for the evolving world state, and reveal a causally plausible outcome as the world continues to evolve, even out of sight. As a baseline reference solution, we introduce EvoWorld, a brain-renderer collaboration framework in which a vision-language model serves as the world brain: it observes generated segments, maintains an evolving account of the world, and reasons about plausible state transitions, while a segment-autoregressive video renderer realizes them visually. We further introduce EvoWorld-Data and EvoWorld-Bench for training and evaluating event realization, causal reactions, and revisit consistency. Experiments show that EvoWorld substantially improves event-conditioned causal evolution over recent models while preserving visual quality and consistency, establishing a reference point for reasoning-guided, autonomous world evolution beyond the current field of view.