從生成到模擬：世界模型距離真正的模擬器還有多遠？

摘要

隨著擴散模型與大規模影片生成的快速進展，生成式世界模型日益被期待能取代傳統模擬器，包括物理引擎、遊戲引擎與強化學習環境。然而，從生成到模擬之間的距離仍缺乏系統性的評估。我們提出一項基於能力的研究，採用外部衡量基準：傳統模擬器的八項能力，即資產建構、物理引擎、互動、可控性、穩定性、狀態回饋、多樣性與評估指標。我們追溯三條主要技術路線——潛在動態、影片生成與聯合嵌入預測——並將2018年至2026年6月間發表的200篇具代表性研究精確對應至上述各項能力。我們的分析顯示，世界模型在特定場景的互動與可控性方面已實現功能性取代，但在物理定律的形式化保證、結構化狀態回饋，以及可重現的長時程演化方面，仍與傳統模擬器存在差距。狀態回饋是跨路線中最受忽視的弱點：163篇實作論文中僅有6篇提供用於查詢實體狀態或物理參數的執行期介面。我們提出六個研究方向：形式化物理、統一動作介面、第一級狀態回饋、長時程穩定性、下游效用評估，以及跨路線混合。專案頁面：https://github.com/AtongWang/world-model-simulators

English

With the rapid progress of diffusion models and large-scale video generation, generative world models are increasingly expected to replace traditional simulators, including physics engines, game engines, and reinforcement-learning environments. Yet the remaining distance from generation to simulation lacks a systematic assessment. We present a capability-based study using an external yardstick: eight capabilities of a traditional simulator, namely asset construction, physics engine, interaction, controllability, stability, state feedback, diversity, and evaluation metrics. We trace three main technical routes--latent dynamics, video generation, and joint-embedding prediction--and map exactly 200 representative works published from 2018 to June 2026 onto these capabilities. Our analysis shows that world models have achieved functional substitution in interaction and controllability for specific scenarios, but remain short of traditional simulators in formal guarantees of physical laws, structured state feedback, and reproducible long-horizon evolution. State feedback is the most neglected cross-route shortcoming: only 6 of 163 implementation papers expose a runtime interface for querying entity states or physical parameters. We identify six research directions: formalized physics, a unified action interface, first-class state feedback, long-horizon stability, downstream-utility evaluation, and cross-route hybridization. Project page: https://github.com/AtongWang/world-model-simulators