學習從行為到世界建模的可遷移動態先驗

摘要

我們研究以動作條件化的世界模型作為一種可擴展的方式，以學習具遷移性的動態先驗知識，應用於機器人學習。透過預訓練一個模型來預測動作如何驅動視覺場景的演變，所產生的世界模型能夠捕捉超越外觀層次影片生成的可重複使用互動動態。具體而言，我們在大規模的機器人操作資料上，使用真實動作標註，預訓練了一個多視角互動基礎擴散世界模型 A2World。我們從兩個互補的角度驗證所學得的動態先驗知識。首先，我們將 A2World 調整為任務或場景特定的真實世界模擬器 A2World-sim，其長時程推演支援基於模擬器的策略評估，並透過以世界模型推演取代真實機器人推演來進行可擴展的假設分析。其次，從相同的預訓練權重出發，我們將 A2World 調整為影片與動作聯合預測模型 A2World-policy，該模型在視覺與指令條件下預測動作。在模擬基準測試與真實機器人設置上的實驗證明，基於動作條件化的世界模型預訓練能夠產生具遷移性的動態先驗知識，從而有益於以模擬器為中心與以策略為中心的機器人學習。

English

We study action-conditioned world modeling as a scalable way to learn transferable dynamics priors for robot learning. By pretraining a model to predict how actions drive visual scene evolution, the resulting world model captures reusable interaction dynamics beyond appearance-level video generation. Concretely, we pretrain a multi-view interactive base diffusion world model, A2World, on large-scale robot manipulation data with real action annotations. We validate the learned dynamics priors from two complementary perspectives. First, we adapt A2World into a task- or scene-specialized real-world simulator, A2World-sim, whose long-horizon rollouts support simulator-based policy evaluation and scalable what-if analysis by replacing real-robot rollouts with world model rollouts. Second, starting from the same pretrained weights, we adapt A2World into a video-action joint prediction model, A2World-policy, that predicts actions under visual and instruction conditioning. Experiments across simulation benchmarks and real-robot settings demonstrate that action-conditioned world model pretraining yields transferable dynamics priors that benefit both simulator-centric and policy-centric robot learning.