WorldToken：面向機器人模仿學習的時間優先序列建模

摘要

機器人策略在每個決策步驟接收異質觀測，但序列模型在如何隨時間組織這些輸入方面有所不同。我們提出WorldToken，一種時間優先的策略實例化，在每個策略時間步內將多視角影像、本體感覺與任務條件融合為單一世界令牌。因果時序Transformer對產生的世界令牌序列進行建模，擴散動作頭生成動作區塊。在23個RoboCasa任務上，一個除凍結的預訓練CLIP文字編碼器外從零訓練的85.3M參數策略，使用每個任務2,900條生成的示範，達到59.45%的平均閉環成功率。對五種數據集規模、五種模型規模與兩種訓練種子進行的完整全因子掃描顯示，額外目標域數據帶來一致的增益，而超過中等模型規模後則出現邊際效益遞減。在相同檢查點歷史截斷下，將可見歷史減少至一或兩個策略時間步會降低全部50個RoboCasa策略的閉環成功率。在RMBench Blocks Ranking上，將可見歷史從146秒減少至8秒會使評估器成功率從95%降至28%，而探索性擴展推演則可維持參考交換序列超過850秒。這些結果確立了完整WorldToken實例化的實證可行性，並刻畫了其在所測試配置下的數據擴展與時間上下文行為。它們並未確立相對替代序列組織的優越性，也未隔離完整實作中哪些組件驅動了所觀測到的性能。

English

Robot policies receive heterogeneous observations at each decision step, yet sequence models differ in how they organize these inputs over time. We introduce WorldToken, a time-first policy instantiation that fuses multiview images, proprioception, and task conditioning within each policy timestep into one world token. A causal temporal Transformer models the resulting world-token sequence, and a diffusion action head generates action chunks. On 23 RoboCasa tasks, an 85.3M-parameter policy trained from scratch apart from a frozen pretrained CLIP text encoder achieves 59.45% mean closed-loop success using 2,900 generated demonstrations per task. A complete factorial sweep over five dataset sizes, five model sizes, and two training seeds shows consistent gains from additional target-domain data and diminishing returns beyond moderate model size. Under same-checkpoint history truncation, reducing visible history to one or two policy timesteps lowers closed-loop success for all 50 RoboCasa policies. On RMBench Blocks Ranking, reducing visible history from 146 to 8 seconds lowers evaluator success from 95% to 28%, while an exploratory extended rollout sustains the reference swap sequence for over 850 seconds. These results establish the empirical feasibility of the complete WorldToken instantiation and characterize its data-scaling and temporal-context behavior under the tested recipes. They do not establish superiority over alternative sequence organizations or isolate which components of the complete implementation drive the observed performance.