適用於影片世界模型的可定址記憶體

摘要

我們研究互動式影片世界模型中的視覺持久性。這類模型依賴鍵值（Key-Value, KV）快取作為持續成長的視覺記憶，以延續先前生成畫面所攜帶的資訊。然而，我們發現一旦推論軌跡（rollout）超出訓練範圍，模型便無法可靠地定址已儲存的內容；原因是時間性旋轉位置編碼（Rotary Positional Embeddings, RoPE）的偏移量會落在訓練期間未見過的區間，使模型難以透過注意力機制檢索相關的視覺資訊。此外，若直接在 RoPE 旋轉後的空間中樸素地壓縮快取，會因將彼此不相容的位置相位平均在一起而破壞記憶。為了解決此問題，我們提出 WorldTrace，一個無需重新訓練的記憶框架，專為長時程視覺持久性而設計。WorldTrace 為每個摘要槽位指派一個獨特且位於分布內的虛擬位置，使壓縮後的記憶仍可被定址。在此可定址快取中，我們研究了兩種記憶壓縮方法：WorldTrace-Field 壓縮歷史以達成時間一致性，而 WorldTrace-Landmark 則在偵測到的場景轉換處儲存原樣場景軌跡，以支援情節回憶。我們進一步提出 LoopBench 基準，用以評估壓縮快取能否在長時間繞行後重建先前造訪過的場景。在 LoopBench 上，WorldTrace-Field 將時間一致性提升了 +15.5%，WorldTrace-Landmark 則將情節回憶提升了 +19.5%，從而在無需重新訓練的情況下延伸視覺持久生成的能力。

English

We study visual persistence in interactive video world models. These models rely on a Key-Value (KV) cache as a growing visual memory to carry forward previously generated frames. However, we find that models can no longer reliably address stored content once rollouts extend beyond the training horizon, because temporal Rotary Positional Embeddings (RoPE) offsets then fall outside the range seen during training and the model struggles to retrieve the relevant visual information through attention. Moreover, naively compressing the cache in the RoPE-rotated space corrupts memory by averaging together incompatible positional phases. To address this, we propose WorldTrace, a training-free memory framework for long-horizon visual persistence. WorldTrace keeps compressed memory addressable by assigning each summary slot a distinct, in-distribution virtual position. Within this addressable cache, we study two memory compression approaches: WorldTrace-Field compresses history for temporal coherence, while WorldTrace-Landmark stores verbatim scene traces at detected transitions for episodic recall. We further introduce LoopBench, a benchmark evaluating whether a compressed cache can reconstruct a previously visited scene after a long detour. WorldTrace-Field improves temporal consistency by +15.5%, and WorldTrace-Landmark improves episodic recall by +19.5% on LoopBench, extending visually persistent generation without retraining.