VLN-Cache: 視覚・意味ダイナミクスを考慮したVLNモデルのトークンキャッシング
VLN-Cache: Enabling Token Caching for VLN Models with Visual/Semantic Dynamics Awareness
VLNタスクで大規模視覚言語モデルの推論コストを削減するため、視点移動によるトークン位置ずれとタスク進行に伴う意味的変化を考慮したトークンキャッシング手法を提案した。
詳しい要約
1. どんなもの?
2. 先行研究と比べてどこがすごい?
3. 技術・手法の肝は?
4. どうやって有効だと検証した?
5. 議論はある?
6. 次に読むべき論文は?
※ AIが要旨から生成した要約です。正確性は原文をご確認ください。
著者: Zihao Zheng, Zhihao Mao, Xingyue Zhou, Jiayu Chen, Maoliang Li, Xinhao Sun, Hailong Zou, Zhaobo Zhang, Xuanzhe Liu, Donggang Cao, Hong Mei, Xiang Chen
分類: cs.RO, cs.LG
原文アブストラクト
Vision-and-Language Navigation (VLN) increasingly relies on large vision-language models, but their inference cost conflicts with real-time deployment. Token caching is a promising training-free strategy that avoids redundant computation by reusing stable visual tokens across frames. However, existing methods assume a static camera and fixed semantic focus, assumptions that VLN fundamentally violates. We identify two failure modes: (1) visual dynamics, where viewpoint shift displaces token positions across frames, causing position-wise matching to pair misaligned content; (2) semantic dynamics, where token relevance shifts across task stages as navigation progresses, making cached states stale. We propose VLN-Cache, a visual-dynamic-aware and semantic-dynamic-aware caching framework that introduces view-aligned remapping to recover geometric correspondences and a task-relevance saliency filter to veto reuse at semantic transitions. A layer-adaptive entropy policy further balances the per-layer reuse budget. Experiments on the R2R-CE simulation benchmark show up to 1.52x speedup while maintaining competitive navigation success rates.
