CondenseVLA:効率的なマルチフレームVLAのための学習可能な履歴圧縮
CondenseVLA: Learnable History Condensation for Efficient Multi-Frame VLA
VLMのフレーム特徴を学習可能なクエリで固定数のスロットに圧縮し、冗長性を抑えつつ短期視覚文脈を行動デコードに活用するマルチフレームVLAを提案。Simpler-WidowXで58.7%→69.8%、LIBEROで97.2%を達成。
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: Feiyang Hong, Yujie Wei, Bo Zhao, Xiu Su, Hongxun Yao, Shuo Yang
分類: cs.CV, cs.RO
原文アブストラクト
Robot manipulation often depends on short-term cues - such as brief contacts, temporary occlusions, and intermediate object states - that are ambiguous in a single frame and can only be resolved by integrating observations over time. However, injecting historical information solely by enlarging the history window is computationally costly and frequently yields limited gains due to substantial redundancy and task-agnostic, inflexible sampling of observations. We propose CondenseVLA, a multi-frame VLA policy with a learnable decoder-side history condensation module for using short-term visual context in action decoding. Given a sequence of cached per-frame features from the VLM backbone, CondenseVLA distills observation history into a fixed set of slot tokens using learnable queries and multi-head cross-attention. Simple auxiliary regularization penalizes overlap among slot attention maps and decorrelates slot embeddings, encouraging complementary attention and reducing slot redundancy. Across the evaluated tasks, CondenseVLA improves Simpler-WidowX success from 58.7% to 69.8% (+11.1 percentage points), achieves 97.2% on LIBERO, and reaches 52.8% on real-world tasks. It achieves 21.6% success on memory-dependent RMBench tasks. These results support selective condensation of bounded recent visual evidence for action decoding, while harder multi-observation tasks expose the limits of a finite history window. Project page: https://github.com/FeiyangHong/CondenseVLA