TY  - RPRT
TI  - ActionCache: Training-Free Acceleration for Vision-Language-Action Models with Action Caching and Refinement
AU  - Ryuji Oi
AU  - Hikari Otsuka
AU  - Kosuke Matsushima
AU  - Yuki Ichikawa
AU  - Masato Motomura
AU  - Tatsuya Kaneko
AU  - Daichi Fujiki
PY  - 2026
UR  - https://arxiv.org/abs/2607.06370
ID  - 2607.06370
ER  -