TY  - RPRT
TI  - HeiSD: Hybrid Speculative Decoding for Embodied Vision-Language-Action Models with Kinematic Awareness
AU  - Zihao Zheng
AU  - Zhihao Mao
AU  - Sicheng Tian
AU  - Maoliang Li
AU  - Jiayu Chen
AU  - Xinhao Sun
AU  - Zhaobo Zhang
AU  - Xuanzhe Liu
AU  - Donggang Cao
AU  - Hong Mei
AU  - Xiang Chen
PY  - 2026
UR  - https://arxiv.org/abs/2603.17573
ID  - 2603.17573
ER  -