@misc{indiciae41614821cd1f,
  title = {HeiSD: Hybrid Speculative Decoding for Embodied Vision-Language-Action Models with Kinematic Awareness},
  author = {Zihao Zheng and Zhihao Mao and Sicheng Tian and Maoliang Li and Jiayu Chen and Xinhao Sun and Zhaobo Zhang and Xuanzhe Liu and Donggang Cao and Hong Mei and Xiang Chen},
  year = {2026},
  url = {https://arxiv.org/abs/2603.17573},
  note = {Source identifier: 2603.17573}
}