Latency-Tolerant Cloud-Edge Collaborative Vision-Language-Action Models via Emergent Representational Specialization
CloudEdgeVLA is introduced, a cloud-edge policy that treats temporal misalignment as a representation-learning problem and offers a practical route to scalable VLA deployment in which cloud models can grow while edge computation remains lightweight and responsive.
Dao-Jie Peng, Fu-Long Ma, Bing-Tao Wang et al.
· 1 citation