WAM-Diff2: Hierarchical AR-to-Diffusion Distillation for Highly Efficient Autonomous Driving VLA

Excerpt

arXiv:2608.01035v3 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models have emerged as a prominent paradigm for end-to-end autonomous driving; however, their efficient deployment is severely constrained by high computational latency and exposure bias arising from sequential autoregressive decoding. Conversely, while specialized diffusion policies enable low-latency, parallel execution, training them from scratch typically yields narrow, single-task architectures that lack