1 paper · 1 filter
Pengxuan Yang, Yupeng Zheng, Deheng Qian +11
We introduce DreamerAD, the first latent world model framework that enables efficient reinforcement learning for autonomous driving by compressing diffusion sampling from 100 steps…