DreamerAD: Efficient Reinforcement Learning via Latent World Model for Autonomous Driving
This work addresses the high costs and safety risks of training RL policies on real-world driving data, offering a more efficient solution for autonomous driving applications.
The paper tackled the problem of inefficient reinforcement learning for autonomous driving by introducing DreamerAD, a latent world model framework that compresses diffusion sampling from 100 steps to 1, achieving an 80x speedup while maintaining visual interpretability and reaching 87.7 EPDMS on NavSim v2.
We introduce DreamerAD, the first latent world model framework that enables efficient reinforcement learning for autonomous driving by compressing diffusion sampling from 100 steps to 1 - achieving 80x speedup while maintaining visual interpretability. Training RL policies on real-world driving data incurs prohibitive costs and safety risks. While existing pixel-level diffusion world models enable safe imagination-based training, they suffer from multi-step diffusion inference latency (2s/frame) that prevents high-frequency RL interaction. Our approach leverages denoised latent features from video generation models through three key mechanisms: (1) shortcut forcing that reduces sampling complexity via recursive multi-resolution step compression, (2) an autoregressive dense reward model operating directly on latent representations for fine-grained credit assignment, and (3) Gaussian vocabulary sampling for GRPO that constrains exploration to physically plausible trajectories. DreamerAD achieves 87.7 EPDMS on NavSim v2, establishing state-of-the-art performance and demonstrating that latent-space RL is effective for autonomous driving.