LORD: Large Models based Opposite Reward Design for Autonomous Driving

arXiv:2403.18965v1 Announce Type: cross
Abstract: Reinforcement learning (RL) based autonomous driving has emerged as a promising alternative to data-driven imitation learning approaches. However, crafting effective reward functions for RL poses challenges due to the complexity of defining and quantifying good driving behaviors across diverse scenarios. Recently, large pretrained models have gained significant attention as zero-shot reward models for tasks specified with desired linguistic goals. However, the desired linguistic goals for autonomous driving such as "drive safely" are ambiguous and incomprehensible …

abstract arxiv autonomous autonomous driving challenges complexity cs.ai cs.lg cs.ro data data-driven design diverse driving functions good however imitation learning large models pretrained models reinforcement reinforcement learning type