LLaDA：通过大型语言模型策略适应实现无处不在的驾驶

报告摘要

Driving Everywhere with Large Language Model Policy Adaptation arXiv:2402.05932v2 [cs.RO] 10 Apr 2024 Boyi Li1 Yue Wang1,2 Jiageng Mao2 Boris Ivanovic1 Sushant Veer1 Karen Leung1,3 Marco Pavone1,4

核心要点

1. 1. Introduction
2. 1. We propose LLaDA, a training-free mechanism to assist
3. 2. LLaDA can be immediately applied to any autonomous
4. 3. Our method achieves performance improvements over
5. 2. Related Works

报告信息