arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2006.14911cs.LGcs.ROstat.ML
收藏

自动驾驶汽车能识别、恢复和适应分布偏移吗?

Can Autonomous Vehicles Identify, Recover From, and Adapt to Distribution Shifts?

University of Oxford(牛津大学);University of California, Berkeley(加利福尼亚大学伯克利分校)

Angelos Filos, Panagiotis Tigas, Rowan McAllister, Nicholas Rhinehart, Sergey Levine, Yarin Gal

摘要:

训练分布外(OOD)场景是学习代理在部署时的常见挑战,通常会导致任意推断和信息不足的决策。原则上,OOD场景的检测和适应可以减轻其不利影响。在本文中,我们强调了目前的方法,以新颖的驾驶场景的局限性,并提出了一种认知不确定性感知的规划方法,称为\emph{robust imitative planning}(RIP)。我们的方法可以检测和恢复一些分布偏移,减少过度自信和灾难性的外推OOD场景。如果模型的不确定性太大,无法提出安全的行动方案,则模型可以向专家驾驶员查询反馈,从而实现样本高效的在线适应,这是我们的方法的一种变体,我们称之为\emph{adaptive robust imitative planning}(AdaRIP)。我们的方法在nuScenes \emph{prediction}挑战中优于当前最先进的方法,但由于目前没有评估OOD检测和自适应的基准来评估\emph{control},我们引入了一个自动驾驶汽车新场景基准,\texttt{CARNOVEL},以评估驾驶代理对具有分布偏移的一系列任务的鲁棒性。

英文摘要:

Out-of-training-distribution (OOD) scenarios are a common challenge of learning agents at deployment, typically leading to arbitrary deductions and poorly-informed decisions. In principle, detection of and adaptation to OOD scenes can mitigate their adverse effects. In this paper, we highlight the limitations of current approaches to novel driving scenes and propose an epistemic uncertainty-aware planning method, called \emph{robust imitative planning} (RIP). Our method can detect and recover from some distribution shifts, reducing the overconfident and catastrophic extrapolations in OOD scenes. If the model's uncertainty is too great to suggest a safe course of action, the model can instead query the expert driver for feedback, enabling sample-efficient online adaptation, a variant of our method we term \emph{adaptive robust imitative planning} (AdaRIP). Our methods outperform current state-of-the-art approaches in the nuScenes \emph{prediction} challenge, but since no benchmark evaluating OOD detection and adaption currently exists to assess \emph{control}, we introduce an autonomous car novel-scene benchmark, \texttt{CARNOVEL}, to evaluate the robustness of driving agents to a suite of tasks with distribution shifts.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑