arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.18260cs.LG
收藏

主动学习中高斯过程的可高效计算安全界

Efficiently Computable Safety Bounds for Gaussian Processes in Active Learning

OWL University of Applied Sciences and Arts(OWL应用科学与艺术大学);Bosch Center for Artificial Intelligence(博世人工智能中心);RWTH Aachen University(亚琛工业大学);TU Delft(代尔夫特理工大学)

Jörn Tebbe, Christoph Zimmer, Ansgar Steland, Markus Lange-Hegermann, Fabian Mies

AI总结:

针对GP主动学习中安全评估计算成本高的问题,提出基于后验GP上确界自适应采样中位数的可证明安全界方法,在保证精度与探索速度的同时提升评估效率,经仿真与真实发动机案例验证有效。

摘要:

物理系统的主动学习通常需遵守实际安全约束,这限制了对设计空间的探索。高斯过程(Gaussian Processes, GPs)及其经校准的不确定性估计被广泛用于该场景。在许多技术应用中,设计空间通过连续轨迹进行探索,需沿轨迹评估安全性。这对GP方法中的严格安全要求而言极具挑战性,因为该方法采用计算成本高昂的高分位数蒙特卡洛采样。我们通过基于后验GP上确界的自适应采样中位数,提供了可证明的安全界,以应对这些挑战。我们的方法大幅减少了估计高安全概率所需的样本数量,在不牺牲精度和探索速度的前提下实现了更快的评估。我们通过大量仿真验证了该安全主动学习方法的有效性,并借助一个真实发动机实例进行了验证。

英文摘要:

Active learning of physical systems must commonly respect practical safety constraints, which restricts the exploration of the design space. Gaussian Processes (GPs) and their calibrated uncertainty estimations are widely used for this purpose. In many technical applications the design space is explored via continuous trajectories, along which the safety needs to be assessed. This is particularly challenging for strict safety requirements in GP methods, as it employs computationally expensive Monte-Carlo sampling of high quantiles. We address these challenges by providing provable safety bounds based on the adaptively sampled median of the supremum of the posterior GP. Our method significantly reduces the number of samples required for estimating high safety probabilities, resulting in faster evaluation without sacrificing accuracy and exploration speed. The effectiveness of our safe active learning approach is demonstrated through extensive simulations and validated using a real-world engine example.

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。

↑