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 2511.10282cs.LGcs.AI
收藏

Torch-Uncertainty:一个面向不确定性量化的深度学习框架

Torch-Uncertainty: A Deep Learning Framework for Uncertainty Quantification

U2IS, ENSTA, Institut Polytechnique de Paris(U2IS、ENSTA、巴黎理工学院)

Adrien Lafage, Olivier Laurent, Firas Gabetni, Gianni Franchi

AI总结:

针对深度神经网络难以准确量化预测不确定性、现有UQ方法缺乏统一整合工具的问题,本文推出基于PyTorch与Lightning的Torch-Uncertainty框架,可支持多任务UQ方法的训练、评估与基准测试。

摘要:

深度神经网络(DNNs)已在计算机视觉、自然语言处理等多个领域展现出卓越性能,但往往难以准确量化其预测结果的不确定性,这限制了其在关键现实应用中的更广泛落地。深度学习不确定性量化(UQ)旨在通过提供提升不确定性估计可靠性的方法来解决这一挑战。尽管目前已提出大量相关技术,但仍缺乏可提供无缝工作流以评估、整合这些方法的统一工具。为填补这一空白,我们推出Torch-Uncertainty,这是一个基于PyTorch与Lightning构建的框架,旨在简化结合UQ技术与指标的DNN训练及评估流程。本文中我们概述了该库的基础原理,并给出了在分类、分割、回归任务上对多种UQ方法进行基准测试的全面实验结果,该库可在https://github.com/ENSTA-U2IS-AI/Torch-Uncertainty获取。

英文摘要:

Deep Neural Networks (DNNs) have demonstrated remarkable performance across various domains, including computer vision and natural language processing. However, they often struggle to accurately quantify the uncertainty of their predictions, limiting their broader adoption in critical real-world applications. Uncertainty Quantification (UQ) for Deep Learning seeks to address this challenge by providing methods to improve the reliability of uncertainty estimates. Although numerous techniques have been proposed, a unified tool offering a seamless workflow to evaluate and integrate these methods remains lacking. To bridge this gap, we introduce Torch-Uncertainty, a PyTorch and Lightning-based framework designed to streamline DNN training and evaluation with UQ techniques and metrics. In this paper, we outline the foundational principles of our library and present comprehensive experimental results that benchmark a diverse set of UQ methods across classification, segmentation, and regression tasks. Our library is available at https://github.com/ENSTA-U2IS-AI/Torch-Uncertainty

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。

↑