arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.06359cs.LG
收藏

修复动态神经网络中的过度自信问题

Fixing Overconfidence in Dynamic Neural Networks

Aalto University(阿尔托大学);Saab Finland Oy(萨博芬兰公司);NVIDIA(英伟达);Xi’an Jiaotong University(西安交通大学)

Lassi Meronen, Martin Trapp, Andrea Pilzer, Le Yang, Arno Solin

AI总结:

本文针对动态神经网络难以区分难易样本且不确定性估计过度自信的问题,提出对最后几层进行概率化处理的事后不确定性量化方法,以同时建模偶然与认知不确定性,并在多个图像数据集上提升准确率、不确定性捕捉与校准表现。

摘要:

动态神经网络是一项近期技术,它有望通过根据输入难度动态调整计算成本,来缓解现代深度学习模型规模不断增大的问题。通过这种方式,模型可以适应有限的计算预算。然而,深度学习模型中不确定性估计的质量较差,使得难以区分困难样本和简单样本。为应对这一挑战,我们提出了一种计算高效的方法,用于动态神经网络中的事后不确定性量化。我们表明,通过对最后几层进行概率化处理,充分量化并考虑偶然不确定性和认知不确定性,可以提高预测性能,并有助于确定计算预算时的决策。在实验中,我们在 CIFAR-100、ImageNet 和 Caltech-256 上展示了在准确率、不确定性捕捉和校准误差方面的改进。

英文摘要:

Dynamic neural networks are a recent technique that promises a remedy for the increasing size of modern deep learning models by dynamically adapting their computational cost to the difficulty of the inputs. In this way, the model can adjust to a limited computational budget. However, the poor quality of uncertainty estimates in deep learning models makes it difficult to distinguish between hard and easy samples. To address this challenge, we present a computationally efficient approach for post-hoc uncertainty quantification in dynamic neural networks. We show that adequately quantifying and accounting for both aleatoric and epistemic uncertainty through a probabilistic treatment of the last layers improves the predictive performance and aids decision-making when determining the computational budget. In the experiments, we show improvements on CIFAR-100, ImageNet, and Caltech-256 in terms of accuracy, capturing uncertainty, and calibration error.

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。

↑