arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.07954cs.CV
收藏

走向人类可理解的视觉解释:不可感知的高频线索可以更好地被去除

Towards Human-Understandable Visual Explanations:Imperceptible High-frequency Cues Can Better Be Removed

KU Leuven(鲁汶大学);University of Antwerp(安特卫普大学);imec-IDLab

Kaili Wang, Jose Oramas, Tinne Tuytelaars

摘要:

可解释人工智能(XAI)方法专注于解释神经网络学到了什么-换句话说,识别对预测影响最大的特征。在本文中,我们称之为“区别特征”。然而,人类是否能够理解生成的解释还取决于这些特征对人类的感知能力。为了确保解释是人类可以理解的,我们认为,人类的能力,受人类视觉系统(HVS)和心理物理学,需要考虑。我们提出了XAI的{ \em人类可感知性原则},指出为了生成人类可理解的解释,神经网络应该在训练过程中专注于人类可理解的线索。我们进行了一个关于真实与虚假人脸图像分类的案例研究,其中许多由标准神经网络提取的区别特征对人类来说是不可感知的。通过应用所提出的原则,一个神经网络与人类可理解的解释训练,在用户研究中,被证明是更好地符合人类的直觉。这可能会使人工智能更值得信赖,并为人类向机器学习打开大门。在案例研究中,我们具体研究和分析了神经网络和XAI方法中人类无法感知的高空间频率特征的行为。

英文摘要:

Explainable AI (XAI) methods focus on explaining what a neural network has learned - in other words, identifying the features that are the most influential to the prediction. In this paper, we call them "distinguishing features". However, whether a human can make sense of the generated explanation also depends on the perceptibility of these features to humans. To make sure an explanation is human-understandable, we argue that the capabilities of humans, constrained by the Human Visual System (HVS) and psychophysics, need to be taken into account. We propose the {\em human perceptibility principle for XAI}, stating that, to generate human-understandable explanations, neural networks should be steered towards focusing on human-understandable cues during training. We conduct a case study regarding the classification of real vs. fake face images, where many of the distinguishing features picked up by standard neural networks turn out not to be perceptible to humans. By applying the proposed principle, a neural network with human-understandable explanations is trained which, in a user study, is shown to better align with human intuition. This is likely to make the AI more trustworthy and opens the door to humans learning from machines. In the case study, we specifically investigate and analyze the behaviour of the human-imperceptible high spatial frequency features in neural networks and XAI methods.

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。

↑