Despite the enormous success of artificial neural networks (ANNs) in many disciplines, the characterization of their computations and the origin of key properties such as generalization and robustness remain open questions. Recent literature suggests that robust networks with good generalization properties tend to be biased towards processing low frequencies in images. To explore the frequency bias hypothesis further, we develop an algorithm that allows us to learn modulatory masks highlighting the essential input frequencies needed for preserving a trained network's performance. We achieve this by imposing invariance in the loss with respect to such modulations in the input frequencies. We first use our method to test the low-frequency preference hypothesis of adversarially trained or data-augmented networks. Our results suggest that adversarially robust networks indeed exhibit a low-frequency bias but we find this bias is also dependent on directions in frequency space. However, this is not necessarily true for other types of data augmentation. Our results also indicate that the essential frequencies in question are effectively the ones used to achieve generalization in the first place. Surprisingly, images seen through these modulatory masks are not recognizable and resemble texture-like patterns.
Lomonosov Moscow State University(罗蒙诺索夫莫斯科国立大学);Huawei Technologies Co., Ltd.(华为技术有限公司);China Nanhu Academy of Electronics and Information Technology(中国南湖电子信息技术研究院)
Contour-Native Bridge Defect Detection and Compact Digital Archiving with Frequency-Supervised Fourier Contours
State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University(测绘遥感信息工程国家重点实验室,武汉大学);Sustainability X-Lab, The University of Hong Kong(可持续性X实验室,香港大学);Department of Cyber Security, Southeast University(安全与保密系,东南大学);School of Computer Science and Engineering, University of Electronic Science and Technology of China(电子科技大学计算机科学与工程学院);Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology(香港理工大学电子与计算机工程系)
Scaling Artificial Intelligence for Multi-Tumor Early Detection with More Reports, Fewer Masks
Johns Hopkins University(约翰霍普金斯大学);University of Bologna(博洛尼亚大学);Istituto Italiano di Tecnologia(意大利理工学院);Jagiellonian University(雅盖隆大学);University of California, Berkeley(加州大学伯克利分校);Warmian-Masurian Cancer Center(马里安马祖尔癌症中心);University of Zurich(苏黎世大学);ETH AI Center(苏黎世联邦理工学院人工智能中心);Istanbul Medipol University(伊斯坦布尔梅迪波尔大学);University of California, Santa Cruz(加州大学圣克鲁兹分校);Stanford University(斯坦福大学);University of Warmia and Mazury(马里安马祖尔大学);École Polytechnique Fédérale de Lausanne(洛桑联邦理工学院);University of California, San Francisco(加州大学旧金山分校)