The susceptibility of deep neural networks to untrustworthy predictions, including out-of-distribution (OOD) data and adversarial examples, still prevent their widespread use in safety-critical applications. Most existing methods either require a re-training of a given model to achieve robust identification of adversarial attacks or are limited to out-of-distribution sample detection only. In this work, we propose a geometric gradient analysis (GGA) to improve the identification of untrustworthy predictions without retraining of a given model. GGA analyzes the geometry of the loss landscape of neural networks based on the saliency maps of their respective input. To motivate the proposed approach, we provide theoretical connections between gradients' geometrical properties and local minima of the loss function. Furthermore, we demonstrate that the proposed method outperforms prior approaches in detecting OOD data and adversarial attacks, including state-of-the-art and adaptive attacks.
Identifying critical residues of a protein using meaningfully-thresholded Random Geometric Graphs
Department of Mathematics, Brunel University London（布鲁内尔大学数学系）；The Thomas Young Centre for Theory and Simulation of Materials（泰晤士·扬材料理论与模拟中心）；Department of Bioscience and Bioengineering, Indian Institute of Technology（印度理工学院生物科学与生物工程系）；Department of Mathematics, University of York（约克大学数学系）
Importance of localized dilatation and distensibility in identifying determinants of thoracic aortic aneurysm with neural operators
Department of Biomedical Engineering, Yale University（耶鲁大学生物医学工程系）；Department of Civil and Systems Engineering, Johns Hopkins University（约翰霍普金斯大学土木与系统工程系）；Division of Applied Mathematics, Brown University（布朗大学应用数学系）；School of Engineering, Brown University（布朗大学工程学院）；Department of Surgery, Yale School of Medicine（耶鲁医学院外科系）；Vascular Biology and Therapeutics Program, Yale School of Medicine（耶鲁医学院血管生物学与治疗学项目）
Identifying Terrain Physical Parameters from Vision -- Towards Physical-Parameter-Aware Locomotion and Navigation
ETH Zurich（苏黎世联邦理工学院）；Max Planck Institute for Intelligent Systems Tübingen（马克斯·普朗克智能系统研究所（图宾根））；Harbin Institute of Technology（哈尔滨工业大学）；University of Tübingen（图宾根大学）
Tackling Racial Bias in Automated Online Hate Detection: Towards Fair and Accurate Classification of Hateful Online Users Using Geometric Deep Learning
Oxford Internet Institute, University of Oxford（牛津大学互联网研究所）；The Alan Turing Institute（艾伦·图灵研究所）；Meedan（米丹（Meedan））
