Bluff: Interactively Deciphering Adversarial Attacks on Deep Neural\n Networks

2020/09/05 by Nilaksh Das, Haekyu Park, Das, Nilaksh +11 · 1 citation
Computer Science · #Adversarial Robustness in Machine Learning #Anomaly Detection Techniques and Applications #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2009.02608

openalex publication_date 2020/09/05 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28

Abstract

Deep neural networks (DNNs) are now commonly used in many domains. However,\nthey are vulnerable to adversarial attacks: carefully crafted perturbations on\ndata inputs that can fool a model into making incorrect predictions. Despite\nsignificant research on developing DNN attack and defense techniques, people\nstill lack an understanding of how such attacks penetrate a model's internals.\nWe present Bluff, an interactive system for visualizing, characterizing, and\ndeciphering adversarial attacks on vision-based neural networks. Bluff allows\npeople to flexibly visualize and compare the activation pathways for benign and\nattacked images, revealing mechanisms that adversarial attacks employ to\ninflict harm on a model. Bluff is open-sourced and runs in modern web browsers.\n

Cited by

• ChannelExplorer: Exploring Class Separability Through Activation Channel Visualization

Related

• AFLOW: Developing Adversarial Examples under Extremely Noise-limited Settings
• Explaining Away Attacks Against Neural Networks
• Adversarial-Playground: A Visualization Suite Showing How Adversarial Examples Fool Deep Learning
• FenceBox: A Platform for Defeating Adversarial Examples with Data Augmentation Techniques
• AdvDrop: Adversarial Attack to DNNs by Dropping Information