Bluff: Interactively Deciphering Adversarial Attacks on Deep Neural Networks
| Autor: | Zijie J. Wang, Duen Horng Chau, Nilaksh Das, Haekyu Park, Fred Hohman, Robert Firstman, Emily Rogers |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
FOS: Computer and information sciences
Computer Science - Machine Learning Web browser Computer Science - Cryptography and Security Artificial neural network business.industry Computer science 020207 software engineering 02 engineering and technology Computer security computer.software_genre Machine Learning (cs.LG) Data modeling Adversarial system Harm Data visualization Bluff 0202 electrical engineering electronic engineering information engineering Deep neural networks 020201 artificial intelligence & image processing business Cryptography and Security (cs.CR) computer |
| Zdroj: | IEEE VIS (Short Papers) |
| Popis: | Deep neural networks (DNNs) are now commonly used in many domains. However, they are vulnerable to adversarial attacks: carefully crafted perturbations on data inputs that can fool a model into making incorrect predictions. Despite significant research on developing DNN attack and defense techniques, people still lack an understanding of how such attacks penetrate a model's internals. We present Bluff, an interactive system for visualizing, characterizing, and deciphering adversarial attacks on vision-based neural networks. Bluff allows people to flexibly visualize and compare the activation pathways for benign and attacked images, revealing mechanisms that adversarial attacks employ to inflict harm on a model. Bluff is open-sourced and runs in modern web browsers. This paper is accepted at IEEE VIS'20 Short Paper |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|