TY - GEN
T1 - Purifying Adversarial Examples Using an Autoencoder
AU - van Weezel, Thijs
AU - van Ree, Famke
AU - Bos, Tychon
AU - Bastiaanssen, Patrick
AU - Hess, Sibylle
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.
PY - 2025/1/25
Y1 - 2025/1/25
N2 - One of the most prominent security challenges to neural networks are adversarial examples - inputs with often barely perceptible perturbations causing misclassification. In this study, we propose a defense mechanism that uses an autoencoder to restore adversarial examples before classification. That is, the autoencoder purifies input data points from potential adversarial perturbations. The method is titled Autoencoder-based Adversarial Purification (AAP). We demonstrate the effectiveness of AAP on multiple datasets, attack methods, and perturbation levels. While certain limitations exist, this research offers valuable insights and a promising direction for robust defense mechanisms in adversarial deep learning.
AB - One of the most prominent security challenges to neural networks are adversarial examples - inputs with often barely perceptible perturbations causing misclassification. In this study, we propose a defense mechanism that uses an autoencoder to restore adversarial examples before classification. That is, the autoencoder purifies input data points from potential adversarial perturbations. The method is titled Autoencoder-based Adversarial Purification (AAP). We demonstrate the effectiveness of AAP on multiple datasets, attack methods, and perturbation levels. While certain limitations exist, this research offers valuable insights and a promising direction for robust defense mechanisms in adversarial deep learning.
KW - adversarial attack
KW - autoencoder
KW - purification
UR - https://www.scopus.com/pages/publications/85219196330
U2 - 10.1007/978-3-031-78980-9_9
DO - 10.1007/978-3-031-78980-9_9
M3 - Conference contribution
AN - SCOPUS:85219196330
SN - 9783031789793
T3 - Lecture Notes in Computer Science
SP - 134
EP - 148
BT - Discovery Science - 27th International Conference, DS 2024, Proceedings
A2 - Pedreschi, Dino
A2 - Monreale, Anna
A2 - Guidotti, Riccardo
A2 - Naretto, Francesca
A2 - Pellungrini, Roberto
PB - Springer
T2 - 27th International Conference on Discovery Science, DS 2024
Y2 - 14 October 2024 through 16 October 2024
ER -