Extending Defensive Distillation

Abstract

Machine learning is vulnerable to adversarial examples: inputs carefully modified to force misclassification. Designing defenses against such inputs remains largely an open problem. In this work, we revisit defensive distillation---which is one of the mechanisms proposed to mitigate adversarial examples---to address its limitations. We view our results not only as an effective way of addressing some of the recently discovered attacks but also as reinforcing the importance of improved training techniques.

References in corpus (6)

• Distilling the Knowledge in a Neural Network
• The Space of Transferable Adversarial Examples
• Detecting Adversarial Samples from Artifacts
• On the (Statistical) Detection of Adversarial Examples
• Blocking Transferability of Adversarial Examples in Black-Box Learning Systems
• Comment on "Biologically inspired protection of deep networks from adversarial attacks"

Cited by in corpus (29)

• ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Models
• DeepXplore: Automated Whitebox Testing of Deep Learning Systems
• DeepTest: Automated Testing of Deep-Neural-Network-driven Autonomous Cars
• Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey
• How Deep Learning Sees the World: A Survey on Adversarial Attacks & Defenses
• APE-GAN: Adversarial Perturbation Elimination with GAN
• Robust Pre-Training by Adversarial Contrastive Learning
• Domain Robustness in Neural Machine Translation
• Testing and verification of neural-network-based safety-critical control software: A systematic literature review
• Detecting Adversarial Samples for Deep Neural Networks through Mutation Testing
• Adversarially Robust Neural Architectures
• Detecting Adversarial Attacks on Neural Network Policies with Visual Foresight
• Adversarial Robustness: From Self-Supervised Pre-Training to Fine-Tuning
• Toward Robust Image Classification
• Universal Perturbation Attack Against Image Retrieval
• AdvSPADE: Realistic Unrestricted Attacks for Semantic Segmentation
• Gradient Band-based Adversarial Training for Generalized Attack Immunity of A3C Path Finding
• Enhancing Gradient-based Attacks with Symbolic Intervals
• Towards Robust Training of Neural Networks by Regularizing Adversarial Gradients
• Improving Adversarial Robustness via Guided Complement Entropy
• GanDef: A GAN based Adversarial Training Defense for Neural Network Classifier
• Weight Map Layer for Noise and Adversarial Attack Robustness
• Latent Adversarial Defence with Boundary-guided Generation
• A Geometrical Approach to Evaluate the Adversarial Robustness of Deep Neural Networks
• ZK-GanDef: A GAN based Zero Knowledge Adversarial Training Defense for Neural Networks
• Deep Minimax Probability Machine
• Mixing between the Cross Entropy and the Expectation Loss Terms
• Adversarial Attack Attribution: Discovering Attributable Signals in Adversarial ML Attacks
• An Explainable Adversarial Robustness Metric for Deep Learning Neural Networks