Characterizing Audio Adversarial Examples Using Temporal Dependency

Abstract

Recent studies have highlighted adversarial examples as a ubiquitous threat to different neural network models and many downstream applications. Nonetheless, as unique data properties have inspired distinct and powerful learning principles, this paper aims to explore their potentials towards mitigating adversarial inputs. In particular, our results reveal the importance of using the temporal dependency in audio data to gain discriminate power against adversarial examples. Tested on the automatic speech recognition (ASR) tasks and three recent audio adversarial attacks, we find that (i) input transformation developed from image adversarial defense provides limited robustness improvement and is subtle to advanced attacks; (ii) temporal dependency can be exploited to gain discriminative power against audio adversarial examples and is resistant to adaptive attacks considered in our experiments. Our results not only show promising means of improving the robustness of ASR systems, but also offer novel insights in exploiting domain-specific data properties to mitigate negative effects of adversarial examples.

References in corpus (19)

• Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks
• Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
• Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition
• Delving into Transferable Adversarial Examples and Black-box Attacks
• Countering Adversarial Images using Input Transformations
• Conditional Generative Adversarial Networks for Speech Enhancement and Noise-Robust Speaker Verification
• A study of the effect of JPG compression on adversarial images
• Did you hear that? Adversarial Examples Against Automatic Speech Recognition
• Houdini: Fooling Deep Structured Prediction Models
• MagNet and "Efficient Defenses Against Adversarial Attacks" are Not Robust to Adversarial Examples
• Foveation-based Mechanisms Alleviate Adversarial Examples
• Adversarial Example Defenses: Ensembles of Weak Defenses are not Strong
• CommanderSong: A Systematic Approach for Practical Adversarial Voice Recognition
• Invariant Representations for Noisy Speech Recognition
• EAD: Elastic-Net Attacks to Deep Neural Networks via Adversarial Examples
• Attacking Visual Language Grounding with Adversarial Examples: A Case Study on Neural Image Captioning
• On the Limitation of MagNet Defense against $L_1$-based Adversarial Examples
• Using Non-invertible Data Transformations to Build Adversarial-Robust Neural Networks
• Is Robustness the Cost of Accuracy? -- A Comprehensive Study on the Robustness of 18 Deep Image Classification Models

Cited by in corpus (18)

• On Adaptive Attacks to Adversarial Example Defenses
• Black-box Adversarial Attacks on Commercial Speech Platforms with Minimal Information
• Characterizing Speech Adversarial Examples Using Self-Attention U-Net Enhancement
• RAB: Provable Robustness Against Backdoor Attacks
• Weighted-Sampling Audio Adversarial Example Attack
• Universal Adversarial Audio Perturbations
• Identifying Audio Adversarial Examples via Anomalous Pattern Detection
• Anomalous Example Detection in Deep Learning: A Survey
• Towards Evaluating the Robustness of Chinese BERT Classifiers
• On the Detection of Adaptive Adversarial Attacks in Speaker Verification Systems
• Towards Resistant Audio Adversarial Examples
• Detecting Adversarial Attacks On Audiovisual Speech Recognition
• Effects of Loss Functions And Target Representations on Adversarial Robustness
• LanCe: A Comprehensive and Lightweight CNN Defense Methodology against Physical Adversarial Attacks on Embedded Multimedia Applications
• Dompteur: Taming Audio Adversarial Examples
• DoPa: A Comprehensive CNN Detection Methodology against Physical Adversarial Attacks
• On the human evaluation of audio adversarial examples
• Can audio-visual integration strengthen robustness under multimodal attacks?