A Unified Framework for Detecting Audio Adversarial Examples

Du, Xia; Pun, Chi‐Man; Zhang, Zheng

doi:10.1145/3394171.3413603

Cited by 14 publications

(9 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Audio reverberation. Xia et al [12] proposed a robust detection method based on an audio reverberation technique. They pioneered the study of defense against robust over-air adversarial examples and discovered that audio adversarial examples are prone to overfitting and continuity.…”

Section: Defenses Against Audio Adversarial Examples Have Been Extensivelymentioning

confidence: 99%

“…FPR is the rate at which a method mistakes benign audio waves for false adversarial 4 ); otherwise, we use the publicly available adversarial examples generated by the authors (weight-sampling attack 5 and metamorph attack 6 ). We also reproduce three state-of-the-art detection methods [12,16,35] for performance comparison.…”

Section: Experimental Evaluation 51 Experimental Setupmentioning

confidence: 99%

“…Reverb. [12] represents the detection method that adds reverberation noise to the input waves. Reverb.…”

Section: Detection Accuracymentioning

confidence: 99%

“…Although these methods seem to be effective for over-line audio adversarial examples, these research did not consider overair audio adversarial examples. Recently, a new method based on signal reverberations [12] was proposed to detect robust physical audio adversarial examples. However, this method requires multiple wave inferences to evaluate whether the input wave is an audio adversarial example or not; therefore, it has a large time overhead.…”

Section: Introductionmentioning

confidence: 99%

“…The proposed method is easily applicable to existing trained models because it requires no structural change or no additional training procedure. Moreover, as the method requires no additional inference for an input, it reduces the time overhead compared with state-of-the-art detection methods [12,16,35].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Detecting Audio Adversarial Examples with Logit Noising

Park

Kim

2021

Annual Computer Security Applications Conference

View full text Add to dashboard Cite

Automatic speech recognition (ASR) systems are vulnerable to audio adversarial examples that attempt to deceive ASR systems by adding perturbations to benign speech signals. Although an adversarial example and the original benign wave are indistinguishable to humans, the former is transcribed as a malicious target sentence by ASR systems. Several methods have been proposed to generate audio adversarial examples and feed them directly into the ASR system (over-line). Furthermore, many researchers have demonstrated the feasibility of robust physical audio adversarial examples (over-air). To defend against the attacks, several studies have been proposed. However, deploying them in a real-world situation is difficult because of accuracy drop or time overhead.In this paper, we propose a novel method to detect audio adversarial examples by adding noise to the logits before feeding them into the decoder of the ASR. We show that carefully selected noise can significantly impact the transcription results of the audio adversarial examples, whereas it has minimal impact on the transcription results of benign audio waves. Based on this characteristic, we detect audio adversarial examples by comparing the transcription altered by logit noising with its original transcription. The proposed method can be easily applied to ASR systems without any structural changes or additional training. The experimental results show that the proposed method is robust to over-line audio adversarial examples as well as over-air audio adversarial examples compared with state-of-the-art detection methods. CCS CONCEPTS• Security and privacy; • Computing methodologies → Speech recognition;

show abstract

Section: Defenses Against Audio Adversarial Examples Have Been Extensivelymentioning

confidence: 99%

Section: Experimental Evaluation 51 Experimental Setupmentioning

confidence: 99%

“…Reverb. [12] represents the detection method that adds reverberation noise to the input waves. Reverb.…”

Section: Detection Accuracymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Detecting Audio Adversarial Examples with Logit Noising

Park

Kim

2021

Annual Computer Security Applications Conference

View full text Add to dashboard Cite

show abstract

A Security Study of Multimodel Artificial Intelligence System: Adaptive Retention Attack for Object Detection System with Multifocus Image Fusion Model

Gao,

Jin,

Miao

et al. 2024

Advanced Intelligent Systems

View full text Add to dashboard Cite

Image preprocessing models are usually employed as the preceding operations of high‐level vision tasks to improve the performance. The adversarial attack technology makes both these models face severe challenges. Prior research is focused solely on attacking single object detection models, without considering the impact of the preprocessing models (multifocus image fusion) on adversarial perturbations within the object detection system. Multifocus image fusion models work in conjunction with the object detection models to enhance the quality of the images and improve the capability of object detection system. Herein, the problem of attacking object detection system that utilizes multifocus image fusion as its preprocessing models is addressed. To retain the attack capabilities of adversarial samples against as many perturbations as possible, new attack method called adaptive retention attack (ARA) is proposed. Additionally, adversarial perturbations concentration mechanism and image selection mechanism, which, respectively, enhance the transferability and attack capability of ARA‐generated adversarial samples. Extensive experiments have demonstrated the feasibility of the ARA. The results confirm that the ARA method can successfully bypass multifocus image fusion models to attack the object detection model.

show abstract

Adversarial example‐based test case generation for black‐box speech recognition systems

Cai

Zhang

Dong

et al. 2023

Software Testing Verif & Rel

View full text Add to dashboard Cite

Test case generation techniques based on adversarial examples are commonly used to enhance the reliability and robustness of image-based and text-based machine learning applications. However, efficient techniques for speech recognition systems are still absent. This paper proposes a family of methods that generate targeted adversarial examples for speech recognition systems. All are based on the firefly algorithm (F), and are enhanced with gauss mutations and / or gradient estimation (F-GM, F-GE, F-GMGE) to fit the specific problem of targeted adversarial test case generation. We conduct an experimental evaluation on three different types of speech datasets, including Google Command, Common Voice and LibriSpeech. In addition, we recruit volunteers to evaluate the performance of the adversarial examples. The experimental results show that, compared with existing approaches, these approaches can effectively improve the success rate of the targeted adversarial example generation. The code is publicly available at https://github.com/HanboCai/FGMGE.

show abstract

A Unified Framework for Detecting Audio Adversarial Examples

Cited by 14 publications

References 26 publications

Detecting Audio Adversarial Examples with Logit Noising

Detecting Audio Adversarial Examples with Logit Noising

A Security Study of Multimodel Artificial Intelligence System: Adaptive Retention Attack for Object Detection System with Multifocus Image Fusion Model

Adversarial example‐based test case generation for black‐box speech recognition systems

Contact Info

Product

Resources

About