A Simple Neural Network Based Countermeasure for Replay Attack

Pang, Wenfeng; He, Qian

doi:10.1145/3158233.3159308

Cited by 2 publications

(8 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Knowledge distillation (KD) [9]. The KD attack retrains a model from scratch by minimizing the distance between the teacher's and student's soft predictions plus the cross-entropy loss between the student's prediction…”

Section: Methodsmentioning

confidence: 99%

“…According to Shafieinejad et al [33], existing watermark embedding techniques [14], [32] and some fingerprinting solutions [10], [24] cannot withstand model extraction attacks. Distillation [34] was first proposed to distill the knowledge of teacher models into student models and later extended as an attack against methods that protect model copyrights [9]. Distilled models are often able to evade copyright tracking, as demonstrated in works such as Cao et al [24] and Lukas et al [25].…”

Section: Model Stealing Attacksmentioning

confidence: 99%

“…• Black-box attacker. Attackers, who are external entities attempting to exploit the functionality of the victim model, employ various attacks such as model extraction attacks [6] and model distillation attacks [9]. • White-box attacker.…”

Section: Threat Modelmentioning

confidence: 99%

“…Model extraction attack efficiently transfers the functionality of a victim model to a stolen copy using limited query answers. Additionally, attackers, who may be insiders with full access to the victim models, employ techniques such as distillation [9], fine-tuning [10], or pruning [11], [12] in an attempt to reverse-engineer the tracking.…”

Section: Introductionmentioning

confidence: 99%

“…Many stealing mechanisms have minimal side effects on the model's functionality but disable the watermarks from given suspect models. However, WE techniques have certain limitations, including tampering with the training process, potential side effects on model functionality, and vulnerability to watermark erasure attacks [9], [10], [26]. In contrast, the OT methods extract the intrinsic characteristics (fingerprints) of DNN models, making them non-invasive and more resilient to adaptive attacks [15], [23].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Systematic View of Leakage Risks in Deep Neural Network Systems

Hu¹,

Liang²,

Chen³

et al. 2021

Preprint

View full text Add to dashboard Cite

As deep neural networks (DNNs) continue their reach into a wide range of application domains, the neural network architecture of DNN models becomes an increasingly sensitive subject, due to either intellectual property protection or risks of adversarial attacks. In observing the large gap between the architectural surfaces exploration and the model integrity study, this paper first presents the formulated schema of the model leakage risks. Then, we propose DeepSniffer, a learning-based model extraction framework, to obtain the complete model architecture information without any prior knowledge of the victim model. It is robust to architectural and system noises introduced by the complex memory hierarchy and diverse run-time system optimizations. Taking GPU platforms as a showcase, DeepSniffer performs model extraction by learning both the architecture-level execution features of kernels and the inter-layer temporal association information introduced by the common practice of DNN design. We demonstrate that DeepSniffer works experimentally in the context of an off-the-shelf Nvidia GPU platform running a variety of DNN models. The extracted models are directly helpful to the attempting of crafting adversarial inputs. The DeepSniffer project has been released in https://github.com/xinghu7788/DeepSniffer.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Model Stealing Attacksmentioning

confidence: 99%

Section: Threat Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Systematic View of Leakage Risks in Deep Neural Network Systems

Hu¹,

Liang²,

Chen³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

Voice spoofing detection using a neural networks assembly considering spectrograms and mel frequency cepstral coefficients

Hernández-Nava,

Rincón-García,

Lara-Velázquez

et al. 2023

PeerJ Computer Science

View full text Add to dashboard Cite

Nowadays, biometric authentication has gained relevance due to the technological advances that have allowed its inclusion in many daily-use devices. However, this same advantage has also brought dangers, as spoofing attacks are now more common. This work addresses the vulnerabilities of automatic speaker verification authentication systems, which are prone to attacks arising from new techniques for the generation of spoofed audio. In this article, we present a countermeasure for these attacks using an approach that includes easy to implement feature extractors such as spectrograms and mel frequency cepstral coefficients, as well as a modular architecture based on deep neural networks. Finally, we evaluate our proposal using the well-know ASVspoof 2017 V2 database, the experiments show that using the final architecture the best performance is obtained, achieving an equal error rate of 6.66% on the evaluation set.

show abstract

A Simple Neural Network Based Countermeasure for Replay Attack

Cited by 2 publications

References 14 publications

A Systematic View of Leakage Risks in Deep Neural Network Systems

A Systematic View of Leakage Risks in Deep Neural Network Systems

Voice spoofing detection using a neural networks assembly considering spectrograms and mel frequency cepstral coefficients

Contact Info

Product

Resources

About