paper-with-me

Papers

Nearest is Not Dearest: Towards Practical Defense against Quantization-conditioned Backdoor Attacks

2024-05-21 · CVPR 2024 1 · Boheng Li, Yishuo Cai, Haowei Li, Feng Xue, Zhifeng Li, Yiming Li

Model quantization is widely used to compress and accelerate deep neural networks. However, recent studies have revealed the feasibility of weaponizing model quantization via implanting quantization-conditioned backdoors (QCBs). These special backdoors stay dormant on released full-precision models but will come into effect after standard quantization. Due to the peculiarity of QCBs, existing defenses have minor effects on reducing their threats or are even infeasible. In this paper, we conduct the first in-depth analysis of QCBs. We reveal that the activation of existing QCBs primarily stems from the nearest rounding operation and is closely related to the norms of neuron-wise truncation errors (i.e., the difference between the continuous full-precision weights and its quantized version). Motivated by these insights, we propose Error-guided Flipped Rounding with Activation Preservation (EFRAP), an effective and practical defense against QCBs. Specifically, EFRAP learns a non-nearest rounding strategy with neuron-wise error norm and layer-wise activation preservation guidance, flipping the rounding strategies of neurons crucial for backdoor effects but with minimal impact on clean accuracy. Extensive evaluations on benchmark datasets demonstrate that our EFRAP can defeat state-of-the-art QCB attacks under various settings. Code is available at https://github.com/AntigoneRandy/QuantBackdoor_EFRAP.

📄 PDF Abstract BibTeX arXiv:2405.12725

Code (1)

antigonerandy/quantbackdoor_efrap 공식 구현 pytorch

Tasks

Quantization

Similar Papers 제목 키워드 기반

Defense Against Adversarial Images using Web-Scale Nearest-Neighbor Search

2019-03-05 · CVPR 2019 6 · Abhimanyu Dubey, Laurens van der Maaten, Zeki Yalniz, Yixuan Li 외

A plethora of recent work has shown that convolutional networks are not robust to adversarial images: images that are created by perturbing a sample from the data distribution as to maximize the loss on the perturbed exa…

Training-Time Attacks against k-Nearest Neighbors

2022-08-15 · Ara Vartanian, Will Rosenbaum, Scott Alfeld

Nearest neighbor-based methods are commonly used for classification tasks and as subroutines of other data-analysis methods. An attacker with the capability of inserting their own data points into the training set can ma…

ClassificationDimensionality Reduction

QuSecNets: Quantization-based Defense Mechanism for Securing Deep Neural Network against Adversarial Attacks

2018-11-04 · Faiq Khalid, Hassan Ali, Hammad Tariq, Muhammad Abdullah Hanif 외

Adversarial examples have emerged as a significant threat to machine learning algorithms, especially to the convolutional neural networks (CNNs). In this paper, we propose two quantization-based defense mechanisms, Const…

Quantization

Breaking the Limits of Quantization-Aware Defenses: QADT-R for Robustness Against Patch-Based Adversarial Attacks in QNNs

2025-03-10 · Amira Guesmi, Bassem Ouni, Muhammad Shafique

Quantized Neural Networks (QNNs) have emerged as a promising solution for reducing model size and computational costs, making them well-suited for deployment in edge and resource-constrained environments. While quantizat…

Quantization

Certified Robustness of Nearest Neighbors against Data Poisoning and Backdoor Attacks

2020-12-07 · Jinyuan Jia, Yupei Liu, Xiaoyu Cao, Neil Zhenqiang Gong

Data poisoning attacks and backdoor attacks aim to corrupt a machine learning classifier via modifying, adding, and/or removing some carefully selected training examples, such that the corrupted classifier makes incorrec…

Data Poisoning