paper-with-me

Papers

An Embarrassingly Simple Backdoor Attack on Self-supervised Learning

2022-10-13 · ICCV 2023 1 · Changjiang Li, Ren Pang, Zhaohan Xi, Tianyu Du, Shouling Ji, Yuan YAO, Ting Wang

As a new paradigm in machine learning, self-supervised learning (SSL) is capable of learning high-quality representations of complex data without relying on labels. In addition to eliminating the need for labeled data, research has found that SSL improves the adversarial robustness over supervised learning since lacking labels makes it more challenging for adversaries to manipulate model predictions. However, the extent to which this robustness superiority generalizes to other types of attacks remains an open question. We explore this question in the context of backdoor attacks. Specifically, we design and evaluate CTRL, an embarrassingly simple yet highly effective self-supervised backdoor attack. By only polluting a tiny fraction of training data (<= 1%) with indistinguishable poisoning samples, CTRL causes any trigger-embedded input to be misclassified to the adversary's designated class with a high probability (>= 99%) at inference time. Our findings suggest that SSL and supervised learning are comparably vulnerable to backdoor attacks. More importantly, through the lens of CTRL, we study the inherent vulnerability of SSL to backdoor attacks. With both empirical and analytical evidence, we reveal that the representation invariance property of SSL, which benefits adversarial robustness, may also be the very reason making \ssl highly susceptible to backdoor attacks. Our findings also imply that the existing defenses against supervised backdoor attacks are not easily retrofitted to the unique vulnerability of SSL.

📄 PDF Abstract BibTeX arXiv:2210.07346

Code (4)

CCCjiang/CTRL 공식 구현 pytorch
meet-cjli/ctrl 공식 구현 pytorch
aryan-satpathy/backdoor pytorch
jsrdcht/SSL-Backdoor pytorch

Tasks

Adversarial RobustnessBackdoor AttackSelf-Supervised Learning

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
ReLU How Do I Communicate to Expedia? How Do I Communicate to Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Live Support & Special Travel…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…
Gradient Clipping One difficulty that arises with optimization of deep neural networks is that large parameter gradients can lead an SGD optimizer to update…
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…

Similar Papers 제목 키워드 기반

Towards Imperceptible Backdoor Attack in Self-supervised Learning

2024-05-23 · Hanrong Zhang, Zhenting Wang, Tingxu Han, Mingyu Jin 외

Self-supervised learning models are vulnerable to backdoor attacks. Existing backdoor attacks that are effective in self-supervised learning often involve noticeable triggers, like colored patches, which are vulnerable t…

Backdoor AttackSelf-Supervised Learning

Invisible Backdoor Attack against Self-supervised Learning

2025-01-01 · CVPR 2025 1 · Hanrong Zhang, Zhenting Wang, Boheng Li, Fulin Lin 외

Self-supervised learning (SSL) models are vulnerable to backdoor attacks. Existing backdoor attacks that are effective in SSL often involve noticeable triggers, like colored patches or visible noise, which are vulner…

Backdoor AttackSelf-Supervised Learning

Backdoor Attacks on Self-Supervised Learning

2021-05-21 · CVPR 2022 1 · Aniruddha Saha, Ajinkya Tejankar, Soroush Abbasi Koohpayegani, Hamed Pirsiavash

Large-scale unlabeled data has spurred recent progress in self-supervised learning methods that learn rich visual representations. State-of-the-art self-supervised methods for learning representations from images (e.g., …

Backdoor AttackInductive BiasKnowledge DistillationSelf-Supervised Learning

Training set cleansing of backdoor poisoning by self-supervised representation learning

2022-10-19 · H. Wang, S. Karami, O. Dia, H. Ritter 외

A backdoor or Trojan attack is an important type of data poisoning attack against deep neural network (DNN) classifiers, wherein the training dataset is poisoned with a small number of samples that each possess the backd…

Data Poisoningimage-classificationImage ClassificationRepresentation Learning

Towards Understanding How Self-training Tolerates Data Backdoor Poisoning

2023-01-20 · Soumyadeep Pal, Ren Wang, Yuguang Yao, Sijia Liu

Recent studies on backdoor attacks in model training have shown that polluting a small portion of training data is sufficient to produce incorrect manipulated predictions on poisoned test-time data while maintaining high…

backdoor defenseRepresentation Learning