Defenses in Adversarial Machine Learning: A Survey
Preprint
Abstract
Adversarial phenomenon has been widely observed in machine learning (ML) systems, especially in those using deep neural networks, describing that ML systems may produce inconsistent and incomprehensible predictions with humans at some particular cases. This phenomenon poses a serious security threat to the practical application of ML systems, and several advanced attack paradigms have been developed to explore it, mainly including backdoor attacks, weight attacks, and adversarial examples. For each individual attack paradigm, various defense paradigms have been developed to improve the model robustness against the corresponding attack paradigm. However, due to the independence and diversity of these defense paradigms, it is difficult to examine the overall robustness of an ML system against different kinds of attacks.This survey aims to build a systematic review of all existing defense paradigms from a unified perspective. Specifically, from the life-cycle perspective, we factorize a complete machine learning system into five stages, including pre-training, training, post-training, deployment, and inference stages, respectively. Then, we present a clear taxonomy to categorize and review representative defense methods at each individual stage. The unified perspective and presented taxonomies not only facilitate the analysis of the mechanism of each defense paradigm but also help us to understand connections and differences among different defense paradigms, which may inspire future research to develop more advanced, comprehensive defenses.Keywords
All Related Versions
This publication has 26 references indexed in Scilit:
- BadCLIP: Dual-Embedding Guided Backdoor Attack on Multimodal Contrastive Learning2024
- RobustMQ: benchmarking robustness of quantized modelsVisual Intelligence, 2023
- Randomized Adversarial Training via Taylor Expansion2023
- Effective, Efficient and Robust Neural Architecture Search2022
- A survey of robust adversarial training in pattern recognition: Fundamental, theory, and methodologiesPattern Recognition, 2022
- Defense against backdoor attack in federated learningComputers & Security, 2022
- Semi-supervised robust training with generalized perturbed neighborhoodPattern Recognition, 2022
- Backdoor attacks-resilient aggregation based on Robust Filtering of Outliers in federated learning for image classificationKnowledge-Based Systems, 2022
- SpectralDefense: Detecting Adversarial Attacks on CNNs in the Fourier Domain2021
- Neural network laundering: Removing black-box backdoor watermarks from deep neural networksComputers & Security, 2021