Adversarial Attacks and Defences: A Survey

Abstract

Deep learning has emerged as a strong and efficient framework that can be applied to a broad spectrum of complex learning problems which were difficult to solve using the traditional machine learning techniques in the past. In the last few years, deep learning has advanced radically in such a way that it can surpass human-level performance on a number of tasks. As a consequence, deep learning is being extensively used in most of the recent day-to-day applications. However, security of deep learning systems are vulnerable to crafted adversarial examples, which may be imperceptible to the human eye, but can lead the model to misclassify the output. In recent times, different types of adversaries based on their threat model leverage these vulnerabilities to compromise a deep learning system where adversaries have high incentives. Hence, it is extremely important to provide robustness to deep learning algorithms against these adversaries. However, there are only a few strong countermeasures which can be used in all types of attack scenarios to design a robust deep learning system. In this paper, we attempt to provide a detailed discussion on different types of adversarial attacks with various threat models and also elaborate the efficiency and challenges of recent countermeasures against them.

References in corpus (10)

• Distilling the Knowledge in a Neural Network
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Ensemble Adversarial Training: Attacks and Defenses
• Empirical Evaluation of Rectified Activations in Convolutional Network
• Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks
• Transferability in Machine Learning: from Phenomena to Black-Box Attacks using Adversarial Samples
• Data Poisoning Attacks on Factorization-Based Collaborative Filtering
• Blocking Transferability of Adversarial Examples in Black-Box Learning Systems
• Defending against Adversarial Images using Basis Functions Transformations
• A Survey on Resilient Machine Learning

Cited by in corpus (92)

• Adversarial Example Detection for DNN Models: A Review and Experimental Comparison
• Adversarial Machine Learning in Image Classification: A Survey Towards the Defender's Perspective
• Wild Patterns Reloaded: A Survey of Machine Learning Security against Training Data Poisoning
• Privacy in Deep Learning: A Survey
• Toward Open-World Electroencephalogram Decoding Via Deep Learning: A Comprehensive Survey
• BERT-ATTACK: Adversarial Attack Against BERT Using BERT
• Contrastive Learning with Adversarial Examples
• Adversarial Explanations for Understanding Image Classification Decisions and Improved Neural Network Robustness
• Robust in Practice: Adversarial Attacks on Quantum Machine Learning
• A Topology Layer for Machine Learning
• Arms Race in Adversarial Malware Detection: A Survey
• Towards Robust and Reliable Algorithmic Recourse
• Socially Responsible AI Algorithms: Issues, Purposes, and Challenges
• Universal Adversarial Examples and Perturbations for Quantum Classifiers
• Tree tensor network classifiers for machine learning: from quantum-inspired to quantum-assisted
• Adversarial Example in Remote Sensing Image Recognition
• Code Prediction by Feeding Trees to Transformers
• Learning Disentangled Behaviour Patterns for Wearable-based Human Activity Recognition
• Opportunities and Challenges in Deep Learning Adversarial Robustness: A Survey
• A survey on Adversarial Recommender Systems: from Attack/Defense strategies to Generative Adversarial Networks
• Dynamic Adversarial Patch for Evading Object Detection Models
• Enabling AutoML for Zero-Touch Network Security: Use-Case Driven Analysis
• The RFML Ecosystem: A Look at the Unique Challenges of Applying Deep Learning to Radio Frequency Applications
• Probabilistic Margins for Instance Reweighting in Adversarial Training
• White Paper Machine Learning in Certified Systems
• ConAML: Constrained Adversarial Machine Learning for Cyber-Physical Systems
• Adversarial Machine Learning in Wireless Communications using RF Data: A Review
• A case for new neural network smoothness constraints
• Adversarial Attack for Uncertainty Estimation: Identifying Critical Regions in Neural Networks
• Chat as Expected: Learning to Manipulate Black-box Neural Dialogue Models
• Security and Privacy for Artificial Intelligence: Opportunities and Challenges
• Adversarial Example Games
• Post-breach Recovery: Protection against White-box Adversarial Examples for Leaked DNN Models
• Adversarial Learning of Deepfakes in Accounting
• Identifying Audio Adversarial Examples via Anomalous Pattern Detection
• Key Protected Classification for Collaborative Learning
• Manipulating a Learning Defender and Ways to Counteract
• Information Laundering for Model Privacy
• Generating Adversarial Examples in Chinese Texts Using Sentence-Pieces
• Query Attack via Opposite-Direction Feature:Towards Robust Image Retrieval
• Data from Model: Extracting Data from Non-robust and Robust Models
• Adversarial Machine Learning Phases of Matter
• Towards Logical Specification of Statistical Machine Learning
• Adversarial Security Attacks and Perturbations on Machine Learning and Deep Learning Methods
• A Tale of Two Approximations: Tightening Over-Approximation for DNN Robustness Verification via Under-Approximation
• BASAR:Black-box Attack on Skeletal Action Recognition
• Simple Contrastive Representation Adversarial Learning for NLP Tasks
• Adversarial Attack on Network Embeddings via Supervised Network Poisoning
• From Shallow to Deep Interactions Between Knowledge Representation, Reasoning and Machine Learning (Kay R. Amel group)
• Resilience of Autonomous Vehicle Object Category Detection to Universal Adversarial Perturbations
• Estimating the Brittleness of AI: Safety Integrity Levels and the Need for Testing Out-Of-Distribution Performance
• Adversarial Transferability in Wearable Sensor Systems
• Semantic-preserving Reinforcement Learning Attack Against Graph Neural Networks for Malware Detection
• Fooling Adversarial Training with Inducing Noise
• Asymptotic Behavior of Adversarial Training in Binary Classification
• Reviews: Topological Distances and Losses for Brain Networks
• Meta Adversarial Perturbations
• Fine-grained Uncertainty Modeling in Neural Networks
• Drawing Robust Scratch Tickets: Subnetworks with Inborn Robustness Are Found within Randomly Initialized Networks
• Learning while Respecting Privacy and Robustness to Distributional Uncertainties and Adversarial Data
• Understanding the Intrinsic Robustness of Image Distributions using Conditional Generative Models
• Crafting Adversarial Examples for Deep Learning Based Prognostics (Extended Version)
• Ensemble-in-One: Learning Ensemble within Random Gated Networks for Enhanced Adversarial Robustness
• Fortify Machine Learning Production Systems: Detect and Classify Adversarial Attacks
• Enhancing the Antidote: Improved Pointwise Certifications against Poisoning Attacks
• Meta-Learning the Search Distribution of Black-Box Random Search Based Adversarial Attacks
• Enhance Diffusion to Improve Robust Generalization
• Generating Adversarial Inputs Using A Black-box Differential Technique
• Bio-inspired Robustness: A Review
• Medical Image Segmentation with Limited Supervision: A Review of Deep Network Models
• Generative Models for Security: Attacks, Defenses, and Opportunities
• Who is Responsible for Adversarial Defense?
• Adaptive Adversarial Training Does Not Increase Recourse Costs
• Attack Assessment and Augmented Identity Recognition for Human Skeleton Data
• Query complexity of adversarial attacks
• A Survey on Assessing the Generalization Envelope of Deep Neural Networks: Predictive Uncertainty, Out-of-distribution and Adversarial Samples
• Provable Training Set Debugging for Linear Regression
• Online Adversarial Attacks
• Broadly Applicable Targeted Data Sample Omission Attacks
• Feature Importance Guided Attack: A Model Agnostic Adversarial Attack
• Human Imperceptible Attacks and Applications to Improve Fairness
• Adversarial Attacks on Deep Models for Financial Transaction Records
• Poisoning the Search Space in Neural Architecture Search
• Spatially Correlated Patterns in Adversarial Images
• An Empirical Review of Adversarial Defenses
• Adversarial Attack Attribution: Discovering Attributable Signals in Adversarial ML Attacks
• Adversarial Machine Learning in Text Analysis and Generation
• Explainable Adversarial Attacks in Deep Neural Networks Using Activation Profiles
• A Review of Computer Vision Methods in Network Security
• Decoder-free Robustness Disentanglement without (Additional) Supervision
• An Adversarial Attack Defending System for Securing In-Vehicle Networks
• Disconnection-aware Attack Detection in Networked Control Systems