Towards quantum enhanced adversarial robustness in machine learning

Published in

Nature Machine Intelligence(2023)

External Links:

Generate Graph DownloadPDF

TL;DR

An overview of recent developments in quantum adversarial machine learning is given, and key challenges and future research directions to advance the field are outlined.

Abstract

Machine learning algorithms are powerful tools for data-driven tasks such as image classification and feature detection. However, their vulnerability to adversarial examples—input samples manipulated to fool the algorithm—remains a serious challenge. The integration of machine learning with quantum computing has the potential to yield tools offering not only better accuracy and computational efficiency, but also superior robustness against adversarial attacks. Indeed, recent work has employed quantum-mechanical phenomena to defend against adversarial attacks, spurring the rapid development of the field of quantum adversarial machine learning (QAML) and potentially yielding a new source of quantum advantage. Despite promising early results, there remain challenges in building robust real-world QAML tools. In this Perspective, we discuss recent progress in QAML and identify key challenges. We also suggest future research directions that could determine the route to practicality for QAML approaches as quantum computing hardware scales up and noise levels are reduced. To fulfil the potential of quantum machine learning for practical applications in the near future, it needs to be robust against adversarial attacks. West and colleagues give an overview of recent developments in quantum adversarial machine learning, and outline key challenges and future research directions to advance the field.

Authors

Maxwell T. West

1 papers

S. Tsang

1 papers

J. S. Low

1 papers

References98 items

Benchmarking Adversarially Robust Quantum Machine Learning at Scale

Suppressing quantum errors by scaling a surface code logical qubit

Quantum Neural Network Classifiers: A Tutorial

Data compression for quantum machine learning

Experimental quantum adversarial learning with programmable superconducting qubits

Towards quantum enhanced adversarial robustness in machine learning

Published in

Nature Machine Intelligence(2023)

External Links:

Generate Graph DownloadPDF

TL;DR

An overview of recent developments in quantum adversarial machine learning is given, and key challenges and future research directions to advance the field are outlined.

Abstract

Authors

Maxwell T. West

1 papers

S. Tsang

1 papers

J. S. Low

1 papers

References98 items

Benchmarking Adversarially Robust Quantum Machine Learning at Scale

Suppressing quantum errors by scaling a surface code logical qubit

Quantum Neural Network Classifiers: A Tutorial

Data compression for quantum machine learning

Experimental quantum adversarial learning with programmable superconducting qubits

L. Hollenberg

1 papers

Muhammad Usman

1 papers

Generative Adversarial Networks

Quantum advantage in learning from experiments

Generalization in quantum machine learning from few training data

A scalable and fast artificial neural network syndrome decoder for surface codes

Encoding-dependent generalization bounds for parametrized quantum circuits

Universal adversarial examples and perturbations for quantum classifiers

Recent Advances in Adversarial Training for Adversarial Robustness

Quantum machine learning models are kernel methods

Machine learning of high dimensional data on a noisy quantum processor

Defence against adversarial attacks using classical and quantum-enhanced Boltzmann machines

Demonstration of non-Markovian process characterisation and control on a quantum processor

Robust in practice: Adversarial attacks on quantum machine learning

A rigorous and robust quantum speed-up in supervised machine learning

Optimal provable robustness of quantum classification via quantum hypothesis testing

Robustness Verification of Quantum Classifiers

Formal Analysis and Redesign of a Neural Network-Based Aircraft Taxiing System with VerifAI

Cost-optimal single-qubit gate synthesis in the Clifford hierarchy

Verification of Deep Convolutional Neural Networks Using ImageStars

Quantum noise protects quantum classifiers against adversaries

Robust data encodings for quantum classifiers

Adversarial Learning Targeting Deep Neural Network Classification: A Comprehensive Review of Defenses Against Attacks

Fast is better than free: Revisiting adversarial training

Quantum Adversarial Machine Learning

An Abstraction-Based Framework for Neural Network Verification

Adversarial machine learning phases of matter

Experimental exploration of five-qubit quantum error-correcting code with superconducting qubits

Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers

Vulnerability of quantum classification to adversarial perturbations

Adversarial Examples Are Not Bugs, They Are Features

Transfer of Adversarial Robustness Between Perturbation Types

Adversarial Learning in Statistical Classification: A Comprehensive Review of Defenses Against Attacks

Quanvolutional neural networks: powering image recognition with quantum circuits

Quantum Generative Adversarial Networks for learning and loading random distributions

Training deep quantum neural networks

Certified Adversarial Robustness via Randomized Smoothing

Adversarial machine learning

Theoretically Principled Trade-off between Robustness and Accuracy

Making machine learning robust against adversarial inputs

Robust Physical-World Attacks on Deep Learning Visual Classification

Scaling provable adversarial defenses

Robustness May Be at Odds with Accuracy

Supervised learning with quantum-enhanced feature spaces

Quantum Generative Adversarial Learning.

Quantum generative adversarial networks

Barren plateaus in quantum neural network training landscapes

Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Models

Thermometer Encoding: One Hot Way To Resist Adversarial Examples

Certified Robustness to Adversarial Examples with Differential Privacy

Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples

Certified Defenses against Adversarial Examples

Generating Adversarial Examples with Adversarial Networks

One Pixel Attack for Fooling Deep Neural Networks

Generative Adversarial Networks: An Overview

ZOO: Zeroth Order Optimization Based Black-box Attacks to Deep Neural Networks without Training Substitute Models

Differential Privacy in Quantum Computation

APE-GAN: Adversarial Perturbation Elimination with GAN

Towards Deep Learning Models Resistant to Adversarial Attacks

Adversarial Examples Are Not Easily Detected: Bypassing Ten Detection Methods

Detecting Adversarial Samples from Artifacts

Generating Adversarial Malware Examples for Black-Box Attacks Based on GAN

Quantum autoencoders for efficient compression of quantum data

Quantum machine learning

Adversarial Machine Learning at Scale

Accessorize to a Crime: Real and Stealthy Attacks on State-of-the-Art Face Recognition

Safety Verification of Deep Neural Networks

Adversarial examples in the physical world

Unified framework for magic state distillation and multiqubit gate synthesis with reduced resource cost

An efficient magic state approach to small angle rotations

Demonstration of quantum advantage in machine learning

What Is Machine Learning

Explaining and Harnessing Adversarial Examples

Quantum decision tree classifier

Intriguing properties of neural networks

Evasion Attacks against Machine Learning at Test Time

Surface codes: Towards practical large-scale quantum computation

Differential Privacy: A Survey of Results

Statistical decision theory for quantum systems

In Data Mining and Big Data: 7th International Conference, DMBD 2022

a quantum information standpoint

Safety Verification for Deep Neural Networks with Provable Guarantees (Invited Paper)

Countering Adversarial Images using Input Transformations

GradientBased Learning Applied to Document Recognition

The concentration of measure phenomenon

Grammar Engineering for CCG using Ant and XSLT ∗

Gradient-based learning applied to document recognition

Detection Theory and Quantum Mechanics

Scaling IonQ's quantum computers: the roadmap

Our quantum computing journey (Quantumai)

Expanding the IBM Quantum roadmap to anticipate the future of quantum - centric supercomputing ( IBM , 2022 )

Field of Study

Computer SciencePhysics

Journal Information

Name

Nature Machine Intelligence

Page

581 - 589

Volume

Venue Information

Name

Nature Machine Intelligence

Type

journal

URL

https://www.nature.com/natmachintell/

Alternate Names

Nat Mach Intell

TL;DR

Abstract

Authors

References98 items

Benchmarking Adversarially Robust Quantum Machine Learning at Scale

Suppressing quantum errors by scaling a surface code logical qubit

Quantum Neural Network Classifiers: A Tutorial

Data compression for quantum machine learning

Experimental quantum adversarial learning with programmable superconducting qubits

TL;DR

Abstract

Authors

References98 items

Benchmarking Adversarially Robust Quantum Machine Learning at Scale

Suppressing quantum errors by scaling a surface code logical qubit

Quantum Neural Network Classifiers: A Tutorial

Data compression for quantum machine learning

Experimental quantum adversarial learning with programmable superconducting qubits

Generative Adversarial Networks

Quantum advantage in learning from experiments

Generalization in quantum machine learning from few training data

A scalable and fast artificial neural network syndrome decoder for surface codes

Encoding-dependent generalization bounds for parametrized quantum circuits

Universal adversarial examples and perturbations for quantum classifiers

Recent Advances in Adversarial Training for Adversarial Robustness

Quantum machine learning models are kernel methods

Machine learning of high dimensional data on a noisy quantum processor

Defence against adversarial attacks using classical and quantum-enhanced Boltzmann machines

Demonstration of non-Markovian process characterisation and control on a quantum processor

Robust in practice: Adversarial attacks on quantum machine learning

A rigorous and robust quantum speed-up in supervised machine learning

Optimal provable robustness of quantum classification via quantum hypothesis testing

Robustness Verification of Quantum Classifiers

Formal Analysis and Redesign of a Neural Network-Based Aircraft Taxiing System with VerifAI

Cost-optimal single-qubit gate synthesis in the Clifford hierarchy

Verification of Deep Convolutional Neural Networks Using ImageStars

Quantum noise protects quantum classifiers against adversaries

Robust data encodings for quantum classifiers

Adversarial Learning Targeting Deep Neural Network Classification: A Comprehensive Review of Defenses Against Attacks

Fast is better than free: Revisiting adversarial training

Quantum Adversarial Machine Learning

An Abstraction-Based Framework for Neural Network Verification

Adversarial machine learning phases of matter

Experimental exploration of five-qubit quantum error-correcting code with superconducting qubits

Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers

Vulnerability of quantum classification to adversarial perturbations

Adversarial Examples Are Not Bugs, They Are Features

Transfer of Adversarial Robustness Between Perturbation Types

Adversarial Learning in Statistical Classification: A Comprehensive Review of Defenses Against Attacks

Quanvolutional neural networks: powering image recognition with quantum circuits

Quantum Generative Adversarial Networks for learning and loading random distributions

Training deep quantum neural networks

Certified Adversarial Robustness via Randomized Smoothing

Adversarial machine learning

Theoretically Principled Trade-off between Robustness and Accuracy

Making machine learning robust against adversarial inputs

Robust Physical-World Attacks on Deep Learning Visual Classification

Scaling provable adversarial defenses

Robustness May Be at Odds with Accuracy

Supervised learning with quantum-enhanced feature spaces

Quantum Generative Adversarial Learning.

Quantum generative adversarial networks

Barren plateaus in quantum neural network training landscapes

Survey results

Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Models

Thermometer Encoding: One Hot Way To Resist Adversarial Examples

Certified Robustness to Adversarial Examples with Differential Privacy

Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples

Certified Defenses against Adversarial Examples

Generating Adversarial Examples with Adversarial Networks

One Pixel Attack for Fooling Deep Neural Networks

Generative Adversarial Networks: An Overview

ZOO: Zeroth Order Optimization Based Black-box Attacks to Deep Neural Networks without Training Substitute Models

Differential Privacy in Quantum Computation

APE-GAN: Adversarial Perturbation Elimination with GAN

Towards Deep Learning Models Resistant to Adversarial Attacks

Adversarial Examples Are Not Easily Detected: Bypassing Ten Detection Methods

Detecting Adversarial Samples from Artifacts

Generating Adversarial Malware Examples for Black-Box Attacks Based on GAN

Quantum autoencoders for efficient compression of quantum data