Filters








25,948 Hits in 5.1 sec

Adversarial Attacks and Defenses: An Interpretation Perspective [article]

Ninghao Liu, Mengnan Du, Ruocheng Guo, Huan Liu, Xia Hu
2020 arXiv   pre-print
In this paper, we review recent work on adversarial attacks and defenses, particularly from the perspective of machine learning interpretation.  ...  The goal of model interpretation, or interpretable machine learning, is to extract human-understandable terms for the working mechanism of models.  ...  attack and defense approaches from the perspective of interpretable machine learning.  ... 
arXiv:2004.11488v2 fatcat:r4nhbnc72ng67bzv4o7ce5ohxu

Adversarial machine learning

Ling Huang, Anthony D. Joseph, Blaine Nelson, Benjamin I.P. Rubinstein, J. D. Tygar
2011 Proceedings of the 4th ACM workshop on Security and artificial intelligence - AISec '11  
In this paper (expanded from an invited talk at AISEC 2010), we discuss an emerging field of study: adversarial machine learning-the study of effective machine learning techniques against an adversarial  ...  for modeling an adversary's capabilities; explore the limits of an adversary's knowledge about the algorithm, feature space, training, and input data; explore vulnerabilities in machine learning algorithms  ...  machine learning.  ... 
doi:10.1145/2046684.2046692 dblp:conf/ccs/HuangJNRT11 fatcat:d6wcto4tmvbbrec35cjdengxby

Adversarial Machine Learning

J.D. Tygar
2011 IEEE Internet Computing  
In this paper (expanded from an invited talk at AISEC 2010), we discuss an emerging field of study: adversarial machine learning-the study of effective machine learning techniques against an adversarial  ...  for modeling an adversary's capabilities; explore the limits of an adversary's knowledge about the algorithm, feature space, training, and input data; explore vulnerabilities in machine learning algorithms  ...  machine learning.  ... 
doi:10.1109/mic.2011.112 fatcat:wb3bt4r67zd4teikanwo6nfzba

Law and Adversarial Machine Learning [article]

Ram Shankar Siva Kumar, David R. O'Brien, Kendra Albert, Salome Vilojen
2018 arXiv   pre-print
When machine learning systems fail because of adversarial manipulation, how should society expect the law to respond?  ...  learning in the context of civil liberties.  ...  Acknowledgments An interdisciplinary paper such as this would not have been possible without fruitful discussions and feedback from ML researchers (Aleksandr Madry, Momin Malik, Gretchen Greene, Sharon  ... 
arXiv:1810.10731v3 fatcat:ylgab2xk3zaivgmaixb7afszaa

Mental Models of Adversarial Machine Learning [article]

Lukas Bieringer, Kathrin Grosse, Michael Backes, Battista Biggio, Katharina Krombholz
2022 arXiv   pre-print
Jointly with our additional findings, these two facets provide a foundation to substantiate mental models for machine learning security and have implications for the integration of adversarial machine  ...  Firstly, practitioners often confuse machine learning security with threats and defences that are not directly related to machine learning.  ...  Introduction Adversarial machine learning (AML) studies the reliability of learning based systems in the context of an adversary [6, 12, 69] .  ... 
arXiv:2105.03726v4 fatcat:p3hgu7vcp5advkjkzprgvifwpu

Adversarial Attacks on Machine Learning Systems for High-Frequency Trading [article]

Micah Goldblum, Avi Schwarzschild, Ankit B. Patel, Tom Goldstein
2020 arXiv   pre-print
We study valuation models for algorithmic trading from the perspective of adversarial machine learning.  ...  Finally, we investigate the feasibility of realistic adversarial attacks in which an adversarial trader fools automated trading systems into making inaccurate predictions.  ...  INTRODUCTION Machine learning serves an increasingly large role in financial applications.  ... 
arXiv:2002.09565v3 fatcat:wcm3e3idvnhhppf2eq6fgkkbtm

Adversarial Machine Learning Phases of Matter [article]

Si Jiang, Sirui Lu, Dong-Ling Deng
2021 arXiv   pre-print
We study the robustness of machine learning approaches to adversarial perturbations, with a focus on supervised learning scenarios.  ...  Our results provide valuable guidance for both theoretical and experimental future studies on applying machine learning techniques to condensed matter physics.  ...  Machine learning, which has achieved dramatic success recently in a broad range of artificial intelligence applications, may bring an unprecedented perspective for this challenging task.  ... 
arXiv:1910.13453v2 fatcat:n2jihtiy3rer7flytuoi54icpu

Explaining Vulnerabilities to Adversarial Machine Learning through Visual Analytics

Yuxin Ma, Tiankai Xie, Jundong Li, Ross Maciejewski
2019 IEEE Transactions on Visualization and Computer Graphics  
As the deployment of artificial intelligence technologies becomes ubiquitous, it is unsurprising that adversaries have begun developing methods to manipulate machine learning models to their advantage.  ...  Machine learning models are currently being deployed in a variety of real-world applications where model predictions are used to make decisions about healthcare, bank loans, and numerous other critical  ...  Goal: In adversarial machine learning, an attacker's goal can be separated into two major categories: targeted attacks and reliability attacks.  ... 
doi:10.1109/tvcg.2019.2934631 pmid:31478859 fatcat:u36fkkxspjdldbys67jnizc5nm

Security Matters: A Survey on Adversarial Machine Learning [article]

Guofu Li, Pengjia Zhu, Jin Li, Zhemin Yang, Ning Cao, Zhiyi Chen
2018 arXiv   pre-print
Adversarial machine learning is a fast growing research area, which considers the scenarios when machine learning systems may face potential adversarial attackers, who intentionally synthesize input data  ...  The earliest studies on the adversarial examples for machine learning algorithms start from the information security area, which considers a much wider varieties of attacking methods.  ...  However, both class of machine learning algorithms are fundamentally based on function fitting. From the RL's perspective, Huang et al.  ... 
arXiv:1810.07339v2 fatcat:dbqngduvdnbyxchvxejvm6iq2m

The Dimpled Manifold Model of Adversarial Examples in Machine Learning [article]

Adi Shamir, Odelia Melamed, Oriel BenShmuel
2022 arXiv   pre-print
adversarial training is just to deepen the generated dimples in the decision boundary.  ...  Finally, we discuss and demonstrate the very different properties of on-manifold and off-manifold adversarial perturbations.  ...  The best known alternative perspective is the one proposed in Ilyas 8 Open problems Table 2 : 2 The mean distance of an adversarial example In Table2one can see a big difference in the adversarial distances  ... 
arXiv:2106.10151v2 fatcat:svwrkeqvnjcb5poov6daqadxy4

Randomized Prediction Games for Adversarial Machine Learning

Samuel Rota Bulo, Battista Biggio, Ignazio Pillai, Marcello Pelillo, Fabio Roli
2017 IEEE Transactions on Neural Networks and Learning Systems  
Recent work has proposed game-theoretical formulations to learn secure classifiers, by simulating different evasion attacks and modifying the classification function accordingly.  ...  Interestingly, randomization has also been proposed to improve security of learning algorithms against evasion attacks, as it results in hiding information about the classifier to the attacker.  ...  Evaluating security of machine learning against such attacks and devising suitable countermeasures, are two among the main open issues under investigation in the field of adversarial machine learning  ... 
doi:10.1109/tnnls.2016.2593488 pmid:27514067 fatcat:xwntvgy5ajhq3j7mjgcdsldjke

Robust in Practice: Adversarial Attacks on Quantum Machine Learning [article]

Haoran Liao, Ian Convy, William J. Huggins, K. Birgitta Whaley
2021 arXiv   pre-print
A more severe vulnerability has been noted for quantum machine learning (QML) models classifying Haar-random pure states.  ...  In order to provide insights into the adversarial robustness of a quantum classifier on real-world classification tasks, we focus on the adversarial robustness in classifying a subset of encoded states  ...  INTRODUCTION Quantum machine learning (QML) protocols, by exploiting quantum mechanics principles, such as superposition, tunneling, and entanglement [1] , have given hope of outperforming their classical  ... 
arXiv:2010.08544v2 fatcat:vitmg7e3m5hldgishykus3c2t4

Adversarial Machine Learning in Text Analysis and Generation [article]

Izzat Alsmadi
2021 arXiv   pre-print
The research field of adversarial machine learning witnessed a significant interest in the last few years.  ...  This paper focuses on studying aspects and research trends in adversarial machine learning specifically in text analysis and generation.  ...  machine learning.  ... 
arXiv:2101.08675v1 fatcat:73b3v35oebefnhzuuuo52jpdtu

Wild patterns: Ten years after the rise of adversarial machine learning

Battista Biggio, Fabio Roli
2018 Pattern Recognition  
adversarial machine learning.  ...  The vulnerability of machine learning to such wild patterns (also referred to as adversarial examples), along with the design of suitable countermeasures, have been investigated in the research field of  ...  (see [21, 38] and references therein) were providing an initial overview of the vulnerabilities of machine learning from a more general perspective, highlighting the need for adversarial machine learning  ... 
doi:10.1016/j.patcog.2018.07.023 fatcat:adgnesv7rrarjptsxxqa7t6cr4

Simulation-based Adversarial Test Generation for Autonomous Vehicles with Machine Learning Components [article]

Cumhur Erkan Tuncali, Georgios Fainekos, Hisahiro Ito, James Kapinski
2019 arXiv   pre-print
One of the main challenges is that many autonomous driving systems have machine learning components, such as deep neural networks, for which formal properties are difficult to characterize.  ...  We demonstrate how the framework can be used to evaluate closed-loop properties of an autonomous driving system model that includes the ML components, all within a virtual environment.  ...  One reason for this is that the current designs usually include some machine learning (ML) components, such as deep neural networks (DNNs), which are notoriously difficult to test and verify.  ... 
arXiv:1804.06760v4 fatcat:jxy5mrzrjjhvferrzm7jltncne
« Previous Showing results 1 — 15 out of 25,948 results