1,082 Hits in 9.4 sec

Cleaning up the neighborhood: A full classification for adversarial partial monitoring [article]

Tor Lattimore, Csaba Szepesvari
2018 arXiv   pre-print
We complete the classification of finite adversarial partial monitoring to include all games, solving an open problem posed by Bartok et al. [2014].  ...  Partial monitoring is a generalization of the well-known multi-armed bandit framework where the loss is not directly observed by the learner.  ...  In contrast to bandit and full information problems the loss in partial monitoring is not observed by the learner, even for the action played.  ... 
arXiv:1805.09247v1 fatcat:m66zgmsjlvfdffjgbbmgrb2qea

A Robust Context-Based Deep Learning Approach for Highly Imbalanced Hyperspectral Classification

Juan F. Ramirez Rochac, Nian Zhang, Lara A. Thompson, Tolessa Deksissa, Anastasios D. Doulamis
2021 Computational Intelligence and Neuroscience  
Our robust context-based DCN was able to outperform others in the presence of noise and was able to maintain a comparable classification accuracy on clean hyperspectral images.  ...  In this paper, we proposed a feature augmentation approach to increase noise resistance in imbalanced hyperspectral classification.  ...  Acknowledgments is work was supported in part by the National Science  ... 
doi:10.1155/2021/9923491 fatcat:okbqiqownzbp7gqwy6kxd4575y

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

David J. Miller, Zhen Xiang, George Kesidis
2019 arXiv   pre-print
There is great potential for damage from adversarial learning (AL) attacks on machine-learning based systems.  ...  ; 4) validity of the universal assumption that a TTE attacker knows the ground-truth class for the example to be attacked; 5) black, grey, or white box attacks as the standard for defense evaluation; 6  ...  This research was supported in part by an AFOSR DDDAS grant, a Cisco Systems URP gift, and an AWS credits gift.  ... 
arXiv:1904.06292v3 fatcat:dguztg5w5neirgggg5irh6doci

Adversarial Machine Learning in Wireless Communications using RF Data: A Review [article]

Damilola Adesina, Chung-Chu Hsieh, Yalin E. Sagduyu, Lijun Qian
2021 arXiv   pre-print
This paper presents a comprehensive review of the latest research efforts focused on AML in wireless communications while accounting for the unique characteristics of wireless systems.  ...  However, ML in general and DL in particular have been found vulnerable to manipulations thus giving rise to a field of study called adversarial machine learning (AML).  ...  that the results are in an -neighborhood of the original input data.  ... 
arXiv:2012.14392v2 fatcat:4d3x2scwjvh33drc745mmc4gvy

Design and analysis of a social botnet

Yazan Boshmaf, Ildar Muslukhov, Konstantin Beznosov, Matei Ripeanu
2013 Computer Networks  
Usually, an adversary starts off by running an infiltration campaign using hijacked or adversary-owned OSN accounts, with an objective to connect with a large number of users in the targeted OSN.  ...  We show that it is not difficult for an adversary to establish arbitrarily many social connections with arbitrary users in OSNs such as Facebook.  ...  the adversary has a knowledge of the used classification techniques.  ... 
doi:10.1016/j.comnet.2012.06.006 fatcat:56yxc3yspngmheu4bkhdj6i4xi

Information Directed Sampling for Linear Partial Monitoring [article]

Johannes Kirschner, Tor Lattimore, Andreas Krause
2020 arXiv   pre-print
We introduce information directed sampling (IDS) for stochastic partial monitoring with a linear reward and observation structure.  ...  Partial monitoring is a rich framework for sequential decision making under uncertainty that generalizes many well known bandit models, including linear, combinatorial and dueling bandits.  ...  Acknowledgments This project has received funding from the European Research Council (ERC) under the European Unions Horizon 2020 research and innovation programme grant agreement No 815943.  ... 
arXiv:2002.11182v1 fatcat:hrrttli2fng4dofheg23vsoyq4

MetaPoison: Practical General-purpose Clean-label Data Poisoning [article]

W. Ronny Huang, Jonas Geiping, Liam Fowl, Gavin Taylor, Tom Goldstein
2021 arXiv   pre-print
MetaPoison can achieve arbitrary adversary goals -- like using poisons of one class to make a target image don the label of another arbitrarily chosen class.  ...  MetaPoison is effective: it outperforms previous clean-label poisoning methods by a large margin.  ...  Acknowledgments supplied necessary tools for monitoring and logging of our large number of experiments and datasets and graciously provided storage and increased bandwidth for the unique requirements  ... 
arXiv:2004.00225v2 fatcat:a3fms4d3lbcpxkzniagesxd63m

Certification of embedded systems based on Machine Learning: A survey [article]

Guillaume Vidot
2021 arXiv   pre-print
This article provides an overview of the main challenges raised by the use ML in the demonstration of compliance with regulation requirements, and a survey of literature relevant to these challenges, with  ...  Advances in machine learning (ML) open the way to innovating functions in the avionic domain, such as navigation/surveillance assistance (e.g. vision-based navigation, obstacle sensing, virtual sensing  ...  Hence, at the end of the backpropagation, we end up with the importance of each feature for a given prediction.  ... 
arXiv:2106.07221v2 fatcat:rm7dapri7jfltd2ek2hpkhedka

Robustness of Graph Neural Networks at Scale [article]

Simon Geisler, Tobias Schmidt, Hakan Şirin, Daniel Zügner, Aleksandar Bojchevski, Stephan Günnemann
2021 arXiv   pre-print
We show that common surrogate losses are not well-suited for global attacks on GNNs. Our alternatives can double the attack strength.  ...  We propose two sparsity-aware first-order optimization attacks that maintain an efficient representation despite optimizing over a number of parameters which is quadratic in the number of nodes.  ...  Acknowledgments and Disclosure of Funding This research was supported by the Helmholtz Association under the joint research school "Munich School for Data Science -MUDS".  ... 
arXiv:2110.14038v3 fatcat:umiz3dcl4bcndkszki64saxu4m

Achieving Adversarial Robustness via Sparsity [article]

Shufan Wang, Ningyi Liao, Liyao Xiang, Nanyang Ye, Quanshi Zhang
2020 arXiv   pre-print
Through experiments on a variety of adversarial pruning methods, we find that weights sparsity will not hurt but improve robustness, where both weights inheritance from the lottery ticket and adversarial  ...  However, how the pruning process affects a network's robustness and the working mechanism behind remain unresolved.  ...  Since there is a tradeoff between accuracy and robustness , and some models tend to sacrifice one for the other, we choose to report the performance where the sum of adversarial accuracy and clean accuracy  ... 
arXiv:2009.05423v1 fatcat:lhfd77lukfaqzcisnqniqit7te

A Survey of Label-noise Representation Learning: Past, Present and Future [article]

Bo Han, Quanming Yao, Tongliang Liu, Gang Niu, Ivor W. Tsang, James T. Kwok, Masashi Sugiyama
2021 arXiv   pre-print
Classical machine learning implicitly assumes that labels of the training data are sampled from a clean distribution, which can be too restrictive for real-world scenarios.  ...  We first clarify a formal definition for LNRL from the perspective of machine learning.  ...  First, the model W I predicts the true label functionỹ(x) for all input x that lie within ε 0 neighborhood of a cluster center {c k } K k=1 .  ... 
arXiv:2011.04406v2 fatcat:76np6wyzvvag7ehy23cwyzdozm

Threats to Training: A Survey of Poisoning Attacks and Defenses on Machine Learning Systems

Zhibo Wang, Jingjing Ma, Xue Wang, Jiahui Hu, Zhan Qin, Kui Ren
2022 ACM Computing Surveys  
for adversaries to exploit.  ...  Machine learning (ML) has been universally adopted for automated decisions in a variety of fields, including recognition and classification applications, recommendation systems, natural language processing  ...  With regard to clean-label attacks, the proposed strategy in [66] large enough neighborhood around a data point.  ... 
doi:10.1145/3538707 fatcat:pcxpqbsrgzgidb7ngrcb5ggeoa

Classifier evaluation and attribute selection against active adversaries

Murat Kantarcıoğlu, Bowei Xi, Chris Clifton
2010 Data mining and knowledge discovery  
Hence a main assumption for the existing classification techniques no longer holds and initially successful classifiers degrade easily.  ...  We develop a game theoretic framework where equilibrium behavior of adversarial classification applications can be analyzed, and provide solutions for finding an equilibrium point.  ...  Acknowledgements We thank the reviewers and the editors for their helpful comments that improved the presentation and the content of the article.  ... 
doi:10.1007/s10618-010-0197-3 fatcat:afg6x2pbwbh4nkad2ccsuakzlm

Comprehensive Review of Deep Learning-Based 3D Point Cloud Completion Processing and Analysis [article]

Ben Fei, Weidong Yang, Wenming Chen, Zhijun Li, Yikang Li, Tao Ma, Xing Hu, Lipeng Ma
2022 arXiv   pre-print
Point cloud completion is a generation and estimation issue derived from the partial point clouds, which plays a vital role in the applications in 3D computer vision.  ...  Besides, this review sums up the commonly used datasets and illustrates the applications of point cloud completion.  ...  For 3D shape classification, the complete point clouds are ultimately needed by recovering from partial observations.  ... 
arXiv:2203.03311v2 fatcat:e2kvryolufearetp4ujlw2gwwy

Comparative Study of Traditional and Deep-Learning Denoising Approaches for Image-Based Petrophysical Characterization of Porous Media

Miral S. Tawfik, Amogh Subbakrishna Adishesha, Yuhan Hsi, Prakash Purswani, Russell T. Johns, Parisa Shokouhi, Xiaolei Huang, Zuleima T. Karpyn
2022 Frontiers in Water  
cycle consistent generative adversarial network (CCGAN)—which require a clean reference (ground truth), as well as noise-to-noise (N2N) and noise-to-void (N2V)—which do not require a clean reference.  ...  N2N75, which is a newly proposed semi-supervised variation of the N2N model, where 75% of the clean reference data is used for training, shows very promising outcomes for both traditional denoising performance  ...  cycle consistent generative adversarial network (CCGAN)—which require a clean reference (ground truth), as well as noise-to-noise (N2N) and noise-to-void (N2V)—which do not require a clean reference.  ... 
doi:10.3389/frwa.2021.800369 fatcat:pb2sqqdddneefeylddpy4gfju4
« Previous Showing results 1 — 15 out of 1,082 results