Filters








1,188 Hits in 10.3 sec

Toward Smart Security Enhancement of Federated Learning Networks [article]

Junjie Tan, Ying-Chang Liang, Nguyen Cong Luong, Dusit Niyato
2020 arXiv   pre-print
been proposed as a promising alternative paradigm to support the training of machine learning (ML) models.  ...  As traditional centralized learning networks (CLNs) are facing increasing challenges in terms of privacy preservation, communication overheads, and scalability, federated learning networks (FLNs) have  ...  If a vulnerable ED is under attack, the labels of training data will be flipped randomly. Moreover, a scale factor is imposed to magnify the poisoned model updates by 20 times.  ... 
arXiv:2008.08330v1 fatcat:7nnkpn45oncbnfl2ef5wfexzjq

Data Poisoning Attacks Against Federated Learning Systems [article]

Vale Tolpegin, Stacey Truex, Mehmet Emre Gursoy, Ling Liu
2020 arXiv   pre-print
Federated learning (FL) is an emerging paradigm for distributed training of large-scale deep neural networks in which participants' data remains on their own devices with only model updates being shared  ...  We additionally show that the attacks can be targeted, i.e., they have a large negative impact only on classes that are under attack.  ...  Label Flipping Attacks in Federated Learning We use a label flipping attack to implement targeted data poisoning in FL.  ... 
arXiv:2007.08432v2 fatcat:sod5gwqwdnaxzgui5j5a57rcta

Vulnerabilities in Federated Learning

Nader Bouacida, Prasant Mohapatra
2021 IEEE Access  
A new decentralized training paradigm, known as Federated Learning (FL), enables multiple clients located at different geographical locations to learn a machine learning model collaboratively without sharing  ...  We highlight the vulnerabilities sources, key attacks on FL, defenses, as well as their unique challenges, and discuss promising future research directions towards more robust FL.  ...  One typical example of a dirty-label poisoning attack is label-flipping. Fig. 2 highlights how an adversary might corrupt the trained model by flipping labels.  ... 
doi:10.1109/access.2021.3075203 doaj:5e62c955db514036939a1c65011f46b8 fatcat:viv7tij6cffnlev4l52wggkxfe

Byzantine-Resilient Federated Learning with Heterogeneous Data Distribution [article]

Saurav Prakash, Hanieh Hashemi, Yongqin Wang, Murali Annavaram, Salman Avestimehr
2022 arXiv   pre-print
, and perform secure aggregation for global model update.  ...  For mitigating Byzantine behaviors in federated learning (FL), most state-of-the-art approaches, such as Bulyan, tend to leverage the similarity of updates from the benign clients.  ...  Results for same value and label flip attacks are included in Section 8.  ... 
arXiv:2010.07541v4 fatcat:w6ujccz5kjdxresmixh6wj2iii

Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses [article]

Micah Goldblum, Dimitris Tsipras, Chulin Xie, Xinyun Chen, Avi Schwarzschild, Dawn Song, Aleksander Madry, Bo Li, Tom Goldstein
2021 arXiv   pre-print
models.  ...  The goal of this work is to systematically categorize and discuss a wide range of dataset vulnerabilities and exploits, approaches for defending against these threats, and an array of open problems in  ...  Additionally, support for Li and Xie was provided by NSF grant CCF-1910100 and the Amazon research award program.  ... 
arXiv:2012.10544v4 fatcat:2tpz6l2dpbgrjcyf5yxxv3pvii

Advances and Open Problems in Federated Learning [article]

Peter Kairouz, H. Brendan McMahan, Brendan Avent, Aurélien Bellet, Mehdi Bennis, Arjun Nitin Bhagoji, Kallista Bonawitz, Zachary Charles, Graham Cormode, Rachel Cummings, Rafael G.L. D'Oliveira, Hubert Eichner (+47 others)
2021 arXiv   pre-print
Federated learning (FL) is a machine learning setting where many clients (e.g. mobile devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g. service  ...  FL embodies the principles of focused data collection and minimization, and can mitigate many of the systemic privacy risks and costs resulting from traditional, centralized machine learning and data science  ...  Acknowledgments The authors would like to thank Alex Ingerman and David Petrou for their useful suggestions and insightful comments during the review process.  ... 
arXiv:1912.04977v3 fatcat:efkbqh4lwfacfeuxpe5pp7mk6a

Federated and Transfer Learning: A Survey on Adversaries and Defense Mechanisms [article]

Ehsan Hallaji, Roozbeh Razavi-Far, Mehrdad Saif
2022 arXiv   pre-print
The advent of federated learning has facilitated large-scale data exchange amongst machine learning models while maintaining privacy.  ...  One of the most significant advancements in this domain is the incorporation of transfer learning into federated learning, which overcomes fundamental constraints of primary federated learning, particularly  ...  Acknowledgements This work is supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) under funding reference numbers CGSD3-569341-2022 and RGPIN-2021-02968.  ... 
arXiv:2207.02337v1 fatcat:rf4fdiunnnehjpvjhbmncrt3ka

From Distributed Machine Learning to Federated Learning: A Survey [article]

Ji Liu, Jizhou Huang, Yang Zhou, Xuhong Li, Shilei Ji, Haoyi Xiong, Dejing Dou
2022 arXiv   pre-print
In this paper, we provide a comprehensive survey of existing works for federated learning. We propose a functional architecture of federated learning systems and a taxonomy of related techniques.  ...  Federated learning emerges as an efficient approach to exploit distributed data and computing resources, so as to collaboratively train machine learning models, while obeying the laws and regulations and  ...  The modification of the labels is denoted by the label flipping attack. As a result, the accuracy of the trained model has low accuracy in terms of Class C [165] .  ... 
arXiv:2104.14362v3 fatcat:sgip6r7vy5djpfapk74jygldni

A Systematic Review of Federated Learning in the Healthcare Area: From the Perspective of Data Properties and Applications

Prayitno, Chi-Ren Shyu, Karisma Trinanda Putra, Hsing-Chung Chen, Yuan-Yu Tsai, K. S. M. Tozammel Hossain, Wei Jiang, Zon-Yin Shae
2021 Applied Sciences  
Excellent deep learning models are heavily data-driven. The more data trained, the more robust and more generalizable the performance of the deep learning model.  ...  Federated learning resolves the previous challenges with a shared global deep learning model using a central aggregator server.  ...  Acknowledgments: This study was supported by the Ministry of Science and Technology of Taiwan under the project grants MOST 110-2321-B-468-001 and MOST 110-2511-H-468-005.  ... 
doi:10.3390/app112311191 fatcat:m6aq2o22cfbp5o3y4cgo7yk4au

AI for Beyond 5G Networks: A Cyber-Security Defense or Offense Enabler? [article]

C. Benzaid, T. Taleb
2022 arXiv   pre-print
envisioned to play a pivotal role in empowering intelligent, adaptive and autonomous security management in 5G and beyond networks, thanks to its potential to uncover hidden patterns from a large set of time-varying  ...  This strategy can be used against models that leverage distributed learning (e.g., federated learning), which relies on several agents for training.  ...  Serving Phase Training and Test Phase • The addition of noise to the execution time of the ML model. B.  ... 
arXiv:2201.02730v1 fatcat:upuk2pjcfzag5bjs5woeiwkxe4

Editorial Special Issue on AI Innovations in Intelligent Transportation Systems

Tai-Hoon Kim
2022 IEEE transactions on intelligent transportation systems (Print)  
It is shown to be robust against various potential attacks through detailed security analysis including the simulation-based formal security verification.  ...  propose a semi-supervised federated learning (SSFL) framework that can accurately identify travel modes without using users' raw trajectories data or relying on notable data labels.  ... 
doi:10.1109/tits.2022.3152067 fatcat:w5qyxfyp7zfzjckdkhsmddvzwm

Orchestrating the Development Lifecycle of Machine Learning-Based IoT Applications: A Taxonomy and Survey

Bin Qian, Jie Su, Zhenyu Wen, Devki Nandan Jha, Yinhao Li, Yu Guan, Deepak Puthal, Philip James, Renyu Yang, Albert Y. Zomaya, Omer Rana, Lizhe Wang (+2 others)
2020 ACM Computing Surveys  
For supervised cases when a class label is available for training, TML aims to map the input data to the labels by optimising a model, which can be used to infer unseen data at the test stage.  ...  To begin with modeling, it is essential for users to choose a suitable learning concept at the first stage.  ...  A label flipping attack attempts to add a noise label to the training data.  ... 
doi:10.1145/3398020 fatcat:zzgfcjxjxbhnhf53dmlo63rs3i

Edge Intelligence: Architectures, Challenges, and Applications [article]

Dianlei Xu, Tong Li, Yong Li, Xiang Su, Sasu Tarkoma, Tao Jiang, Jon Crowcroft, Pan Hui
2020 arXiv   pre-print
Edge intelligence refers to a set of connected systems and devices for data collection, caching, processing, and analysis in locations close to where data is captured based on artificial intelligence.  ...  We then aim for a systematic classification of the state of the solutions by examining research results and observations for each of the four components and present a taxonomy that includes practical problems  ...  IoT devices connect to the Internet through a gateway. They design two models for IoT device identification and anomaly detection.  ... 
arXiv:2003.12172v2 fatcat:xbrylsvb7bey5idirunacux6pe

Orchestrating the Development Lifecycle of Machine Learning-Based IoT Applications: A Taxonomy and Survey [article]

Bin Qian, Jie Su, Zhenyu Wen, Devki Nandan Jha, Yinhao Li, Yu Guan, Deepak Puthal, Philip James, Renyu Yang, Albert Y. Zomaya, Omer Rana, Lizhe Wang (+2 others)
2020 arXiv   pre-print
This paper provides a comprehensive and systematic survey on the development lifecycle of ML-based IoT application.  ...  sensors into ML models, thereby employing results to improve their business processes and services.  ...  Bonawitz against SVMs which utilizes label flipping to poison the training data [363] . A label flipping attack attempts to add a noise label to the training data.  ... 
arXiv:1910.05433v5 fatcat:ffvjipmylve6feuzdbav2syxfu

Towards Security Threats of Deep Learning Systems: A Survey [article]

Yingzhe He and Guozhu Meng and Kai Chen and Xingbo Hu and Jinwen He
2020 arXiv   pre-print
In particular, we focus on four types of attacks associated with security threats of deep learning: model extraction attack, model inversion attack, poisoning attack and adversarial attack.  ...  In order to unveil the security weaknesses and aid in the development of a robust deep learning system, we undertake an investigation on attacks towards deep learning, and analyze these attacks to conclude  ...  Poisoning Attack Approach Manipulating Mislabeled Data Learning model usually experiences training under labeled data in advance.  ... 
arXiv:1911.12562v2 fatcat:m3lyece44jgdbp6rlcpj6dz2gm
« Previous Showing results 1 — 15 out of 1,188 results