Filters








1,620 Hits in 5.3 sec

Byzantine-robust Federated Learning through Spatial-temporal Analysis of Local Model Updates [article]

Zhuohang Li, Luyang Liu, Jiaxin Zhang, Jian Liu
2021 arXiv   pre-print
However, Byzantine clients that send incorrect or disruptive updates due to system failures or adversarial attacks may disturb the joint learning process, consequently degrading the performance of the  ...  Federated Learning (FL) enables multiple distributed clients (e.g., mobile devices) to collaboratively train a centralized model while keeping the training data locally on the client.  ...  Federated Learning Federated Learning (FL) (or Collaborative Learning) is a distributed learning framework that allows multiple clients to collaboratively train a machine learning model under the coordination  ... 
arXiv:2107.01477v2 fatcat:e4zvglywcfbmbmi2kkyeohlwna

PIRATE: A Blockchain-based Secure Framework of Distributed Machine Learning in 5G Networks [article]

Sicong Zhou, Huawei Huang, Wuhui Chen, Zibin Zheng, Song Guo
2019 arXiv   pre-print
This is because the distributed learning system is prone to suffering from byzantine attacks during the stages of updating model parameters and aggregating gradients amongst multiple learning participants  ...  Finally, we also envision some open issues and challenges based on the proposed byzantine-resilient learning framework.  ...  30% attack High Medium Unknown High Resiliency under 30% attack (FL) Low Unknown High Unknown Resiliency under majority attack Low Medium Unknown High Resiliency under majority attack (  ... 
arXiv:1912.07860v1 fatcat:f77hexw7ybbb5nfhleujbaji5q

DeFL: Decentralized Weight Aggregation for Cross-silo Federated Learning [article]

Jialiang Han, Yudong Han, Gang Huang, Yun Ma
2022 arXiv   pre-print
Federated learning (FL) is an emerging promising paradigm of privacy-preserving machine learning (ML).  ...  However, the central server may be vulnerable to malicious attacks or software failures in practice.  ...  Therefore, the stability, fairness, and security of the central server are crucial to FL. Decentralized Federated Learning.  ... 
arXiv:2208.00848v1 fatcat:s64duytwufbu5ffijh4zwo7fzy

BARFED: Byzantine Attack-Resistant Federated Averaging Based on Outlier Elimination [article]

Ece Isik-Polat, Gorkem Polat, Altan Kocyigit
2021 arXiv   pre-print
Although many defense algorithms have recently been proposed to address these attacks, they often make strong assumptions that do not agree with the nature of federated learning, such as Non-IID datasets  ...  In federated learning, each participant trains its local model with its own data and a global model is formed at a trusted server by aggregating model updates coming from these participants.  ...  Attacks Table 5 : 5 Accuracies under Byzantine attacks at different attacker ratios in IID settings.  ... 
arXiv:2111.04550v1 fatcat:fjvje47d2jhs3nypskfnhisrny

Suppressing Poisoning Attacks on Federated Learning for Medical Imaging [article]

Naif Alkhunaizi, Dmitry Kamzolov, Martin Takáč, Karthik Nandakumar
2022 arXiv   pre-print
Federated Learning (FL) is a promising solution that enables collaborative training through exchange of model parameters instead of raw data.  ...  However, most existing FL solutions work under the assumption that participating clients are honest and thus can fail against poisoning attacks from malicious parties, whose goal is to deteriorate the  ...  Fig. 1 : 1 Fig. 1: Federated Learning (FL) in the presence of a malicious client.  ... 
arXiv:2207.10804v1 fatcat:xfkuofk3nfbiriepbem5s2sab4

A Decentralized Federated Learning Framework via Committee Mechanism with Convergence Guarantee [article]

Chunjiang Che, Xiaoli Li, Chuan Chen, Xiaoyu He, Zibin Zheng
2021 arXiv   pre-print
However, this distributed machine learning training method is prone to attacks from Byzantine clients, which interfere with the training of the global model by modifying the model or uploading the false  ...  Extensive experiments illustrate that CMFL achieves faster convergence and better accuracy than the typical Federated Learning, in the meanwhile obtaining better robustness than the traditional Byzantine-tolerant  ...  In this paper, we comprehensively consider Byzantine attacks of both clients and the central server, and design a serverless Federated Learning framework under Committee Mechanism (CMFL), in which some  ... 
arXiv:2108.00365v1 fatcat:w2t6pmftxfditnxqbmlgu46zci

Byzantine-Robust Federated Machine Learning through Adaptive Model Averaging [article]

Luis Muñoz-González, Kenneth T. Co, Emil C. Lupu
2019 arXiv   pre-print
Standard federated learning techniques are vulnerable to Byzantine failures, biased local datasets, and poisoning attacks.  ...  In this paper we introduce Adaptive Federated Averaging, a novel algorithm for robust federated learning that is designed to detect failures, attacks, and bad updates provided by participants in a collaborative  ...  In this paper we introduce Adaptive Federated Averaging (AFA) a novel approach for Byzantine-robust federated learning.  ... 
arXiv:1909.05125v1 fatcat:ngeyqqlqivfwdnavutq52eju5m

Privacy and Robustness in Federated Learning: Attacks and Defenses [article]

Lingjuan Lyu, Han Yu, Xingjun Ma, Chen Chen, Lichao Sun, Jun Zhao, Qiang Yang, Philip S. Yu
2022 arXiv   pre-print
Recently, federated learning (FL) has emerged as an alternative solution and continue to thrive in this new reality.  ...  Finally, we discuss promising future research directions towards robust and privacy-preserving federated learning.  ...  federated learning (HFL), vertically federated learning (VFL) and federated transfer learning (FTL) [17] .  ... 
arXiv:2012.06337v3 fatcat:f5aflxnsdrdcdf4kvoa6yzseqq

Byzantine Fault Tolerance in Distributed Machine Learning : a Survey [article]

Djamila Bouhata, Hamouma Moumen
2022 arXiv   pre-print
Byzantine Fault Tolerance (BFT) is among the most challenging problems in Distributed Machine Learning (DML).  ...  Byzantine failures are still difficult to tackle due to their unrestricted nature; as a result, the possibility of generating arbitrary data.  ...  This filter acts as a self-stabilizer by leveraging the Lipschitzness of cost functions to protect the SGD model from adversarial attacks.  ... 
arXiv:2205.02572v1 fatcat:h2hkcgz3w5cvrnro6whl2rpvby

Byzantine-Robust Learning on Heterogeneous Datasets via Bucketing [article]

Sai Praneeth Karimireddy, Lie He, Martin Jaggi
2022 arXiv   pre-print
In Byzantine robust distributed or federated learning, a central server wants to train a machine learning model over data distributed across multiple workers.  ...  Our work is the first to establish guaranteed convergence for the non-iid Byzantine robust problem under realistic assumptions.  ...  Bucketing against general Byzantine attacks. In Figure 1 , we present thorough experiments on non-iid data over 25 workers with 5 Byzantine workers, under different attacks.  ... 
arXiv:2006.09365v5 fatcat:aeqfsarizfgflkw2dxkwi44peu

Robust Federated Recommendation System [article]

Chen Chen, Jingfeng Zhang, Anthony K. H. Tung, Mohan Kankanhalli, Gang Chen
2020 arXiv   pre-print
In this paper, we develop a novel federated recommendation technique that is robust against the poisoning attack where Byzantine clients prevail.  ...  Theoretically, we justify our robust learning strategy by our proposed definition of Byzantine resilience.  ...  Local model poisoning attacks to byzantine-robust federated learning. arXiv preprint arXiv:1911.11815, 2019. [14] Mihajlo Grbovic and Haibin Cheng.  ... 
arXiv:2006.08259v1 fatcat:boav3q2s5zgv5o3u5v5sxya6ti

Federated Learning via Plurality Vote [article]

Kai Yue, Richeng Jin, Chau-Wai Wong, Huaiyu Dai
2022 arXiv   pre-print
The model parameters are aggregated via weighted voting to enhance the resilience against Byzantine attacks.  ...  Federated learning allows collaborative workers to solve a machine learning problem while preserving data privacy.  ...  INTRODUCTION Federated learning enables multiple workers to solve a machine learning problem under the coordination of a central server (Kairouz et al., 2021) .  ... 
arXiv:2110.02998v2 fatcat:k6jepi5slffitfco547bdco77a

Federated Learning: A Distributed Shared Machine Learning Method

Kai Hu, Yaogen Li, Min Xia, Jiasheng Wu, Meixia Lu, Shuai Zhang, Liguo Weng, Siew Ann Cheong
2021 Complexity  
Federated learning (FL) is a distributed machine learning (ML) framework.  ...  On the basis of classical FL algorithms, several federated machine learning algorithms are briefly introduced, with emphasis on deep learning and classification and comparisons of those algorithms are  ...  Byzantine Prevention of Federated Learning.  ... 
doi:10.1155/2021/8261663 fatcat:ahr2rpg2indqzg4h3zzda3co5a

Distributed Momentum for Byzantine-resilient Learning [article]

El-Mahdi El-Mhamdi, Rachid Guerraoui, Sébastien Rouault
2020 arXiv   pre-print
We first prove that computing momentum at the workers reduces the variance-norm ratio of the gradient estimation at the server, strengthening Byzantine resilient aggregation rules.  ...  For instance, in the context of federated learning, recent work has shown that Byzantine fault tolerance serves as a good basis to study poisoning (Bagdasaryan et al., 2018; Sun et al., 2019) .  ...  This approach is however based on replication, which is unsuitable for distributed machine learning and stochastic gradient descent, such as federated learning (Konecný et al., 2015) .  ... 
arXiv:2003.00010v2 fatcat:ykr3ay2jinbd3co3zfpd4lfefe

Learning from History for Byzantine Robust Optimization [article]

Sai Praneeth Karimireddy, Lie He, Martin Jaggi
2021 arXiv   pre-print
Byzantine robustness has received significant attention recently given its importance for distributed and federated learning.  ...  First, we show realistic examples where current state of the art robust aggregation rules fail to converge even in the absence of any Byzantine attackers.  ...  IPM uses 11 Byzantine workers while others use 5. The dashed brown line is average aggregator under no attacks (δ = 0). Momentum generally improves all methods, with larger momentum adding stability.  ... 
arXiv:2012.10333v3 fatcat:vgpkiqzb75aetc3ipaniocjvmu
« Previous Showing results 1 — 15 out of 1,620 results