Matches in SemOpenAlex for { <https://semopenalex.org/work/W4327500531> ?p ?o ?g. }
Showing items 1 to 64 of
64
with 100 items per page.
- W4327500531 abstract "Federated learning (FL) enables multiple participants to collaboratively train a deep neural network (DNN) model. To combat malicious participants in FL, Byzantine-resilient aggregation rules (AGRs) have been developed. However, although Byzantine-resilient AGRs are effective against untargeted attacks, they become suboptimal when attacks are stealthy and targeted. In this paper, we study the problem of defending against targeted data poisoning attacks in FL and make three main contributions. First, we propose a method for selective extraction of DNN parameters from FL participants’ update vectors that are indicative of attack, and embedding them into low-dimensional latent space. We show that the effectiveness of Byzantine-resilient AGRs such as Trimmed Mean and Krum can be improved if they are used in combination with our proposed method. Second, we develop a clustering-based defense using X-Means for separating items into malicious versus benign clusters in latent space. Such separation allows identification of malicious versus benign updates. Third, using the separation from the previous step, we show that a clean model (i.e., a model that is not negatively impacted by the attack) can be trained using only the benign updates. We experimentally evaluate our defense methods on Fashion-MNIST and CIFAR-10 datasets. Results show that our methods can achieve up to 95% true positive rate and 99% accuracy in malicious update identification across various settings. In addition, the clean models trained using our approach achieve similar accuracy compared to a baseline scenario without poisoning." @default.
- W4327500531 created "2023-03-17" @default.
- W4327500531 creator A5052765649 @default.
- W4327500531 creator A5091167202 @default.
- W4327500531 date "2022-12-01" @default.
- W4327500531 modified "2023-09-23" @default.
- W4327500531 title "Defending Against Targeted Poisoning Attacks in Federated Learning" @default.
- W4327500531 cites W2295124130 @default.
- W4327500531 cites W2559840118 @default.
- W4327500531 cites W3004155269 @default.
- W4327500531 cites W3027749727 @default.
- W4327500531 cites W3040685212 @default.
- W4327500531 cites W3043758338 @default.
- W4327500531 cites W3087391814 @default.
- W4327500531 cites W3091870957 @default.
- W4327500531 cites W3138153888 @default.
- W4327500531 cites W3199119343 @default.
- W4327500531 cites W3203600060 @default.
- W4327500531 cites W4226016526 @default.
- W4327500531 cites W4229455429 @default.
- W4327500531 cites W4312996082 @default.
- W4327500531 doi "https://doi.org/10.1109/tps-isa56441.2022.00033" @default.
- W4327500531 hasPublicationYear "2022" @default.
- W4327500531 type Work @default.
- W4327500531 citedByCount "0" @default.
- W4327500531 crossrefType "proceedings-article" @default.
- W4327500531 hasAuthorship W4327500531A5052765649 @default.
- W4327500531 hasAuthorship W4327500531A5091167202 @default.
- W4327500531 hasConcept C116834253 @default.
- W4327500531 hasConcept C119857082 @default.
- W4327500531 hasConcept C154945302 @default.
- W4327500531 hasConcept C190502265 @default.
- W4327500531 hasConcept C41008148 @default.
- W4327500531 hasConcept C41608201 @default.
- W4327500531 hasConcept C50644808 @default.
- W4327500531 hasConcept C59822182 @default.
- W4327500531 hasConcept C73555534 @default.
- W4327500531 hasConcept C86803240 @default.
- W4327500531 hasConceptScore W4327500531C116834253 @default.
- W4327500531 hasConceptScore W4327500531C119857082 @default.
- W4327500531 hasConceptScore W4327500531C154945302 @default.
- W4327500531 hasConceptScore W4327500531C190502265 @default.
- W4327500531 hasConceptScore W4327500531C41008148 @default.
- W4327500531 hasConceptScore W4327500531C41608201 @default.
- W4327500531 hasConceptScore W4327500531C50644808 @default.
- W4327500531 hasConceptScore W4327500531C59822182 @default.
- W4327500531 hasConceptScore W4327500531C73555534 @default.
- W4327500531 hasConceptScore W4327500531C86803240 @default.
- W4327500531 hasLocation W43275005311 @default.
- W4327500531 hasOpenAccess W4327500531 @default.
- W4327500531 hasPrimaryLocation W43275005311 @default.
- W4327500531 hasRelatedWork W2597787948 @default.
- W4327500531 hasRelatedWork W2791824431 @default.
- W4327500531 hasRelatedWork W2791952757 @default.
- W4327500531 hasRelatedWork W2904174853 @default.
- W4327500531 hasRelatedWork W2904372345 @default.
- W4327500531 hasRelatedWork W2951786554 @default.
- W4327500531 hasRelatedWork W2978290780 @default.
- W4327500531 hasRelatedWork W3186840088 @default.
- W4327500531 hasRelatedWork W4287064118 @default.
- W4327500531 hasRelatedWork W1629725936 @default.
- W4327500531 isParatext "false" @default.
- W4327500531 isRetracted "false" @default.
- W4327500531 workType "article" @default.