Matches in SemOpenAlex for { <https://semopenalex.org/work/W3209507148> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W3209507148 abstract "Byzantine attacks hinder the deployment of federated learning algorithms. Although we know that the benign gradients and Byzantine attacked gradients are distributed differently, to detect the malicious gradients is challenging due to (1) the gradient is high-dimensional and each dimension has its unique distribution and (2) the benign gradients and the attacked gradients are always mixed (two-sample test methods cannot apply directly). To address the above, for the first time, we propose MANDERA which is theoretically guaranteed to efficiently detect all malicious gradients under Byzantine attacks with no prior knowledge or history about the number of attacked nodes. More specifically, we transfer the original updating gradient space into a ranking matrix. By such an operation, the scales of different dimensions of the gradients in the ranking space become identical. The high-dimensional benign gradients and the malicious gradients can be easily separated. The effectiveness of MANDERA is further confirmed by experimentation on four Byzantine attack implementations (Gaussian, Zero Gradient, Sign Flipping, Shifted Mean), comparing with state-of-the-art defenses. The experiments cover both IID and Non-IID datasets." @default.
- W3209507148 created "2021-11-08" @default.
- W3209507148 creator A5016670591 @default.
- W3209507148 creator A5026035989 @default.
- W3209507148 creator A5032717607 @default.
- W3209507148 creator A5039679896 @default.
- W3209507148 date "2021-10-22" @default.
- W3209507148 modified "2023-10-17" @default.
- W3209507148 title "MANDERA: Malicious Node Detection in Federated Learning via Ranking" @default.
- W3209507148 cites W2614254310 @default.
- W3209507148 cites W2750384547 @default.
- W3209507148 cites W2752689052 @default.
- W3209507148 cites W2946633206 @default.
- W3209507148 cites W2963165390 @default.
- W3209507148 cites W2964279249 @default.
- W3209507148 cites W3003426262 @default.
- W3209507148 cites W3035498193 @default.
- W3209507148 cites W3036791758 @default.
- W3209507148 cites W3037024761 @default.
- W3209507148 cites W3046449784 @default.
- W3209507148 cites W3048715803 @default.
- W3209507148 cites W3087391814 @default.
- W3209507148 cites W3107632872 @default.
- W3209507148 cites W3118608800 @default.
- W3209507148 cites W3173401850 @default.
- W3209507148 doi "https://doi.org/10.48550/arxiv.2110.11736" @default.
- W3209507148 hasPublicationYear "2021" @default.
- W3209507148 type Work @default.
- W3209507148 sameAs 3209507148 @default.
- W3209507148 citedByCount "0" @default.
- W3209507148 crossrefType "posted-content" @default.
- W3209507148 hasAuthorship W3209507148A5016670591 @default.
- W3209507148 hasAuthorship W3209507148A5026035989 @default.
- W3209507148 hasAuthorship W3209507148A5032717607 @default.
- W3209507148 hasAuthorship W3209507148A5039679896 @default.
- W3209507148 hasBestOaLocation W32095071481 @default.
- W3209507148 hasConcept C114614502 @default.
- W3209507148 hasConcept C119857082 @default.
- W3209507148 hasConcept C121332964 @default.
- W3209507148 hasConcept C154945302 @default.
- W3209507148 hasConcept C163716315 @default.
- W3209507148 hasConcept C189430467 @default.
- W3209507148 hasConcept C33676613 @default.
- W3209507148 hasConcept C33923547 @default.
- W3209507148 hasConcept C41008148 @default.
- W3209507148 hasConcept C62520636 @default.
- W3209507148 hasConcept C62611344 @default.
- W3209507148 hasConceptScore W3209507148C114614502 @default.
- W3209507148 hasConceptScore W3209507148C119857082 @default.
- W3209507148 hasConceptScore W3209507148C121332964 @default.
- W3209507148 hasConceptScore W3209507148C154945302 @default.
- W3209507148 hasConceptScore W3209507148C163716315 @default.
- W3209507148 hasConceptScore W3209507148C189430467 @default.
- W3209507148 hasConceptScore W3209507148C33676613 @default.
- W3209507148 hasConceptScore W3209507148C33923547 @default.
- W3209507148 hasConceptScore W3209507148C41008148 @default.
- W3209507148 hasConceptScore W3209507148C62520636 @default.
- W3209507148 hasConceptScore W3209507148C62611344 @default.
- W3209507148 hasLocation W32095071481 @default.
- W3209507148 hasOpenAccess W3209507148 @default.
- W3209507148 hasPrimaryLocation W32095071481 @default.
- W3209507148 hasRelatedWork W2103764822 @default.
- W3209507148 hasRelatedWork W2961085424 @default.
- W3209507148 hasRelatedWork W3046775127 @default.
- W3209507148 hasRelatedWork W3107474891 @default.
- W3209507148 hasRelatedWork W3209574120 @default.
- W3209507148 hasRelatedWork W4205958290 @default.
- W3209507148 hasRelatedWork W4286629047 @default.
- W3209507148 hasRelatedWork W4306321456 @default.
- W3209507148 hasRelatedWork W4306674287 @default.
- W3209507148 hasRelatedWork W4224009465 @default.
- W3209507148 isParatext "false" @default.
- W3209507148 isRetracted "false" @default.
- W3209507148 magId "3209507148" @default.
- W3209507148 workType "article" @default.