Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312578439> ?p ?o ?g. }
Showing items 1 to 80 of
80
with 100 items per page.
- W4312578439 abstract "Federated Learning (FL) allows individual clients to train a global model by aggregating local model updates each round. This results in collaborative model training while main-taining the privacy of clients' sensitive data. However, malicious clients can join the training process and train with poisoned data or send artificial model updates in targeted poisoning attacks. Many defenses to targeted poisoning attacks rely on anomaly-detection based metrics which remove participants that deviate from the majority. Similarly, aggregation-based defenses aim to reduce the impact of outliers, while L2-norm clipping tries to scale down the impact of malicious models. However, oftentimes these defenses misidentify benign clients as malicious or only work under specific attack conditions. In our paper, we examine the effectiveness of two anomaly -detection metrics on three different aggregation methods, in addition to the presence of L2-norm clipping and weight selection, across two different types of attacks. We also combine different defenses in order to examine their interaction and examine each defense when no attack is present. We found minimum aggregation to be the most effective defense against label-flipping attacks, whereas both minimum aggregation and geometric median worked well against distributed backdoor attacks. Using random weight selection significantly deteriorated defenses against both attacks, whereas the use of clipping made little difference. Finally, the main task accuracy was directly correlated with the BA in the label-flipping attack and generally was close to the MA in benign scenarios. However, in the DBA the MA and BA are inversely correlated and the MA fluctuates greatly." @default.
- W4312578439 created "2023-01-05" @default.
- W4312578439 creator A5014182467 @default.
- W4312578439 creator A5079273195 @default.
- W4312578439 creator A5081844561 @default.
- W4312578439 date "2022-10-01" @default.
- W4312578439 modified "2023-09-30" @default.
- W4312578439 title "Evaluation of Various Defense Techniques Against Targeted Poisoning Attacks in Federated Learning" @default.
- W4312578439 cites W2125908420 @default.
- W4312578439 cites W2559840118 @default.
- W4312578439 cites W2767079719 @default.
- W4312578439 cites W2883059862 @default.
- W4312578439 cites W2914583895 @default.
- W4312578439 cites W2936608441 @default.
- W4312578439 cites W2963183964 @default.
- W4312578439 cites W2963318081 @default.
- W4312578439 cites W2982426954 @default.
- W4312578439 cites W4221129260 @default.
- W4312578439 cites W4308536713 @default.
- W4312578439 doi "https://doi.org/10.1109/mass56207.2022.00102" @default.
- W4312578439 hasPublicationYear "2022" @default.
- W4312578439 type Work @default.
- W4312578439 citedByCount "0" @default.
- W4312578439 crossrefType "proceedings-article" @default.
- W4312578439 hasAuthorship W4312578439A5014182467 @default.
- W4312578439 hasAuthorship W4312578439A5079273195 @default.
- W4312578439 hasAuthorship W4312578439A5081844561 @default.
- W4312578439 hasConcept C111919701 @default.
- W4312578439 hasConcept C119857082 @default.
- W4312578439 hasConcept C127413603 @default.
- W4312578439 hasConcept C138885662 @default.
- W4312578439 hasConcept C154945302 @default.
- W4312578439 hasConcept C17744445 @default.
- W4312578439 hasConcept C191795146 @default.
- W4312578439 hasConcept C199539241 @default.
- W4312578439 hasConcept C201995342 @default.
- W4312578439 hasConcept C2776848632 @default.
- W4312578439 hasConcept C2780451532 @default.
- W4312578439 hasConcept C2781045450 @default.
- W4312578439 hasConcept C38652104 @default.
- W4312578439 hasConcept C41008148 @default.
- W4312578439 hasConcept C41895202 @default.
- W4312578439 hasConcept C739882 @default.
- W4312578439 hasConcept C79337645 @default.
- W4312578439 hasConcept C98045186 @default.
- W4312578439 hasConceptScore W4312578439C111919701 @default.
- W4312578439 hasConceptScore W4312578439C119857082 @default.
- W4312578439 hasConceptScore W4312578439C127413603 @default.
- W4312578439 hasConceptScore W4312578439C138885662 @default.
- W4312578439 hasConceptScore W4312578439C154945302 @default.
- W4312578439 hasConceptScore W4312578439C17744445 @default.
- W4312578439 hasConceptScore W4312578439C191795146 @default.
- W4312578439 hasConceptScore W4312578439C199539241 @default.
- W4312578439 hasConceptScore W4312578439C201995342 @default.
- W4312578439 hasConceptScore W4312578439C2776848632 @default.
- W4312578439 hasConceptScore W4312578439C2780451532 @default.
- W4312578439 hasConceptScore W4312578439C2781045450 @default.
- W4312578439 hasConceptScore W4312578439C38652104 @default.
- W4312578439 hasConceptScore W4312578439C41008148 @default.
- W4312578439 hasConceptScore W4312578439C41895202 @default.
- W4312578439 hasConceptScore W4312578439C739882 @default.
- W4312578439 hasConceptScore W4312578439C79337645 @default.
- W4312578439 hasConceptScore W4312578439C98045186 @default.
- W4312578439 hasFunder F4320306076 @default.
- W4312578439 hasLocation W43125784391 @default.
- W4312578439 hasOpenAccess W4312578439 @default.
- W4312578439 hasPrimaryLocation W43125784391 @default.
- W4312578439 hasRelatedWork W2460485004 @default.
- W4312578439 hasRelatedWork W2769077153 @default.
- W4312578439 hasRelatedWork W2886026457 @default.
- W4312578439 hasRelatedWork W2901091351 @default.
- W4312578439 hasRelatedWork W2982977353 @default.
- W4312578439 hasRelatedWork W2996061416 @default.
- W4312578439 hasRelatedWork W3044458868 @default.
- W4312578439 hasRelatedWork W4226370762 @default.
- W4312578439 hasRelatedWork W4226413265 @default.
- W4312578439 hasRelatedWork W4312578439 @default.
- W4312578439 isParatext "false" @default.
- W4312578439 isRetracted "false" @default.
- W4312578439 workType "article" @default.