FedMID: A Data-Free Method for Using Intermediate Outputs as a Defense Mechanism Against Poisoning Attacks in Federated Learning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914759640612864 |
|---|---|
| author | Han, Sungwon Song, Hyeonho Park, Sungwon Cha, Meeyoung |
| author_facet | Han, Sungwon Song, Hyeonho Park, Sungwon Cha, Meeyoung |
| contents | Federated learning combines local updates from clients to produce a global model, which is susceptible to poisoning attacks. Most previous defense strategies relied on vectors derived from projections of local updates on a Euclidean space; however, these methods fail to accurately represent the functionality and structure of local models, resulting in inconsistent performance. Here, we present a new paradigm to defend against poisoning attacks in federated learning using functional mappings of local models based on intermediate outputs. Experiments show that our mechanism is robust under a broad range of computing conditions and advanced attack scenarios, enabling safer collaboration among data-sensitive participants via federated learning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_11905 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | FedMID: A Data-Free Method for Using Intermediate Outputs as a Defense Mechanism Against Poisoning Attacks in Federated Learning Han, Sungwon Song, Hyeonho Park, Sungwon Cha, Meeyoung Machine Learning Cryptography and Security Federated learning combines local updates from clients to produce a global model, which is susceptible to poisoning attacks. Most previous defense strategies relied on vectors derived from projections of local updates on a Euclidean space; however, these methods fail to accurately represent the functionality and structure of local models, resulting in inconsistent performance. Here, we present a new paradigm to defend against poisoning attacks in federated learning using functional mappings of local models based on intermediate outputs. Experiments show that our mechanism is robust under a broad range of computing conditions and advanced attack scenarios, enabling safer collaboration among data-sensitive participants via federated learning. |
| title | FedMID: A Data-Free Method for Using Intermediate Outputs as a Defense Mechanism Against Poisoning Attacks in Federated Learning |
| topic | Machine Learning Cryptography and Security |
| url | https://arxiv.org/abs/2404.11905 |