Schroth, Christian A. ; Vlaski, Stefan ; Zoubir, Abdelhak M. (2023)
Attacks on Robust Distributed Learning Schemes via Sensitivity Curve Maximization.
24th International Conference on Digital Signal Processing. Rhodes, Greece (11.-13.06.2023)
doi: 10.1109/DSP58604.2023.10167919
Konferenzveröffentlichung, Bibliographie
Kurzbeschreibung (Abstract)
Distributed learning paradigms, such as federated or decentralized learning, allow a collection of agents to solve global learning and optimization problems through limited local interactions. Most such strategies rely on a mixture of local adaptation and aggregation steps, either among peers or at a central fusion center. Classically, aggregation in distributed learning is based on averaging, which is statistically efficient, but susceptible to attacks by even a small number of malicious agents. This observation has motivated a number of recent works, which develop robust aggregation schemes by employing robust variations of the mean. We present a new attack based on sensitivity curve maximization (SCM), and demonstrate that it is able to disrupt existing robust aggregation schemes by injecting small, but effective perturbations.
| Typ des Eintrags: | Konferenzveröffentlichung |
|---|---|
| Erschienen: | 2023 |
| Autor(en): | Schroth, Christian A. ; Vlaski, Stefan ; Zoubir, Abdelhak M. |
| Art des Eintrags: | Bibliographie |
| Titel: | Attacks on Robust Distributed Learning Schemes via Sensitivity Curve Maximization |
| Sprache: | Englisch |
| Publikationsjahr: | 5 Juli 2023 |
| Verlag: | IEEE |
| Buchtitel: | 24th DSP 2023: 2023 24th International Conference on Digital Signal Processing |
| Veranstaltungstitel: | 24th International Conference on Digital Signal Processing |
| Veranstaltungsort: | Rhodes, Greece |
| Veranstaltungsdatum: | 11.-13.06.2023 |
| DOI: | 10.1109/DSP58604.2023.10167919 |
| Zugehörige Links: | |
| Kurzbeschreibung (Abstract): | Distributed learning paradigms, such as federated or decentralized learning, allow a collection of agents to solve global learning and optimization problems through limited local interactions. Most such strategies rely on a mixture of local adaptation and aggregation steps, either among peers or at a central fusion center. Classically, aggregation in distributed learning is based on averaging, which is statistically efficient, but susceptible to attacks by even a small number of malicious agents. This observation has motivated a number of recent works, which develop robust aggregation schemes by employing robust variations of the mean. We present a new attack based on sensitivity curve maximization (SCM), and demonstrate that it is able to disrupt existing robust aggregation schemes by injecting small, but effective perturbations. |
| Freie Schlagworte: | emergenCITY_CPS, emergenCITY |
| Fachbereich(e)/-gebiet(e): | 18 Fachbereich Elektrotechnik und Informationstechnik 18 Fachbereich Elektrotechnik und Informationstechnik > Institut für Nachrichtentechnik 18 Fachbereich Elektrotechnik und Informationstechnik > Institut für Nachrichtentechnik > Signalverarbeitung LOEWE LOEWE > LOEWE-Zentren LOEWE > LOEWE-Zentren > emergenCITY |
| Hinterlegungsdatum: | 10 Jul 2023 10:16 |
| Letzte Änderung: | 10 Jul 2023 10:16 |
| PPN: | |
| Zugehörige Links: | |
| Export: | |
| Suche nach Titel in: | TUfind oder in Google |
 |
Frage zum Eintrag |
Optionen (nur für Redakteure)
 |
Redaktionelle Details anzeigen |
Drucken
Impressum