GradMDM : Adversarial Attack on Dynamic Networks
Dynamic neural networks can greatly reduce computation redundancy without compromising accuracy by adapting their structures based on the input. In this paper, we explore the robustness of dynamic neural networks against energy-oriented attacks targeted at reducing their efficiency. Specifically, we...
Ausführliche Beschreibung
Bibliographische Detailangaben
| Veröffentlicht in: | IEEE transactions on pattern analysis and machine intelligence. - 1979. - 45(2023), 9 vom: 29. Sept., Seite 11374-11381
|
|---|
| 1. Verfasser: |
Pan, Jianhong
(VerfasserIn) |
|---|
| Weitere Verfasser: |
Foo, Lin Geng,
Zheng, Qichen,
Fan, Zhipeng,
Rahmani, Hossein,
Ke, Qiuhong,
Liu, Jun |
|---|
| Format: | Online-Aufsatz
|
|---|
| Sprache: | English |
|---|
| Veröffentlicht: |
2023
|
|---|
| Zugriff auf das übergeordnete Werk: | IEEE transactions on pattern analysis and machine intelligence
|
|---|
| Schlagworte: | Journal Article |
|---|