DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Kim, Changick | - |
dc.contributor.advisor | 김창익 | - |
dc.contributor.author | Kim, Hee-Seon | - |
dc.date.accessioned | 2023-06-26T19:33:52Z | - |
dc.date.available | 2023-06-26T19:33:52Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1032882&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/309876 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 전기및전자공학부, 2023.2,[iii, 19 p. :] | - |
dc.description.abstract | Deep neural networks perform well in various computer vision tasks, but they are vulnerable to adversarial attacks. Since adversarial examples are transferable between models, adversaries can induce a model to predict a wrong class even if the target model's interior is hidden. For these transfer-based attacks, preventing adversarial examples from overfitting the source model is the most important challenge in improving transferability. To improve the adversarial transferability, I introduce the Frequency-Blended Image(FBI) method, which diversifies inputs when generating an adversarial image. Specifically, sensitive-frequency and insensitive-frequency components to the model are extracted from the original image and mixed into the original image with different weights to further increase input diversity. Extensive experiments on the ImageNet dataset show that the attack with the FBI significantly improves the transfer-based attack success rates. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Adversarial attacks▼aTransfer-based attack▼aTransferability▼aInput transformations | - |
dc.subject | 적대적 공격▼a전이 기반 공격▼a전이성▼a입력 변환 | - |
dc.title | Adversarial attack with frequency blended image | - |
dc.title.alternative | 주파수 혼합 이미지를 이용한 적대적 공격 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :전기및전자공학부, | - |
dc.contributor.alternativeauthor | 김희선 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.