Matches in SemOpenAlex for { <https://semopenalex.org/work/W4285605767> ?p ?o ?g. }
Showing items 1 to 70 of
70
with 100 items per page.
- W4285605767 abstract "Backdoor attacks are rapidly emerging threats to deep neural networks (DNNs). In the backdoor attack scenario, attackers usually implant the backdoor into the target model by manipulating the training dataset or training process. Then, the compromised model behaves normally for benign input yet makes mistakes when the pre-defined trigger appears. In this paper, we analyze the drawbacks of existing attack approaches and propose a novel imperceptible backdoor attack. We treat the trigger pattern as a special kind of noise following a multinomial distribution. A U-net-based network is employed to generate concrete parameters of multinomial distribution for each benign input. This elaborated trigger ensures that our approach is invisible to both humans and statistical detection. Besides the design of the trigger, we also consider the robustness of our approach against model diagnose-based defences. We force the feature representation of malicious input stamped with the trigger to be entangled with the benign one. We demonstrate the effectiveness and robustness against multiple state-of-the-art defences through extensive datasets and networks. Our trigger only modifies less than 1% pixels of a benign image while the modification magnitude is 1. Our source code is available at https://github.com/Ekko-zn/IJCAI2022-Backdoor." @default.
- W4285605767 created "2022-07-16" @default.
- W4285605767 creator A5010605370 @default.
- W4285605767 creator A5015994736 @default.
- W4285605767 creator A5042370642 @default.
- W4285605767 creator A5069600278 @default.
- W4285605767 date "2022-07-01" @default.
- W4285605767 modified "2023-10-16" @default.
- W4285605767 title "Hierarchical Bilevel Learning with Architecture and Loss Search for Hadamard-based Image Restoration" @default.
- W4285605767 doi "https://doi.org/10.24963/ijcai.2022/242" @default.
- W4285605767 hasPublicationYear "2022" @default.
- W4285605767 type Work @default.
- W4285605767 citedByCount "3" @default.
- W4285605767 countsByYear W42856057672023 @default.
- W4285605767 crossrefType "proceedings-article" @default.
- W4285605767 hasAuthorship W4285605767A5010605370 @default.
- W4285605767 hasAuthorship W4285605767A5015994736 @default.
- W4285605767 hasAuthorship W4285605767A5042370642 @default.
- W4285605767 hasAuthorship W4285605767A5069600278 @default.
- W4285605767 hasBestOaLocation W42856057671 @default.
- W4285605767 hasConcept C104317684 @default.
- W4285605767 hasConcept C105795698 @default.
- W4285605767 hasConcept C119857082 @default.
- W4285605767 hasConcept C138885662 @default.
- W4285605767 hasConcept C153180895 @default.
- W4285605767 hasConcept C154945302 @default.
- W4285605767 hasConcept C185592680 @default.
- W4285605767 hasConcept C192065140 @default.
- W4285605767 hasConcept C2776401178 @default.
- W4285605767 hasConcept C2781045450 @default.
- W4285605767 hasConcept C33923547 @default.
- W4285605767 hasConcept C38652104 @default.
- W4285605767 hasConcept C41008148 @default.
- W4285605767 hasConcept C41895202 @default.
- W4285605767 hasConcept C55493867 @default.
- W4285605767 hasConcept C59404180 @default.
- W4285605767 hasConcept C63479239 @default.
- W4285605767 hasConceptScore W4285605767C104317684 @default.
- W4285605767 hasConceptScore W4285605767C105795698 @default.
- W4285605767 hasConceptScore W4285605767C119857082 @default.
- W4285605767 hasConceptScore W4285605767C138885662 @default.
- W4285605767 hasConceptScore W4285605767C153180895 @default.
- W4285605767 hasConceptScore W4285605767C154945302 @default.
- W4285605767 hasConceptScore W4285605767C185592680 @default.
- W4285605767 hasConceptScore W4285605767C192065140 @default.
- W4285605767 hasConceptScore W4285605767C2776401178 @default.
- W4285605767 hasConceptScore W4285605767C2781045450 @default.
- W4285605767 hasConceptScore W4285605767C33923547 @default.
- W4285605767 hasConceptScore W4285605767C38652104 @default.
- W4285605767 hasConceptScore W4285605767C41008148 @default.
- W4285605767 hasConceptScore W4285605767C41895202 @default.
- W4285605767 hasConceptScore W4285605767C55493867 @default.
- W4285605767 hasConceptScore W4285605767C59404180 @default.
- W4285605767 hasConceptScore W4285605767C63479239 @default.
- W4285605767 hasLocation W42856057671 @default.
- W4285605767 hasOpenAccess W4285605767 @default.
- W4285605767 hasPrimaryLocation W42856057671 @default.
- W4285605767 hasRelatedWork W2382607599 @default.
- W4285605767 hasRelatedWork W2546942002 @default.
- W4285605767 hasRelatedWork W2573334707 @default.
- W4285605767 hasRelatedWork W2592385986 @default.
- W4285605767 hasRelatedWork W2792059834 @default.
- W4285605767 hasRelatedWork W2905846897 @default.
- W4285605767 hasRelatedWork W2944661354 @default.
- W4285605767 hasRelatedWork W2970216048 @default.
- W4285605767 hasRelatedWork W2998168123 @default.
- W4285605767 hasRelatedWork W4287995534 @default.
- W4285605767 isParatext "false" @default.
- W4285605767 isRetracted "false" @default.
- W4285605767 workType "article" @default.