Matches in SemOpenAlex for { <https://semopenalex.org/work/W4366308436> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W4366308436 abstract "Deep neural networks (DNNs) have lately shown tremendous performance in various applications. However, along-side their superiority in these tasks, recent studies have demon-strated that DNNs are easily fooled by adversarial attacks. To guard against adversarial examples, we provide a new solution to hardening DNNs through Block-term Dropout (BT-Dropout), an adversarial defense technique that leverages a latent high-order factorization of the network. Specifically, we impose low-rank block-term tensor structure on the weights of fully-connected layer to obtain compact networks, and then apply BT-Dropout in the latent subspace without pruning the weights directly. Meanwhile, for activation tensor fed into fully-connected layer, Tucker Dropout which can be viewed as a special case of BT-Dropout is introduced to preserve multilinear structure of activations. Furthermore, we show that BT-Dropout implicitly regularizes the tensor decomposition. Comprehensive experiments have demonstrated the effectiveness of our proposed method to improve the adversarial robustness for the models on standard image classification benchmarks." @default.
- W4366308436 created "2023-04-20" @default.
- W4366308436 creator A5019681146 @default.
- W4366308436 creator A5038006108 @default.
- W4366308436 creator A5090602006 @default.
- W4366308436 date "2022-10-01" @default.
- W4366308436 modified "2023-09-26" @default.
- W4366308436 title "Block-term Dropout For Robust Adversarial Defense" @default.
- W4366308436 cites W1963826206 @default.
- W4366308436 cites W1980287119 @default.
- W4366308436 cites W2000045479 @default.
- W4366308436 cites W2000215628 @default.
- W4366308436 cites W2076063813 @default.
- W4366308436 cites W2194775991 @default.
- W4366308436 cites W2963857521 @default.
- W4366308436 cites W2963995504 @default.
- W4366308436 cites W3096366160 @default.
- W4366308436 cites W3100646226 @default.
- W4366308436 cites W3137884659 @default.
- W4366308436 cites W3157471480 @default.
- W4366308436 doi "https://doi.org/10.1109/ictai56018.2022.00097" @default.
- W4366308436 hasPublicationYear "2022" @default.
- W4366308436 type Work @default.
- W4366308436 citedByCount "0" @default.
- W4366308436 crossrefType "proceedings-article" @default.
- W4366308436 hasAuthorship W4366308436A5019681146 @default.
- W4366308436 hasAuthorship W4366308436A5038006108 @default.
- W4366308436 hasAuthorship W4366308436A5090602006 @default.
- W4366308436 hasConcept C104317684 @default.
- W4366308436 hasConcept C11413529 @default.
- W4366308436 hasConcept C119857082 @default.
- W4366308436 hasConcept C154945302 @default.
- W4366308436 hasConcept C185592680 @default.
- W4366308436 hasConcept C202444582 @default.
- W4366308436 hasConcept C2524010 @default.
- W4366308436 hasConcept C2776145597 @default.
- W4366308436 hasConcept C2777210771 @default.
- W4366308436 hasConcept C32834561 @default.
- W4366308436 hasConcept C33923547 @default.
- W4366308436 hasConcept C37736160 @default.
- W4366308436 hasConcept C41008148 @default.
- W4366308436 hasConcept C55493867 @default.
- W4366308436 hasConcept C63479239 @default.
- W4366308436 hasConcept C84392682 @default.
- W4366308436 hasConceptScore W4366308436C104317684 @default.
- W4366308436 hasConceptScore W4366308436C11413529 @default.
- W4366308436 hasConceptScore W4366308436C119857082 @default.
- W4366308436 hasConceptScore W4366308436C154945302 @default.
- W4366308436 hasConceptScore W4366308436C185592680 @default.
- W4366308436 hasConceptScore W4366308436C202444582 @default.
- W4366308436 hasConceptScore W4366308436C2524010 @default.
- W4366308436 hasConceptScore W4366308436C2776145597 @default.
- W4366308436 hasConceptScore W4366308436C2777210771 @default.
- W4366308436 hasConceptScore W4366308436C32834561 @default.
- W4366308436 hasConceptScore W4366308436C33923547 @default.
- W4366308436 hasConceptScore W4366308436C37736160 @default.
- W4366308436 hasConceptScore W4366308436C41008148 @default.
- W4366308436 hasConceptScore W4366308436C55493867 @default.
- W4366308436 hasConceptScore W4366308436C63479239 @default.
- W4366308436 hasConceptScore W4366308436C84392682 @default.
- W4366308436 hasFunder F4320321001 @default.
- W4366308436 hasLocation W43663084361 @default.
- W4366308436 hasOpenAccess W4366308436 @default.
- W4366308436 hasPrimaryLocation W43663084361 @default.
- W4366308436 hasRelatedWork W2937211408 @default.
- W4366308436 hasRelatedWork W2952919291 @default.
- W4366308436 hasRelatedWork W3035729345 @default.
- W4366308436 hasRelatedWork W3114774860 @default.
- W4366308436 hasRelatedWork W3193857078 @default.
- W4366308436 hasRelatedWork W4287549955 @default.
- W4366308436 hasRelatedWork W4303857474 @default.
- W4366308436 hasRelatedWork W4306177400 @default.
- W4366308436 hasRelatedWork W4311734044 @default.
- W4366308436 hasRelatedWork W4312741812 @default.
- W4366308436 isParatext "false" @default.
- W4366308436 isRetracted "false" @default.
- W4366308436 workType "article" @default.