Matches in SemOpenAlex for { <https://semopenalex.org/work/W2897083899> ?p ?o ?g. }
- W2897083899 abstract "Deep neural networks (DNNs) are vulnerable to maliciously generated adversarial examples. These examples are intentionally designed by making imperceptible perturbations and often mislead a DNN into making an incorrect prediction. This phenomenon means that there is significant risk in applying DNNs to safety-critical applications, such as driverless cars. To address this issue, we present a visual analytics approach to explain the primary cause of the wrong predictions introduced by adversarial examples. The key is to analyze the datapaths of the adversarial examples and compare them with those of the normal examples. A datapath is a group of critical neurons and their connections. To this end, we formulate the datapath extraction as a subset selection problem and approximately solve it based on back-propagation. A multi-level visualization consisting of a segmented DAG (layer level), an Euler diagram (feature map level), and a heat map (neuron level), has been designed to help experts investigate datapaths from the high-level layers to the detailed neuron activations. Two case studies are conducted that demonstrate the promise of our approach in support of explaining the working mechanism of adversarial examples." @default.
- W2897083899 created "2018-10-26" @default.
- W2897083899 creator A5003419563 @default.
- W2897083899 creator A5003713785 @default.
- W2897083899 creator A5015663174 @default.
- W2897083899 creator A5038775008 @default.
- W2897083899 creator A5077341571 @default.
- W2897083899 date "2018-10-01" @default.
- W2897083899 modified "2023-10-10" @default.
- W2897083899 title "Analyzing the Noise Robustness of Deep Neural Networks" @default.
- W2897083899 cites W1928219753 @default.
- W2897083899 cites W2009858410 @default.
- W2897083899 cites W2027855569 @default.
- W2897083899 cites W2030522750 @default.
- W2897083899 cites W2046336997 @default.
- W2897083899 cites W2071353749 @default.
- W2897083899 cites W2074930186 @default.
- W2897083899 cites W2116420093 @default.
- W2897083899 cites W2117006370 @default.
- W2897083899 cites W2117539524 @default.
- W2897083899 cites W2124809398 @default.
- W2897083899 cites W2159205441 @default.
- W2897083899 cites W2159729957 @default.
- W2897083899 cites W2161114148 @default.
- W2897083899 cites W2163922914 @default.
- W2897083899 cites W2183341477 @default.
- W2897083899 cites W2194775991 @default.
- W2897083899 cites W2243397390 @default.
- W2897083899 cites W2342045095 @default.
- W2897083899 cites W2343061342 @default.
- W2897083899 cites W2508572344 @default.
- W2897083899 cites W2512304460 @default.
- W2897083899 cites W2529597677 @default.
- W2897083899 cites W2542889218 @default.
- W2897083899 cites W2543927648 @default.
- W2897083899 cites W2607223307 @default.
- W2897083899 cites W2751298778 @default.
- W2897083899 cites W2751305043 @default.
- W2897083899 cites W2751746637 @default.
- W2897083899 cites W2752194699 @default.
- W2897083899 cites W2752332392 @default.
- W2897083899 cites W2753539354 @default.
- W2897083899 cites W2753713840 @default.
- W2897083899 cites W2792641098 @default.
- W2897083899 cites W2888685797 @default.
- W2897083899 cites W2919115771 @default.
- W2897083899 cites W2962700793 @default.
- W2897083899 cites W2963446712 @default.
- W2897083899 cites W2964200170 @default.
- W2897083899 cites W4243718952 @default.
- W2897083899 doi "https://doi.org/10.1109/vast.2018.8802509" @default.
- W2897083899 hasPublicationYear "2018" @default.
- W2897083899 type Work @default.
- W2897083899 sameAs 2897083899 @default.
- W2897083899 citedByCount "34" @default.
- W2897083899 countsByYear W28970838992018 @default.
- W2897083899 countsByYear W28970838992019 @default.
- W2897083899 countsByYear W28970838992020 @default.
- W2897083899 countsByYear W28970838992021 @default.
- W2897083899 countsByYear W28970838992022 @default.
- W2897083899 countsByYear W28970838992023 @default.
- W2897083899 crossrefType "proceedings-article" @default.
- W2897083899 hasAuthorship W2897083899A5003419563 @default.
- W2897083899 hasAuthorship W2897083899A5003713785 @default.
- W2897083899 hasAuthorship W2897083899A5015663174 @default.
- W2897083899 hasAuthorship W2897083899A5038775008 @default.
- W2897083899 hasAuthorship W2897083899A5077341571 @default.
- W2897083899 hasBestOaLocation W28970838992 @default.
- W2897083899 hasConcept C104317684 @default.
- W2897083899 hasConcept C119857082 @default.
- W2897083899 hasConcept C149635348 @default.
- W2897083899 hasConcept C154945302 @default.
- W2897083899 hasConcept C185592680 @default.
- W2897083899 hasConcept C2781198647 @default.
- W2897083899 hasConcept C2984842247 @default.
- W2897083899 hasConcept C36464697 @default.
- W2897083899 hasConcept C37736160 @default.
- W2897083899 hasConcept C41008148 @default.
- W2897083899 hasConcept C50644808 @default.
- W2897083899 hasConcept C52622490 @default.
- W2897083899 hasConcept C55493867 @default.
- W2897083899 hasConcept C63479239 @default.
- W2897083899 hasConceptScore W2897083899C104317684 @default.
- W2897083899 hasConceptScore W2897083899C119857082 @default.
- W2897083899 hasConceptScore W2897083899C149635348 @default.
- W2897083899 hasConceptScore W2897083899C154945302 @default.
- W2897083899 hasConceptScore W2897083899C185592680 @default.
- W2897083899 hasConceptScore W2897083899C2781198647 @default.
- W2897083899 hasConceptScore W2897083899C2984842247 @default.
- W2897083899 hasConceptScore W2897083899C36464697 @default.
- W2897083899 hasConceptScore W2897083899C37736160 @default.
- W2897083899 hasConceptScore W2897083899C41008148 @default.
- W2897083899 hasConceptScore W2897083899C50644808 @default.
- W2897083899 hasConceptScore W2897083899C52622490 @default.
- W2897083899 hasConceptScore W2897083899C55493867 @default.
- W2897083899 hasConceptScore W2897083899C63479239 @default.
- W2897083899 hasLocation W28970838991 @default.
- W2897083899 hasLocation W28970838992 @default.
- W2897083899 hasLocation W28970838993 @default.
- W2897083899 hasLocation W28970838994 @default.