Matches in SemOpenAlex for { <https://semopenalex.org/work/W3034855057> ?p ?o ?g. }
- W3034855057 abstract "While deep learning continues to permeate through all fields of signal processing and machine learning, a critical exploit in these frameworks exists and remains unsolved. These exploits, or adversarial examples, are a type of signal attack that can change the output class of a classifier by perturbing the stimulus signal by an imperceptible amount. The attack takes advantage of statistical irregularities within the training data, where the added perturbations can move the image across deep learning decision boundaries. What is even more alarming is the transferability of these attacks to different deep learning models and architectures. This means a successful attack on one model has adversarial effects on other, unrelated models. In a general sense, adversarial attack through perturbations is not a machine learning vulnerability. Human and biological vision can also be fooled by various methods, i.e. mixing high and low frequency images together, by altering semantically related signals, or by sufficiently distorting the input signal. However, the amount and magnitude of such a distortion required to alter biological perception is at a much larger scale. In this work, we explored this gap through the lens of biology and neuroscience in order to understand the robustness exhibited in human perception. Our experiments show that by leveraging sparsity and modeling the biological mechanisms at a cellular level, we are able to mitigate the effect of adversarial alterations to the signal that have no perceptible meaning. Furthermore, we present and illustrate the effects of top-down functional processes that contribute to the inherent immunity in human perception in the context of exploiting these properties to make a more robust machine vision system." @default.
- W3034855057 created "2020-06-19" @default.
- W3034855057 creator A5062980852 @default.
- W3034855057 creator A5067556276 @default.
- W3034855057 creator A5082337387 @default.
- W3034855057 creator A5086534835 @default.
- W3034855057 date "2020-06-01" @default.
- W3034855057 modified "2023-10-02" @default.
- W3034855057 title "Modeling Biological Immunity to Adversarial Examples" @default.
- W3034855057 cites W1976695311 @default.
- W3034855057 cites W1996737327 @default.
- W3034855057 cites W2007709912 @default.
- W3034855057 cites W2010419711 @default.
- W3034855057 cites W2013475237 @default.
- W3034855057 cites W2026738815 @default.
- W3034855057 cites W2036123457 @default.
- W3034855057 cites W2042925217 @default.
- W3034855057 cites W2066340926 @default.
- W3034855057 cites W2070339144 @default.
- W3034855057 cites W2070605650 @default.
- W3034855057 cites W2074210274 @default.
- W3034855057 cites W2078509321 @default.
- W3034855057 cites W2094048200 @default.
- W3034855057 cites W2096926732 @default.
- W3034855057 cites W2103113084 @default.
- W3034855057 cites W2105464873 @default.
- W3034855057 cites W2112246162 @default.
- W3034855057 cites W2116360511 @default.
- W3034855057 cites W2151035455 @default.
- W3034855057 cites W2164137768 @default.
- W3034855057 cites W2194775991 @default.
- W3034855057 cites W2295237794 @default.
- W3034855057 cites W2344384217 @default.
- W3034855057 cites W2495710720 @default.
- W3034855057 cites W2576960802 @default.
- W3034855057 cites W2586433236 @default.
- W3034855057 cites W2883238243 @default.
- W3034855057 cites W2964246953 @default.
- W3034855057 cites W2969340765 @default.
- W3034855057 cites W2973159797 @default.
- W3034855057 cites W3106224614 @default.
- W3034855057 cites W4249474355 @default.
- W3034855057 doi "https://doi.org/10.1109/cvpr42600.2020.00472" @default.
- W3034855057 hasPublicationYear "2020" @default.
- W3034855057 type Work @default.
- W3034855057 sameAs 3034855057 @default.
- W3034855057 citedByCount "16" @default.
- W3034855057 countsByYear W30348550572020 @default.
- W3034855057 countsByYear W30348550572021 @default.
- W3034855057 countsByYear W30348550572023 @default.
- W3034855057 crossrefType "proceedings-article" @default.
- W3034855057 hasAuthorship W3034855057A5062980852 @default.
- W3034855057 hasAuthorship W3034855057A5067556276 @default.
- W3034855057 hasAuthorship W3034855057A5082337387 @default.
- W3034855057 hasAuthorship W3034855057A5086534835 @default.
- W3034855057 hasConcept C104317684 @default.
- W3034855057 hasConcept C108583219 @default.
- W3034855057 hasConcept C119857082 @default.
- W3034855057 hasConcept C154945302 @default.
- W3034855057 hasConcept C165696696 @default.
- W3034855057 hasConcept C169760540 @default.
- W3034855057 hasConcept C185592680 @default.
- W3034855057 hasConcept C26760741 @default.
- W3034855057 hasConcept C37736160 @default.
- W3034855057 hasConcept C38652104 @default.
- W3034855057 hasConcept C41008148 @default.
- W3034855057 hasConcept C55493867 @default.
- W3034855057 hasConcept C63479239 @default.
- W3034855057 hasConcept C86803240 @default.
- W3034855057 hasConcept C95623464 @default.
- W3034855057 hasConceptScore W3034855057C104317684 @default.
- W3034855057 hasConceptScore W3034855057C108583219 @default.
- W3034855057 hasConceptScore W3034855057C119857082 @default.
- W3034855057 hasConceptScore W3034855057C154945302 @default.
- W3034855057 hasConceptScore W3034855057C165696696 @default.
- W3034855057 hasConceptScore W3034855057C169760540 @default.
- W3034855057 hasConceptScore W3034855057C185592680 @default.
- W3034855057 hasConceptScore W3034855057C26760741 @default.
- W3034855057 hasConceptScore W3034855057C37736160 @default.
- W3034855057 hasConceptScore W3034855057C38652104 @default.
- W3034855057 hasConceptScore W3034855057C41008148 @default.
- W3034855057 hasConceptScore W3034855057C55493867 @default.
- W3034855057 hasConceptScore W3034855057C63479239 @default.
- W3034855057 hasConceptScore W3034855057C86803240 @default.
- W3034855057 hasConceptScore W3034855057C95623464 @default.
- W3034855057 hasLocation W30348550571 @default.
- W3034855057 hasLocation W30348550572 @default.
- W3034855057 hasOpenAccess W3034855057 @default.
- W3034855057 hasPrimaryLocation W30348550571 @default.
- W3034855057 hasRelatedWork W3193857078 @default.
- W3034855057 hasRelatedWork W3208304128 @default.
- W3034855057 hasRelatedWork W4221150850 @default.
- W3034855057 hasRelatedWork W4221152960 @default.
- W3034855057 hasRelatedWork W4223943233 @default.
- W3034855057 hasRelatedWork W4312200629 @default.
- W3034855057 hasRelatedWork W4360585206 @default.
- W3034855057 hasRelatedWork W4364306694 @default.
- W3034855057 hasRelatedWork W4380075502 @default.
- W3034855057 hasRelatedWork W4382463655 @default.
- W3034855057 isParatext "false" @default.