Matches in SemOpenAlex for { <https://semopenalex.org/work/W2828141535> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W2828141535 abstract "Machine learning algorithms can be fooled by small well-designed adversarial perturbations. This is reminiscent of cellular decision-making where ligands (called antagonists) prevent correct signalling, like in early immune recognition. We draw a formal analogy between neural networks used in machine learning and models of cellular decision-making (adaptive proofreading). We apply attacks from machine learning to simple decision-making models, and show explicitly the correspondence to antagonism by weakly bound ligands. Such antagonism is absent in more nonlinear models, which inspired us to implement a biomimetic defence in neural networks filtering out adversarial perturbations. We then apply a gradient-descent approach from machine learning to different cellular decision-making models, and we reveal the existence of two regimes characterized by the presence or absence of a critical point for the gradient. This critical point causes the strongest antagonists to lie close to the decision boundary. This is validated in the loss landscapes of robust neural networks and cellular decision-making models, and observed experimentally for immune cells. For both regimes, we explain how associated defence mechanisms shape the geometry of the loss landscape, and why different adversarial attacks are effective in different regimes. Our work connects evolved cellular decision-making to machine learning, and motivates the design of a general theory of adversarial perturbations, both for in vivo and in silico systems." @default.
- W2828141535 created "2018-07-19" @default.
- W2828141535 creator A5019976122 @default.
- W2828141535 creator A5027310054 @default.
- W2828141535 creator A5056287697 @default.
- W2828141535 date "2018-07-10" @default.
- W2828141535 modified "2023-09-26" @default.
- W2828141535 title "Fooling the classifier: Ligand antagonism and adversarial examples" @default.
- W2828141535 hasPublicationYear "2018" @default.
- W2828141535 type Work @default.
- W2828141535 sameAs 2828141535 @default.
- W2828141535 citedByCount "0" @default.
- W2828141535 crossrefType "posted-content" @default.
- W2828141535 hasAuthorship W2828141535A5019976122 @default.
- W2828141535 hasAuthorship W2828141535A5027310054 @default.
- W2828141535 hasAuthorship W2828141535A5056287697 @default.
- W2828141535 hasConcept C119857082 @default.
- W2828141535 hasConcept C138885662 @default.
- W2828141535 hasConcept C154945302 @default.
- W2828141535 hasConcept C37736160 @default.
- W2828141535 hasConcept C41008148 @default.
- W2828141535 hasConcept C41895202 @default.
- W2828141535 hasConcept C42023084 @default.
- W2828141535 hasConcept C50644808 @default.
- W2828141535 hasConcept C521332185 @default.
- W2828141535 hasConcept C95623464 @default.
- W2828141535 hasConceptScore W2828141535C119857082 @default.
- W2828141535 hasConceptScore W2828141535C138885662 @default.
- W2828141535 hasConceptScore W2828141535C154945302 @default.
- W2828141535 hasConceptScore W2828141535C37736160 @default.
- W2828141535 hasConceptScore W2828141535C41008148 @default.
- W2828141535 hasConceptScore W2828141535C41895202 @default.
- W2828141535 hasConceptScore W2828141535C42023084 @default.
- W2828141535 hasConceptScore W2828141535C50644808 @default.
- W2828141535 hasConceptScore W2828141535C521332185 @default.
- W2828141535 hasConceptScore W2828141535C95623464 @default.
- W2828141535 hasLocation W28281415351 @default.
- W2828141535 hasOpenAccess W2828141535 @default.
- W2828141535 hasPrimaryLocation W28281415351 @default.
- W2828141535 hasRelatedWork W160531206 @default.
- W2828141535 hasRelatedWork W2001431571 @default.
- W2828141535 hasRelatedWork W2104706909 @default.
- W2828141535 hasRelatedWork W2143104085 @default.
- W2828141535 hasRelatedWork W2168399023 @default.
- W2828141535 hasRelatedWork W2177180613 @default.
- W2828141535 hasRelatedWork W2247294682 @default.
- W2828141535 hasRelatedWork W2263057803 @default.
- W2828141535 hasRelatedWork W2470418787 @default.
- W2828141535 hasRelatedWork W2739241668 @default.
- W2828141535 hasRelatedWork W2748504025 @default.
- W2828141535 hasRelatedWork W2753758104 @default.
- W2828141535 hasRelatedWork W2914877261 @default.
- W2828141535 hasRelatedWork W2915160854 @default.
- W2828141535 hasRelatedWork W2943653773 @default.
- W2828141535 hasRelatedWork W2946617578 @default.
- W2828141535 hasRelatedWork W2949228990 @default.
- W2828141535 hasRelatedWork W3037581782 @default.
- W2828141535 hasRelatedWork W3113119677 @default.
- W2828141535 hasRelatedWork W3201628275 @default.
- W2828141535 isParatext "false" @default.
- W2828141535 isRetracted "false" @default.
- W2828141535 magId "2828141535" @default.
- W2828141535 workType "article" @default.