Matches in SemOpenAlex for { <https://semopenalex.org/work/W3039594701> ?p ?o ?g. }
Showing items 1 to 84 of
84
with 100 items per page.
- W3039594701 abstract "Test-time adversarial attacks have posed serious challenges to the robustness of machine learning models, and in many settings the adversarial manipulation needs not be bounded by small $ell_p$-norms. Motivated by semantic-preserving attacks in security domain, we investigate logical adversaries, a broad class of attackers who create adversarial examples within a reflexive-transitive closure of a logical relation. We analyze the conditions for robustness and propose normalize-and-predict -- a learning framework with provable robustness guarantee. We compare our approach with adversarial training and derive a unified framework that provides the benefits of both approaches.Driven by the theoretical findings, we apply our framework to malware detection. We use our framework to learn new detectors and propose two generic logical attacks to validate model robustness. Experiment results on real-world data set show that attacks using logical relations can evade existing detectors, and our unified framework can significantly enhance model robustness." @default.
- W3039594701 created "2020-07-10" @default.
- W3039594701 creator A5027198310 @default.
- W3039594701 creator A5047581480 @default.
- W3039594701 creator A5050855162 @default.
- W3039594701 creator A5088826068 @default.
- W3039594701 date "2020-07-01" @default.
- W3039594701 modified "2023-09-27" @default.
- W3039594701 title "Robust Learning against Logical Adversaries." @default.
- W3039594701 cites W1553177637 @default.
- W3039594701 cites W2039529128 @default.
- W3039594701 cites W2243397390 @default.
- W3039594701 cites W2594877703 @default.
- W3039594701 cites W2626166918 @default.
- W3039594701 cites W2744095836 @default.
- W3039594701 cites W2886812138 @default.
- W3039594701 cites W2911919851 @default.
- W3039594701 cites W2913266441 @default.
- W3039594701 cites W2963207607 @default.
- W3039594701 cites W2964153729 @default.
- W3039594701 cites W2995404272 @default.
- W3039594701 cites W3037012516 @default.
- W3039594701 hasPublicationYear "2020" @default.
- W3039594701 type Work @default.
- W3039594701 sameAs 3039594701 @default.
- W3039594701 citedByCount "0" @default.
- W3039594701 crossrefType "posted-content" @default.
- W3039594701 hasAuthorship W3039594701A5027198310 @default.
- W3039594701 hasAuthorship W3039594701A5047581480 @default.
- W3039594701 hasAuthorship W3039594701A5050855162 @default.
- W3039594701 hasAuthorship W3039594701A5088826068 @default.
- W3039594701 hasConcept C104317684 @default.
- W3039594701 hasConcept C119857082 @default.
- W3039594701 hasConcept C134306372 @default.
- W3039594701 hasConcept C154945302 @default.
- W3039594701 hasConcept C185592680 @default.
- W3039594701 hasConcept C33923547 @default.
- W3039594701 hasConcept C34388435 @default.
- W3039594701 hasConcept C37736160 @default.
- W3039594701 hasConcept C38652104 @default.
- W3039594701 hasConcept C41008148 @default.
- W3039594701 hasConcept C55493867 @default.
- W3039594701 hasConcept C63479239 @default.
- W3039594701 hasConcept C80444323 @default.
- W3039594701 hasConceptScore W3039594701C104317684 @default.
- W3039594701 hasConceptScore W3039594701C119857082 @default.
- W3039594701 hasConceptScore W3039594701C134306372 @default.
- W3039594701 hasConceptScore W3039594701C154945302 @default.
- W3039594701 hasConceptScore W3039594701C185592680 @default.
- W3039594701 hasConceptScore W3039594701C33923547 @default.
- W3039594701 hasConceptScore W3039594701C34388435 @default.
- W3039594701 hasConceptScore W3039594701C37736160 @default.
- W3039594701 hasConceptScore W3039594701C38652104 @default.
- W3039594701 hasConceptScore W3039594701C41008148 @default.
- W3039594701 hasConceptScore W3039594701C55493867 @default.
- W3039594701 hasConceptScore W3039594701C63479239 @default.
- W3039594701 hasConceptScore W3039594701C80444323 @default.
- W3039594701 hasLocation W30395947011 @default.
- W3039594701 hasOpenAccess W3039594701 @default.
- W3039594701 hasPrimaryLocation W30395947011 @default.
- W3039594701 hasRelatedWork W1515104600 @default.
- W3039594701 hasRelatedWork W166153009 @default.
- W3039594701 hasRelatedWork W2116683577 @default.
- W3039594701 hasRelatedWork W2157301192 @default.
- W3039594701 hasRelatedWork W2402894935 @default.
- W3039594701 hasRelatedWork W2500043887 @default.
- W3039594701 hasRelatedWork W2512472178 @default.
- W3039594701 hasRelatedWork W2888981498 @default.
- W3039594701 hasRelatedWork W2898268570 @default.
- W3039594701 hasRelatedWork W2914925440 @default.
- W3039594701 hasRelatedWork W2950496531 @default.
- W3039594701 hasRelatedWork W2963952467 @default.
- W3039594701 hasRelatedWork W2969320744 @default.
- W3039594701 hasRelatedWork W2989093880 @default.
- W3039594701 hasRelatedWork W3006165421 @default.
- W3039594701 hasRelatedWork W3022774529 @default.
- W3039594701 hasRelatedWork W3097002519 @default.
- W3039594701 hasRelatedWork W3152222778 @default.
- W3039594701 hasRelatedWork W69351110 @default.
- W3039594701 hasRelatedWork W71289362 @default.
- W3039594701 isParatext "false" @default.
- W3039594701 isRetracted "false" @default.
- W3039594701 magId "3039594701" @default.
- W3039594701 workType "article" @default.