Matches in SemOpenAlex for { <https://semopenalex.org/work/W2963753390> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W2963753390 abstract "Adversarial samples are maliciously created inputs that lead a learning-based classifier to produce incorrect output labels. An adversarial sample is often generated by adding adversarial perturbation (AP) to a normal test sample. Recent studies that tried to analyze classifiers under such AP are mostly empirical and provide little understanding of why. To fill this gap, we propose a theoretical framework for analyzing learning-based classifiers, especially deep neural networks (DNN) in the face of such AP. By using concepts from topology, this framework brings forth the key reasons why an adversarial can fool a classifier ($f_1$) and suggests a new focus on its oracle ($f_2$, like human annotators of that specific task). By investigating the topology relationship between two (pseudo)metric spaces corresponding to predictor $f_1$ and oracle $f_2$, we develop several ideal conditions that can determine if $f_1$ is always robust (strong-robust) against adversarial samples according to its $f_2$. The theoretical framework leads to a set of novel and complementary insights that have not been uncovered by the literature. Surprisingly our theorems find that just one extra irrelevant feature can make a classifier not strong-robust, and the right feature representation learning is the key to getting a classifier that is both accurate and strong robust. Empirically we find that Siamese architecture can be used to help DNN models get close to the desired topological relationship for strong-robustness, which in turn effectively improves its performance against AP." @default.
- W2963753390 created "2019-07-30" @default.
- W2963753390 creator A5009957868 @default.
- W2963753390 creator A5013925367 @default.
- W2963753390 creator A5071517081 @default.
- W2963753390 date "2017-04-24" @default.
- W2963753390 modified "2023-09-23" @default.
- W2963753390 title "A Theoretical Framework for Robustness of (Deep) Classifiers against Adversarial Samples" @default.
- W2963753390 hasPublicationYear "2017" @default.
- W2963753390 type Work @default.
- W2963753390 sameAs 2963753390 @default.
- W2963753390 citedByCount "8" @default.
- W2963753390 countsByYear W29637533902017 @default.
- W2963753390 countsByYear W29637533902018 @default.
- W2963753390 countsByYear W29637533902019 @default.
- W2963753390 countsByYear W29637533902020 @default.
- W2963753390 crossrefType "proceedings-article" @default.
- W2963753390 hasAuthorship W2963753390A5009957868 @default.
- W2963753390 hasAuthorship W2963753390A5013925367 @default.
- W2963753390 hasAuthorship W2963753390A5071517081 @default.
- W2963753390 hasConcept C104317684 @default.
- W2963753390 hasConcept C111919701 @default.
- W2963753390 hasConcept C115903868 @default.
- W2963753390 hasConcept C119857082 @default.
- W2963753390 hasConcept C154945302 @default.
- W2963753390 hasConcept C185592680 @default.
- W2963753390 hasConcept C199845137 @default.
- W2963753390 hasConcept C2984842247 @default.
- W2963753390 hasConcept C37736160 @default.
- W2963753390 hasConcept C41008148 @default.
- W2963753390 hasConcept C50644808 @default.
- W2963753390 hasConcept C55166926 @default.
- W2963753390 hasConcept C55493867 @default.
- W2963753390 hasConcept C63479239 @default.
- W2963753390 hasConcept C95623464 @default.
- W2963753390 hasConceptScore W2963753390C104317684 @default.
- W2963753390 hasConceptScore W2963753390C111919701 @default.
- W2963753390 hasConceptScore W2963753390C115903868 @default.
- W2963753390 hasConceptScore W2963753390C119857082 @default.
- W2963753390 hasConceptScore W2963753390C154945302 @default.
- W2963753390 hasConceptScore W2963753390C185592680 @default.
- W2963753390 hasConceptScore W2963753390C199845137 @default.
- W2963753390 hasConceptScore W2963753390C2984842247 @default.
- W2963753390 hasConceptScore W2963753390C37736160 @default.
- W2963753390 hasConceptScore W2963753390C41008148 @default.
- W2963753390 hasConceptScore W2963753390C50644808 @default.
- W2963753390 hasConceptScore W2963753390C55166926 @default.
- W2963753390 hasConceptScore W2963753390C55493867 @default.
- W2963753390 hasConceptScore W2963753390C63479239 @default.
- W2963753390 hasConceptScore W2963753390C95623464 @default.
- W2963753390 hasLocation W29637533901 @default.
- W2963753390 hasOpenAccess W2963753390 @default.
- W2963753390 hasPrimaryLocation W29637533901 @default.
- W2963753390 hasRelatedWork W2180612164 @default.
- W2963753390 hasRelatedWork W2194775991 @default.
- W2963753390 hasRelatedWork W2243397390 @default.
- W2963753390 hasRelatedWork W2774644650 @default.
- W2963753390 hasRelatedWork W2947792938 @default.
- W2963753390 hasRelatedWork W2963178695 @default.
- W2963753390 hasRelatedWork W2963207607 @default.
- W2963753390 hasRelatedWork W2963739340 @default.
- W2963753390 hasRelatedWork W2963744840 @default.
- W2963753390 hasRelatedWork W2963857521 @default.
- W2963753390 hasRelatedWork W2964082701 @default.
- W2963753390 hasRelatedWork W2964153729 @default.
- W2963753390 hasRelatedWork W2964253222 @default.
- W2963753390 hasRelatedWork W2995645057 @default.
- W2963753390 hasRelatedWork W3008051686 @default.
- W2963753390 hasRelatedWork W3028525609 @default.
- W2963753390 hasRelatedWork W3033258370 @default.
- W2963753390 hasRelatedWork W3122142687 @default.
- W2963753390 hasRelatedWork W3174229740 @default.
- W2963753390 hasRelatedWork W3200942016 @default.
- W2963753390 isParatext "false" @default.
- W2963753390 isRetracted "false" @default.
- W2963753390 magId "2963753390" @default.
- W2963753390 workType "article" @default.