Matches in SemOpenAlex for { <https://semopenalex.org/work/W3098020164> ?p ?o ?g. }
- W3098020164 endingPage "1134" @default.
- W3098020164 startingPage "1119" @default.
- W3098020164 abstract "In the past five years, deep learning methods have become state-of-the-art in solving various inverse problems. Before such approaches can find application in safety-critical fields, a verification of their reliability appears mandatory. Recent works have pointed out instabilities of deep neural networks for several image reconstruction tasks. In analogy to adversarial attacks in classification, it was shown that slight distortions in the input domain may cause severe artifacts. The present article sheds new light on this concern, by conducting an extensive study of the robustness of deep-learning-based algorithms for solving underdetermined inverse problems. This covers compressed sensing with Gaussian measurements as well as image recovery from Fourier and Radon measurements, including a real-world scenario for magnetic resonance imaging (using the NYU-fastMRI dataset). Our main focus is on computing adversarial perturbations of the measurements that maximize the reconstruction error. A distinctive feature of our approach is the quantitative and qualitative comparison with total-variation minimization, which serves as a provably robust reference method. In contrast to previous findings, our results reveal that standard end-to-end network architectures are not only resilient against statistical noise, but also against adversarial perturbations. All considered networks are trained by common deep learning techniques, without sophisticated defense strategies." @default.
- W3098020164 created "2020-11-23" @default.
- W3098020164 creator A5027021592 @default.
- W3098020164 creator A5030864586 @default.
- W3098020164 creator A5065171277 @default.
- W3098020164 date "2023-01-01" @default.
- W3098020164 modified "2023-10-17" @default.
- W3098020164 title "Solving Inverse Problems With Deep Neural Networks – Robustness Included?" @default.
- W3098020164 cites W143004564 @default.
- W3098020164 cites W1639211413 @default.
- W3098020164 cites W1885185971 @default.
- W3098020164 cites W1972150100 @default.
- W3098020164 cites W2011687187 @default.
- W3098020164 cites W2025223969 @default.
- W3098020164 cites W2029816571 @default.
- W3098020164 cites W2039939700 @default.
- W3098020164 cites W2045079045 @default.
- W3098020164 cites W2070665556 @default.
- W3098020164 cites W2098874108 @default.
- W3098020164 cites W2103559027 @default.
- W3098020164 cites W2107034604 @default.
- W3098020164 cites W2111406701 @default.
- W3098020164 cites W2112796928 @default.
- W3098020164 cites W2128633079 @default.
- W3098020164 cites W2133665775 @default.
- W3098020164 cites W2145096794 @default.
- W3098020164 cites W2194775991 @default.
- W3098020164 cites W2467545770 @default.
- W3098020164 cites W2559597482 @default.
- W3098020164 cites W2560447554 @default.
- W3098020164 cites W2562637781 @default.
- W3098020164 cites W2570202822 @default.
- W3098020164 cites W2574952845 @default.
- W3098020164 cites W2584483805 @default.
- W3098020164 cites W2594014149 @default.
- W3098020164 cites W2604388535 @default.
- W3098020164 cites W2611467245 @default.
- W3098020164 cites W2617128058 @default.
- W3098020164 cites W2743780012 @default.
- W3098020164 cites W2798302089 @default.
- W3098020164 cites W2890437351 @default.
- W3098020164 cites W2902695846 @default.
- W3098020164 cites W2911290743 @default.
- W3098020164 cites W2919115771 @default.
- W3098020164 cites W2924551358 @default.
- W3098020164 cites W2947704577 @default.
- W3098020164 cites W2952020389 @default.
- W3098020164 cites W2963178695 @default.
- W3098020164 cites W2963399478 @default.
- W3098020164 cites W2963446712 @default.
- W3098020164 cites W2963857521 @default.
- W3098020164 cites W2999511788 @default.
- W3098020164 cites W3002289736 @default.
- W3098020164 cites W3003650457 @default.
- W3098020164 cites W3021094251 @default.
- W3098020164 cites W3046819731 @default.
- W3098020164 cites W3100730608 @default.
- W3098020164 cites W3101765447 @default.
- W3098020164 cites W3103586216 @default.
- W3098020164 cites W3104324122 @default.
- W3098020164 cites W3122499440 @default.
- W3098020164 cites W3125195138 @default.
- W3098020164 cites W4210477505 @default.
- W3098020164 cites W613352991 @default.
- W3098020164 doi "https://doi.org/10.1109/tpami.2022.3148324" @default.
- W3098020164 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/35119999" @default.
- W3098020164 hasPublicationYear "2023" @default.
- W3098020164 type Work @default.
- W3098020164 sameAs 3098020164 @default.
- W3098020164 citedByCount "34" @default.
- W3098020164 countsByYear W30980201642020 @default.
- W3098020164 countsByYear W30980201642021 @default.
- W3098020164 countsByYear W30980201642022 @default.
- W3098020164 countsByYear W30980201642023 @default.
- W3098020164 crossrefType "journal-article" @default.
- W3098020164 hasAuthorship W3098020164A5027021592 @default.
- W3098020164 hasAuthorship W3098020164A5030864586 @default.
- W3098020164 hasAuthorship W3098020164A5065171277 @default.
- W3098020164 hasBestOaLocation W30980201642 @default.
- W3098020164 hasConcept C104317684 @default.
- W3098020164 hasConcept C108583219 @default.
- W3098020164 hasConcept C11413529 @default.
- W3098020164 hasConcept C119857082 @default.
- W3098020164 hasConcept C134306372 @default.
- W3098020164 hasConcept C135252773 @default.
- W3098020164 hasConcept C141379421 @default.
- W3098020164 hasConcept C147764199 @default.
- W3098020164 hasConcept C153180895 @default.
- W3098020164 hasConcept C154945302 @default.
- W3098020164 hasConcept C179690561 @default.
- W3098020164 hasConcept C185592680 @default.
- W3098020164 hasConcept C199360897 @default.
- W3098020164 hasConcept C2781067378 @default.
- W3098020164 hasConcept C2984842247 @default.
- W3098020164 hasConcept C33923547 @default.
- W3098020164 hasConcept C41008148 @default.
- W3098020164 hasConcept C50644808 @default.
- W3098020164 hasConcept C55493867 @default.