Matches in SemOpenAlex for { <https://semopenalex.org/work/W3007517706> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W3007517706 endingPage "103550" @default.
- W3007517706 startingPage "103550" @default.
- W3007517706 abstract "In many real-world applications of Machine Learning it is of paramount importance not only to provide accurate predictions, but also to ensure certain levels of robustness. Adversarial Training is a training procedure aiming at providing models that are robust to worst-case perturbations around predefined points. Unfortunately, one of the main issues in adversarial training is that robustness w.r.t. gradient-based attackers is always achieved at the cost of prediction accuracy. In this paper, a new algorithm, called Wasserstein Projected Gradient Descent (WPGD), for adversarial training is proposed. WPGD provides a simple way to obtain cost-sensitive robustness, resulting in a finer control of the robustness-accuracy trade-off. Moreover, WPGD solves an optimal transport problem on the output space of the network and it can efficiently discover directions where robustness is required, allowing to control the directional trade-off between accuracy and robustness. The proposed WPGD is validated in this work on image recognition tasks with different benchmark datasets and architectures. Moreover, real world-like datasets are often imbalanced: this paper shows that when dealing with such type of datasets, the performance of adversarial training are mainly affected in term of standard accuracy." @default.
- W3007517706 created "2020-03-06" @default.
- W3007517706 creator A5006038667 @default.
- W3007517706 creator A5026617079 @default.
- W3007517706 creator A5032614224 @default.
- W3007517706 date "2020-05-01" @default.
- W3007517706 modified "2023-10-16" @default.
- W3007517706 title "Directional adversarial training for cost sensitive deep learning classification applications" @default.
- W3007517706 cites W1990283121 @default.
- W3007517706 cites W1999393241 @default.
- W3007517706 cites W2081580037 @default.
- W3007517706 cites W2112796928 @default.
- W3007517706 cites W2795863078 @default.
- W3007517706 cites W2808993481 @default.
- W3007517706 cites W2884001105 @default.
- W3007517706 cites W2893602881 @default.
- W3007517706 cites W2898365215 @default.
- W3007517706 cites W2905810301 @default.
- W3007517706 cites W2912581782 @default.
- W3007517706 cites W2919115771 @default.
- W3007517706 cites W2956698489 @default.
- W3007517706 cites W2963739340 @default.
- W3007517706 doi "https://doi.org/10.1016/j.engappai.2020.103550" @default.
- W3007517706 hasPublicationYear "2020" @default.
- W3007517706 type Work @default.
- W3007517706 sameAs 3007517706 @default.
- W3007517706 citedByCount "17" @default.
- W3007517706 countsByYear W30075177062020 @default.
- W3007517706 countsByYear W30075177062021 @default.
- W3007517706 countsByYear W30075177062022 @default.
- W3007517706 countsByYear W30075177062023 @default.
- W3007517706 crossrefType "journal-article" @default.
- W3007517706 hasAuthorship W3007517706A5006038667 @default.
- W3007517706 hasAuthorship W3007517706A5026617079 @default.
- W3007517706 hasAuthorship W3007517706A5032614224 @default.
- W3007517706 hasBestOaLocation W30075177062 @default.
- W3007517706 hasConcept C104317684 @default.
- W3007517706 hasConcept C119857082 @default.
- W3007517706 hasConcept C153258448 @default.
- W3007517706 hasConcept C154945302 @default.
- W3007517706 hasConcept C185592680 @default.
- W3007517706 hasConcept C37736160 @default.
- W3007517706 hasConcept C41008148 @default.
- W3007517706 hasConcept C50644808 @default.
- W3007517706 hasConcept C55493867 @default.
- W3007517706 hasConcept C63479239 @default.
- W3007517706 hasConceptScore W3007517706C104317684 @default.
- W3007517706 hasConceptScore W3007517706C119857082 @default.
- W3007517706 hasConceptScore W3007517706C153258448 @default.
- W3007517706 hasConceptScore W3007517706C154945302 @default.
- W3007517706 hasConceptScore W3007517706C185592680 @default.
- W3007517706 hasConceptScore W3007517706C37736160 @default.
- W3007517706 hasConceptScore W3007517706C41008148 @default.
- W3007517706 hasConceptScore W3007517706C50644808 @default.
- W3007517706 hasConceptScore W3007517706C55493867 @default.
- W3007517706 hasConceptScore W3007517706C63479239 @default.
- W3007517706 hasFunder F4320309480 @default.
- W3007517706 hasFunder F4320310598 @default.
- W3007517706 hasLocation W30075177061 @default.
- W3007517706 hasLocation W30075177062 @default.
- W3007517706 hasOpenAccess W3007517706 @default.
- W3007517706 hasPrimaryLocation W30075177061 @default.
- W3007517706 hasRelatedWork W2979999556 @default.
- W3007517706 hasRelatedWork W3007517706 @default.
- W3007517706 hasRelatedWork W3046550781 @default.
- W3007517706 hasRelatedWork W3046843850 @default.
- W3007517706 hasRelatedWork W4288094329 @default.
- W3007517706 hasRelatedWork W4309156448 @default.
- W3007517706 hasRelatedWork W4311734044 @default.
- W3007517706 hasRelatedWork W4315780078 @default.
- W3007517706 hasRelatedWork W4379258830 @default.
- W3007517706 hasRelatedWork W4386716251 @default.
- W3007517706 hasVolume "91" @default.
- W3007517706 isParatext "false" @default.
- W3007517706 isRetracted "false" @default.
- W3007517706 magId "3007517706" @default.
- W3007517706 workType "article" @default.