Matches in SemOpenAlex for { <https://semopenalex.org/work/W2904967782> ?p ?o ?g. }
- W2904967782 abstract "We propose a robust adversarial prediction framework for general multiclass classification. Our method seeks predictive distributions that robustly optimize non-convex and non-continuous multiclass loss metrics against the worst-case conditional label distributions (the adversarial distributions) that (approximately) match the statistics of the training data. Although the optimized loss metrics are non-convex and non-continuous, the dual formulation of the framework is a convex optimization problem that can be recast as a risk minimization model with a prescribed convex surrogate loss we call the adversarial surrogate loss. We show that the adversarial surrogate losses fill an existing gap in surrogate loss construction for general multiclass classification problems, by simultaneously aligning better with the original multiclass loss, guaranteeing Fisher consistency, enabling a way to incorporate rich feature spaces via the kernel trick, and providing competitive performance in practice." @default.
- W2904967782 created "2018-12-22" @default.
- W2904967782 creator A5005784479 @default.
- W2904967782 creator A5023212176 @default.
- W2904967782 creator A5023610050 @default.
- W2904967782 creator A5024686404 @default.
- W2904967782 creator A5038911681 @default.
- W2904967782 creator A5047206360 @default.
- W2904967782 creator A5068807179 @default.
- W2904967782 creator A5077576020 @default.
- W2904967782 date "2018-12-18" @default.
- W2904967782 modified "2023-09-27" @default.
- W2904967782 title "Consistent Robust Adversarial Prediction for General Multiclass Classification." @default.
- W2904967782 cites W1480538416 @default.
- W2904967782 cites W1530144005 @default.
- W2904967782 cites W1585021961 @default.
- W2904967782 cites W1590693676 @default.
- W2904967782 cites W1698155719 @default.
- W2904967782 cites W1968355947 @default.
- W2904967782 cites W1982032418 @default.
- W2904967782 cites W1985077192 @default.
- W2904967782 cites W1988790447 @default.
- W2904967782 cites W1992736075 @default.
- W2904967782 cites W2023508744 @default.
- W2904967782 cites W2041615247 @default.
- W2904967782 cites W2076680677 @default.
- W2904967782 cites W2087347434 @default.
- W2904967782 cites W2107490382 @default.
- W2904967782 cites W2116444583 @default.
- W2904967782 cites W2116458520 @default.
- W2904967782 cites W2119821739 @default.
- W2904967782 cites W2123358413 @default.
- W2904967782 cites W2124105163 @default.
- W2904967782 cites W2125993116 @default.
- W2904967782 cites W2127470768 @default.
- W2904967782 cites W2128186735 @default.
- W2904967782 cites W2135219455 @default.
- W2904967782 cites W2142575165 @default.
- W2904967782 cites W2143747826 @default.
- W2904967782 cites W2144846366 @default.
- W2904967782 cites W2155195660 @default.
- W2904967782 cites W2157791002 @default.
- W2904967782 cites W2161813894 @default.
- W2904967782 cites W2169610 @default.
- W2904967782 cites W2188209982 @default.
- W2904967782 cites W2296319761 @default.
- W2904967782 cites W2395964374 @default.
- W2904967782 cites W2398429633 @default.
- W2904967782 cites W2403763698 @default.
- W2904967782 cites W2549334782 @default.
- W2904967782 cites W2555956183 @default.
- W2904967782 cites W2752012466 @default.
- W2904967782 cites W2788508747 @default.
- W2904967782 cites W2804347066 @default.
- W2904967782 cites W2963343251 @default.
- W2904967782 cites W3120740533 @default.
- W2904967782 cites W55145144 @default.
- W2904967782 cites W2586016702 @default.
- W2904967782 hasPublicationYear "2018" @default.
- W2904967782 type Work @default.
- W2904967782 sameAs 2904967782 @default.
- W2904967782 citedByCount "4" @default.
- W2904967782 countsByYear W29049677822019 @default.
- W2904967782 countsByYear W29049677822020 @default.
- W2904967782 crossrefType "posted-content" @default.
- W2904967782 hasAuthorship W2904967782A5005784479 @default.
- W2904967782 hasAuthorship W2904967782A5023212176 @default.
- W2904967782 hasAuthorship W2904967782A5023610050 @default.
- W2904967782 hasAuthorship W2904967782A5024686404 @default.
- W2904967782 hasAuthorship W2904967782A5038911681 @default.
- W2904967782 hasAuthorship W2904967782A5047206360 @default.
- W2904967782 hasAuthorship W2904967782A5068807179 @default.
- W2904967782 hasAuthorship W2904967782A5077576020 @default.
- W2904967782 hasConcept C112680207 @default.
- W2904967782 hasConcept C114614502 @default.
- W2904967782 hasConcept C119857082 @default.
- W2904967782 hasConcept C12267149 @default.
- W2904967782 hasConcept C123860398 @default.
- W2904967782 hasConcept C126255220 @default.
- W2904967782 hasConcept C147764199 @default.
- W2904967782 hasConcept C154945302 @default.
- W2904967782 hasConcept C157972887 @default.
- W2904967782 hasConcept C2524010 @default.
- W2904967782 hasConcept C2776436953 @default.
- W2904967782 hasConcept C33923547 @default.
- W2904967782 hasConcept C37736160 @default.
- W2904967782 hasConcept C39891107 @default.
- W2904967782 hasConcept C41008148 @default.
- W2904967782 hasConcept C74193536 @default.
- W2904967782 hasConceptScore W2904967782C112680207 @default.
- W2904967782 hasConceptScore W2904967782C114614502 @default.
- W2904967782 hasConceptScore W2904967782C119857082 @default.
- W2904967782 hasConceptScore W2904967782C12267149 @default.
- W2904967782 hasConceptScore W2904967782C123860398 @default.
- W2904967782 hasConceptScore W2904967782C126255220 @default.
- W2904967782 hasConceptScore W2904967782C147764199 @default.
- W2904967782 hasConceptScore W2904967782C154945302 @default.
- W2904967782 hasConceptScore W2904967782C157972887 @default.
- W2904967782 hasConceptScore W2904967782C2524010 @default.
- W2904967782 hasConceptScore W2904967782C2776436953 @default.