Matches in SemOpenAlex for { <https://semopenalex.org/work/W2772217839> ?p ?o ?g. }
- W2772217839 abstract "We propose a novel framework called Semantics-Preserving Adversarial Embedding Network (SP-AEN) for zero-shot visual recognition (ZSL), where test images and their classes are both unseen during training. SP-AEN aims to tackle the inherent problem --- semantic loss --- in the prevailing family of embedding-based ZSL, where some semantics would be discarded during training if they are non-discriminative for training classes, but could become critical for recognizing test classes. Specifically, SP-AEN prevents the semantic loss by introducing an independent visual-to-semantic space embedder which disentangles the semantic space into two subspaces for the two arguably conflicting objectives: classification and reconstruction. Through adversarial learning of the two subspaces, SP-AEN can transfer the semantics from the reconstructive subspace to the discriminative one, accomplishing the improved zero-shot recognition of unseen classes. Comparing with prior works, SP-AEN can not only improve classification but also generate photo-realistic images, demonstrating the effectiveness of semantic preservation. On four popular benchmarks: CUB, AWA, SUN and aPY, SP-AEN considerably outperforms other state-of-the-art methods by an absolute performance difference of 12.2%, 9.3%, 4.0%, and 3.6% in terms of harmonic mean values" @default.
- W2772217839 created "2017-12-22" @default.
- W2772217839 creator A5012324763 @default.
- W2772217839 creator A5012943082 @default.
- W2772217839 creator A5017855052 @default.
- W2772217839 creator A5042324027 @default.
- W2772217839 creator A5071037763 @default.
- W2772217839 date "2017-12-05" @default.
- W2772217839 modified "2023-10-18" @default.
- W2772217839 title "Zero-Shot Visual Recognition using Semantics-Preserving Adversarial Embedding Networks" @default.
- W2772217839 cites W1677182931 @default.
- W2772217839 cites W1722318740 @default.
- W2772217839 cites W1849277567 @default.
- W2772217839 cites W1858576077 @default.
- W2772217839 cites W2028285070 @default.
- W2772217839 cites W2070148066 @default.
- W2772217839 cites W2098411764 @default.
- W2772217839 cites W2099471712 @default.
- W2772217839 cites W2102765684 @default.
- W2772217839 cites W2117539524 @default.
- W2772217839 cites W2122528955 @default.
- W2772217839 cites W2123024445 @default.
- W2772217839 cites W2134270519 @default.
- W2772217839 cites W2141350700 @default.
- W2772217839 cites W2161208721 @default.
- W2772217839 cites W2169177311 @default.
- W2772217839 cites W2187089797 @default.
- W2772217839 cites W2250539671 @default.
- W2772217839 cites W2250646737 @default.
- W2772217839 cites W2322020277 @default.
- W2772217839 cites W2331128040 @default.
- W2772217839 cites W2523714292 @default.
- W2772217839 cites W2548275288 @default.
- W2772217839 cites W2557449848 @default.
- W2772217839 cites W2564591810 @default.
- W2772217839 cites W2608015370 @default.
- W2772217839 cites W2611345819 @default.
- W2772217839 cites W2737435850 @default.
- W2772217839 cites W2753880617 @default.
- W2772217839 cites W2765137706 @default.
- W2772217839 cites W2779610669 @default.
- W2772217839 cites W2949503252 @default.
- W2772217839 cites W2949650786 @default.
- W2772217839 cites W2949823873 @default.
- W2772217839 cites W2949906299 @default.
- W2772217839 cites W2949987290 @default.
- W2772217839 cites W2949999304 @default.
- W2772217839 cites W2950015165 @default.
- W2772217839 cites W2950133940 @default.
- W2772217839 cites W2950276680 @default.
- W2772217839 cites W2950319135 @default.
- W2772217839 cites W2950359962 @default.
- W2772217839 cites W2950580379 @default.
- W2772217839 cites W2950700180 @default.
- W2772217839 cites W2951313160 @default.
- W2772217839 cites W2951584201 @default.
- W2772217839 cites W2951871104 @default.
- W2772217839 cites W2951939904 @default.
- W2772217839 cites W2962793481 @default.
- W2772217839 cites W2962835968 @default.
- W2772217839 cites W2963078860 @default.
- W2772217839 cites W2963118547 @default.
- W2772217839 cites W2963174698 @default.
- W2772217839 cites W2963709863 @default.
- W2772217839 cites W3037567775 @default.
- W2772217839 cites W3038022805 @default.
- W2772217839 cites W652269744 @default.
- W2772217839 cites W870084106 @default.
- W2772217839 cites W14333344 @default.
- W2772217839 doi "https://doi.org/10.48550/arxiv.1712.01928" @default.
- W2772217839 hasPublicationYear "2017" @default.
- W2772217839 type Work @default.
- W2772217839 sameAs 2772217839 @default.
- W2772217839 citedByCount "1" @default.
- W2772217839 countsByYear W27722178392018 @default.
- W2772217839 crossrefType "posted-content" @default.
- W2772217839 hasAuthorship W2772217839A5012324763 @default.
- W2772217839 hasAuthorship W2772217839A5012943082 @default.
- W2772217839 hasAuthorship W2772217839A5017855052 @default.
- W2772217839 hasAuthorship W2772217839A5042324027 @default.
- W2772217839 hasAuthorship W2772217839A5071037763 @default.
- W2772217839 hasBestOaLocation W27722178391 @default.
- W2772217839 hasConcept C115961682 @default.
- W2772217839 hasConcept C12362212 @default.
- W2772217839 hasConcept C138885662 @default.
- W2772217839 hasConcept C153180895 @default.
- W2772217839 hasConcept C154945302 @default.
- W2772217839 hasConcept C184337299 @default.
- W2772217839 hasConcept C199360897 @default.
- W2772217839 hasConcept C202444582 @default.
- W2772217839 hasConcept C2780813799 @default.
- W2772217839 hasConcept C32834561 @default.
- W2772217839 hasConcept C33923547 @default.
- W2772217839 hasConcept C37736160 @default.
- W2772217839 hasConcept C41008148 @default.
- W2772217839 hasConcept C41608201 @default.
- W2772217839 hasConcept C41895202 @default.
- W2772217839 hasConcept C97931131 @default.
- W2772217839 hasConceptScore W2772217839C115961682 @default.
- W2772217839 hasConceptScore W2772217839C12362212 @default.