Matches in SemOpenAlex for { <https://semopenalex.org/work/W3134911507> ?p ?o ?g. }
- W3134911507 abstract "Few-shot class incremental learning (FSCIL) portrays the problem of learning new concepts gradually, where only a few examples per concept are available to the learner. Due to the limited number of examples for training, the techniques developed for standard incremental learning cannot be applied verbatim to FSCIL. In this work, we introduce a distillation algorithm to address the problem of FSCIL and propose to make use of semantic information during training. To this end, we make use of word embeddings as semantic information which is cheap to obtain and which facilitate the distillation process. Furthermore, we propose a method based on an attention mechanism on multiple parallel embeddings of visual data to align visual and semantic vectors, which reduces issues related to catastrophic forgetting. Via experiments on MiniImageNet, CUB200, and CIFAR100 dataset, we establish new state-of-the-art results by outperforming existing approaches." @default.
- W3134911507 created "2021-03-15" @default.
- W3134911507 creator A5005651230 @default.
- W3134911507 creator A5007895165 @default.
- W3134911507 creator A5021790939 @default.
- W3134911507 creator A5047166819 @default.
- W3134911507 creator A5058541858 @default.
- W3134911507 creator A5060335798 @default.
- W3134911507 date "2021-03-06" @default.
- W3134911507 modified "2023-10-07" @default.
- W3134911507 title "Semantic-aware Knowledge Distillation for Few-Shot Class-Incremental Learning" @default.
- W3134911507 cites W1522301498 @default.
- W3134911507 cites W1682403713 @default.
- W3134911507 cites W1797268635 @default.
- W3134911507 cites W1821462560 @default.
- W3134911507 cites W2153579005 @default.
- W3134911507 cites W2194775991 @default.
- W3134911507 cites W2250539671 @default.
- W3134911507 cites W2334493732 @default.
- W3134911507 cites W2400717490 @default.
- W3134911507 cites W2473930607 @default.
- W3134911507 cites W2737492962 @default.
- W3134911507 cites W2787501667 @default.
- W3134911507 cites W2796346823 @default.
- W3134911507 cites W2798836702 @default.
- W3134911507 cites W2884282566 @default.
- W3134911507 cites W2939137134 @default.
- W3134911507 cites W2948734064 @default.
- W3134911507 cites W2954929116 @default.
- W3134911507 cites W2962966271 @default.
- W3134911507 cites W2963341924 @default.
- W3134911507 cites W2963390791 @default.
- W3134911507 cites W2963559848 @default.
- W3134911507 cites W2963577698 @default.
- W3134911507 cites W2964189064 @default.
- W3134911507 cites W2970689606 @default.
- W3134911507 cites W2985891137 @default.
- W3134911507 cites W3000133216 @default.
- W3134911507 cites W3009201938 @default.
- W3134911507 cites W3010806304 @default.
- W3134911507 cites W3013325675 @default.
- W3134911507 cites W3021349224 @default.
- W3134911507 cites W3034503642 @default.
- W3134911507 cites W3034856281 @default.
- W3134911507 cites W3035342403 @default.
- W3134911507 cites W3109728025 @default.
- W3134911507 cites W3134946523 @default.
- W3134911507 cites W3175516982 @default.
- W3134911507 cites W3176276396 @default.
- W3134911507 hasPublicationYear "2021" @default.
- W3134911507 type Work @default.
- W3134911507 sameAs 3134911507 @default.
- W3134911507 citedByCount "3" @default.
- W3134911507 countsByYear W31349115072021 @default.
- W3134911507 crossrefType "posted-content" @default.
- W3134911507 hasAuthorship W3134911507A5005651230 @default.
- W3134911507 hasAuthorship W3134911507A5007895165 @default.
- W3134911507 hasAuthorship W3134911507A5021790939 @default.
- W3134911507 hasAuthorship W3134911507A5047166819 @default.
- W3134911507 hasAuthorship W3134911507A5058541858 @default.
- W3134911507 hasAuthorship W3134911507A5060335798 @default.
- W3134911507 hasConcept C111919701 @default.
- W3134911507 hasConcept C119857082 @default.
- W3134911507 hasConcept C138885662 @default.
- W3134911507 hasConcept C154945302 @default.
- W3134911507 hasConcept C178790620 @default.
- W3134911507 hasConcept C185592680 @default.
- W3134911507 hasConcept C204030448 @default.
- W3134911507 hasConcept C204321447 @default.
- W3134911507 hasConcept C2777212361 @default.
- W3134911507 hasConcept C2780735816 @default.
- W3134911507 hasConcept C41008148 @default.
- W3134911507 hasConcept C41895202 @default.
- W3134911507 hasConcept C7149132 @default.
- W3134911507 hasConcept C98045186 @default.
- W3134911507 hasConceptScore W3134911507C111919701 @default.
- W3134911507 hasConceptScore W3134911507C119857082 @default.
- W3134911507 hasConceptScore W3134911507C138885662 @default.
- W3134911507 hasConceptScore W3134911507C154945302 @default.
- W3134911507 hasConceptScore W3134911507C178790620 @default.
- W3134911507 hasConceptScore W3134911507C185592680 @default.
- W3134911507 hasConceptScore W3134911507C204030448 @default.
- W3134911507 hasConceptScore W3134911507C204321447 @default.
- W3134911507 hasConceptScore W3134911507C2777212361 @default.
- W3134911507 hasConceptScore W3134911507C2780735816 @default.
- W3134911507 hasConceptScore W3134911507C41008148 @default.
- W3134911507 hasConceptScore W3134911507C41895202 @default.
- W3134911507 hasConceptScore W3134911507C7149132 @default.
- W3134911507 hasConceptScore W3134911507C98045186 @default.
- W3134911507 hasOpenAccess W3134911507 @default.
- W3134911507 hasRelatedWork W2942862636 @default.
- W3134911507 hasRelatedWork W2964512962 @default.
- W3134911507 hasRelatedWork W2967339675 @default.
- W3134911507 hasRelatedWork W3015113763 @default.
- W3134911507 hasRelatedWork W3021669159 @default.
- W3134911507 hasRelatedWork W3034451759 @default.
- W3134911507 hasRelatedWork W3034987881 @default.
- W3134911507 hasRelatedWork W3036411233 @default.
- W3134911507 hasRelatedWork W3038568084 @default.
- W3134911507 hasRelatedWork W3044396630 @default.