FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2912353637> ?p ?o ?g. }

• W2912353637 abstract "We employ triplet loss as a space embedding regularizer to boost classification performance. Standard architectures, like ResNet and DesneNet, are extended to support both losses with minimal hyper-parameter tuning. This promotes generality while fine-tuning pretrained networks. Triplet loss is a powerful surrogate for recently proposed embedding regularizers. Yet, it is avoided for large batch-size requirement and high computational cost. Through our experiments, we re-assess these assumptions. During inference, our network supports both classification and embedding tasks without any computational overhead. Quantitative evaluation highlights how our approach compares favorably to the existing state of the art on multiple fine-grained recognition datasets. Further evaluation on an imbalanced video dataset achieves significant improvement (>7%). Beyond boosting efficiency, triplet loss brings retrieval and interpretability to classification models." @default.
• W2912353637 created "2019-02-21" @default.
• W2912353637 creator A5019389546 @default.
• W2912353637 creator A5055264952 @default.
• W2912353637 creator A5071123879 @default.
• W2912353637 creator A5084637778 @default.
• W2912353637 date "2019-01-24" @default.
• W2912353637 modified "2023-09-27" @default.
• W2912353637 title "In Defense of the Triplet Loss for Visual Recognition." @default.
• W2912353637 cites W1686810756 @default.
• W2912353637 cites W1846799578 @default.
• W2912353637 cites W1898560071 @default.
• W2912353637 cites W1950843348 @default.
• W2912353637 cites W2097117768 @default.
• W2912353637 cites W2108598243 @default.
• W2912353637 cites W2110015572 @default.
• W2912353637 cites W2138621090 @default.
• W2912353637 cites W2183341477 @default.
• W2912353637 cites W2194775991 @default.
• W2912353637 cites W2270409809 @default.
• W2912353637 cites W2274287116 @default.
• W2912353637 cites W2470322391 @default.
• W2912353637 cites W2520774990 @default.
• W2912353637 cites W2543665857 @default.
• W2912353637 cites W2550462002 @default.
• W2912353637 cites W2598634450 @default.
• W2912353637 cites W2612445135 @default.
• W2912353637 cites W2737701595 @default.
• W2912353637 cites W2740620254 @default.
• W2912353637 cites W2792096654 @default.
• W2912353637 cites W2794497862 @default.
• W2912353637 cites W2809446765 @default.
• W2912353637 cites W2953391683 @default.
• W2912353637 cites W2963026686 @default.
• W2912353637 cites W2963350250 @default.
• W2912353637 cites W2963446712 @default.
• W2912353637 cites W2798338071 @default.
• W2912353637 hasPublicationYear "2019" @default.
• W2912353637 type Work @default.
• W2912353637 sameAs 2912353637 @default.
• W2912353637 citedByCount "2" @default.
• W2912353637 countsByYear W29123536372019 @default.
• W2912353637 countsByYear W29123536372020 @default.
• W2912353637 crossrefType "posted-content" @default.
• W2912353637 hasAuthorship W2912353637A5019389546 @default.
• W2912353637 hasAuthorship W2912353637A5055264952 @default.
• W2912353637 hasAuthorship W2912353637A5071123879 @default.
• W2912353637 hasAuthorship W2912353637A5084637778 @default.
• W2912353637 hasConcept C108583219 @default.
• W2912353637 hasConcept C111919701 @default.
• W2912353637 hasConcept C119857082 @default.
• W2912353637 hasConcept C153180895 @default.
• W2912353637 hasConcept C154945302 @default.
• W2912353637 hasConcept C15744967 @default.
• W2912353637 hasConcept C2776214188 @default.
• W2912353637 hasConcept C2779960059 @default.
• W2912353637 hasConcept C2780767217 @default.
• W2912353637 hasConcept C2781067378 @default.
• W2912353637 hasConcept C2944601119 @default.
• W2912353637 hasConcept C41008148 @default.
• W2912353637 hasConcept C41608201 @default.
• W2912353637 hasConcept C46686674 @default.
• W2912353637 hasConcept C542102704 @default.
• W2912353637 hasConceptScore W2912353637C108583219 @default.
• W2912353637 hasConceptScore W2912353637C111919701 @default.
• W2912353637 hasConceptScore W2912353637C119857082 @default.
• W2912353637 hasConceptScore W2912353637C153180895 @default.
• W2912353637 hasConceptScore W2912353637C154945302 @default.
• W2912353637 hasConceptScore W2912353637C15744967 @default.
• W2912353637 hasConceptScore W2912353637C2776214188 @default.
• W2912353637 hasConceptScore W2912353637C2779960059 @default.
• W2912353637 hasConceptScore W2912353637C2780767217 @default.
• W2912353637 hasConceptScore W2912353637C2781067378 @default.
• W2912353637 hasConceptScore W2912353637C2944601119 @default.
• W2912353637 hasConceptScore W2912353637C41008148 @default.
• W2912353637 hasConceptScore W2912353637C41608201 @default.
• W2912353637 hasConceptScore W2912353637C46686674 @default.
• W2912353637 hasConceptScore W2912353637C542102704 @default.
• W2912353637 hasLocation W29123536371 @default.
• W2912353637 hasOpenAccess W2912353637 @default.
• W2912353637 hasPrimaryLocation W29123536371 @default.
• W2912353637 hasRelatedWork W2606611007 @default.
• W2912353637 hasRelatedWork W2887723086 @default.
• W2912353637 hasRelatedWork W2921286775 @default.
• W2912353637 hasRelatedWork W2937923353 @default.
• W2912353637 hasRelatedWork W2945569004 @default.
• W2912353637 hasRelatedWork W2947620140 @default.
• W2912353637 hasRelatedWork W2963175113 @default.
• W2912353637 hasRelatedWork W2963916311 @default.
• W2912353637 hasRelatedWork W2965843558 @default.
• W2912353637 hasRelatedWork W2986365236 @default.
• W2912353637 hasRelatedWork W2995590249 @default.
• W2912353637 hasRelatedWork W2996406913 @default.
• W2912353637 hasRelatedWork W3000295996 @default.
• W2912353637 hasRelatedWork W3119896582 @default.
• W2912353637 hasRelatedWork W3121615128 @default.
• W2912353637 hasRelatedWork W3125559414 @default.
• W2912353637 hasRelatedWork W3161703143 @default.
• W2912353637 hasRelatedWork W3172429633 @default.
• W2912353637 hasRelatedWork W3194173591 @default.

• next