FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386076647> ?p ?o ?g. }

• W4386076647 abstract "Learning a generalizable and comprehensive similarity metric to depict the semantic discrepancies between images is the foundation of many computer vision tasks. While existing methods approach this goal by learning an ensemble of embeddings with diverse objectives, the backbone network still receives a mix of all the training signals. Differently, we propose a deep factorized metric learning (DFML) method to factorize the training signal and employ different samples to train various components of the backbone network. We factorize the network to different sub-blocks and devise a learnable router to adaptively allocate the training samples to each sub-block with the objective to capture the most information. The metric model trained by DFML capture different characteristics with different sub-blocks and constitutes a generalizable metric when using all the sub-blocks. The proposed DFML achieves state-of-the-art performance on all three benchmarks for deep metric learning including CUB-200-20ll, Cars196, and Stanford Online Products. We also generalize DFML to the image classification task on ImageNet-1K and observe consistent improvement in accuracy/computation trade-off. Specifically, we improve the performance of ViT-B on ImageNet (+0.2% accuracy) with less computation load (-24% FLOPs). <sup xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>1</sup> <sup xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>1</sup> Code is available at: https://github.com/wangck20/DFML." @default.
• W4386076647 created "2023-08-23" @default.
• W4386076647 creator A5006396086 @default.
• W4386076647 creator A5010357390 @default.
• W4386076647 creator A5010456925 @default.
• W4386076647 creator A5077320577 @default.
• W4386076647 creator A5090079801 @default.
• W4386076647 date "2023-06-01" @default.
• W4386076647 modified "2023-09-27" @default.
• W4386076647 title "Deep Factorized Metric Learning" @default.
• W4386076647 cites W1975517671 @default.
• W4386076647 cites W2043806097 @default.
• W4386076647 cites W2059216481 @default.
• W4386076647 cites W2069473864 @default.
• W4386076647 cites W2076434944 @default.
• W4386076647 cites W2108598243 @default.
• W4386076647 cites W2117539524 @default.
• W4386076647 cites W2135293965 @default.
• W4386076647 cites W2138011018 @default.
• W4386076647 cites W2549858646 @default.
• W4386076647 cites W2605102252 @default.
• W4386076647 cites W2606611007 @default.
• W4386076647 cites W2779066530 @default.
• W4386076647 cites W2783165089 @default.
• W4386076647 cites W2798303923 @default.
• W4386076647 cites W2885563408 @default.
• W4386076647 cites W2889625178 @default.
• W4386076647 cites W2895347732 @default.
• W4386076647 cites W2895608766 @default.
• W4386076647 cites W2921310091 @default.
• W4386076647 cites W2935456071 @default.
• W4386076647 cites W2948077755 @default.
• W4386076647 cites W2948638722 @default.
• W4386076647 cites W2953271441 @default.
• W4386076647 cites W2962723992 @default.
• W4386076647 cites W2962869940 @default.
• W4386076647 cites W2963026686 @default.
• W4386076647 cites W2963350250 @default.
• W4386076647 cites W2963988212 @default.
• W4386076647 cites W2964157492 @default.
• W4386076647 cites W2964271799 @default.
• W4386076647 cites W2979329071 @default.
• W4386076647 cites W2982392870 @default.
• W4386076647 cites W2984884190 @default.
• W4386076647 cites W2991234496 @default.
• W4386076647 cites W2991642706 @default.
• W4386076647 cites W2995182785 @default.
• W4386076647 cites W3034202663 @default.
• W4386076647 cites W3034303554 @default.
• W4386076647 cites W3035295689 @default.
• W4386076647 cites W3035748723 @default.
• W4386076647 cites W3108807630 @default.
• W4386076647 cites W3165924482 @default.
• W4386076647 cites W3175616662 @default.
• W4386076647 cites W3203003533 @default.
• W4386076647 cites W4214609630 @default.
• W4386076647 cites W4312533318 @default.
• W4386076647 cites W4313144963 @default.
• W4386076647 doi "https://doi.org/10.1109/cvpr52729.2023.00741" @default.
• W4386076647 hasPublicationYear "2023" @default.
• W4386076647 type Work @default.
• W4386076647 citedByCount "0" @default.
• W4386076647 crossrefType "proceedings-article" @default.
• W4386076647 hasAuthorship W4386076647A5006396086 @default.
• W4386076647 hasAuthorship W4386076647A5010357390 @default.
• W4386076647 hasAuthorship W4386076647A5010456925 @default.
• W4386076647 hasAuthorship W4386076647A5077320577 @default.
• W4386076647 hasAuthorship W4386076647A5090079801 @default.
• W4386076647 hasConcept C103278499 @default.
• W4386076647 hasConcept C104954878 @default.
• W4386076647 hasConcept C108583219 @default.
• W4386076647 hasConcept C11413529 @default.
• W4386076647 hasConcept C115961682 @default.
• W4386076647 hasConcept C119857082 @default.
• W4386076647 hasConcept C154945302 @default.
• W4386076647 hasConcept C162324750 @default.
• W4386076647 hasConcept C173608175 @default.
• W4386076647 hasConcept C176217482 @default.
• W4386076647 hasConcept C195780805 @default.
• W4386076647 hasConcept C204948658 @default.
• W4386076647 hasConcept C21547014 @default.
• W4386076647 hasConcept C2524010 @default.
• W4386076647 hasConcept C2777210771 @default.
• W4386076647 hasConcept C31258907 @default.
• W4386076647 hasConcept C33923547 @default.
• W4386076647 hasConcept C3826847 @default.
• W4386076647 hasConcept C41008148 @default.
• W4386076647 hasConcept C45374587 @default.
• W4386076647 hasConcept C74172769 @default.
• W4386076647 hasConcept C80444323 @default.
• W4386076647 hasConceptScore W4386076647C103278499 @default.
• W4386076647 hasConceptScore W4386076647C104954878 @default.
• W4386076647 hasConceptScore W4386076647C108583219 @default.
• W4386076647 hasConceptScore W4386076647C11413529 @default.
• W4386076647 hasConceptScore W4386076647C115961682 @default.
• W4386076647 hasConceptScore W4386076647C119857082 @default.
• W4386076647 hasConceptScore W4386076647C154945302 @default.
• W4386076647 hasConceptScore W4386076647C162324750 @default.
• W4386076647 hasConceptScore W4386076647C173608175 @default.
• W4386076647 hasConceptScore W4386076647C176217482 @default.

• next