Self-distillation with Online Diﬀusion on Batch Manifolds Improves Deep Metric Learning

doi:doi:10.57702/urle1zln

Self-distillation with Online Diﬀusion on Batch Manifolds Improves Deep Metric Learning

Deep metric learning (DML) methods typically leverage solely class labels to keep positive samples far away from negative ones. However, this type of method normally ignores the crucial knowledge hidden in the data (e.g., intra-class information variation), which is harmful to the generalization of the trained model. To alleviate this problem, in this paper we propose Online Batch Diﬀusion-based Self-Distillation (OBD-SD) for DML.

BibTex:

@dataset{Zelong_Zeng_and_Hong_Liu_and_Fan_Yang_and_Shin'ichi_Satoh_2024,
    abstract = {Deep metric learning (DML) methods typically leverage solely class labels to keep positive samples far away from negative ones. However, this type of method normally ignores the crucial knowledge hidden in the data (e.g., intra-class information variation), which is harmful to the generalization of the trained model. To alleviate this problem, in this paper we propose Online Batch Diﬀusion-based Self-Distillation (OBD-SD) for DML.},
    author = {Zelong Zeng and Hong Liu and Fan Yang and Shin'ichi Satoh},
    doi = {10.57702/urle1zln},
    institution = {No Organization},
    keyword = {'Batch Manifolds', 'Birds', 'CARS196', 'CUB200-2011', 'Cars', 'Deep Metric Learning', 'Image Classification', 'Online Diﬀusion', 'Online Products', 'Self-Distillation', 'Stanford Online Products'},
    month = {dec},
    publisher = {TIB},
    title = {Self-distillation with Online Diﬀusion on Batch Manifolds Improves Deep Metric Learning},
    url = {https://service.tib.eu/ldmservice/dataset/self-distillation-with-online-di-usion-on-batch-manifolds-improves-deep-metric-learning},
    year = {2024}
}