Multi-level Distance Regularization for Deep Metric Learning

• Published in: arXiv.org
• Main Authors: Kim, Yonghyun, Park, Wonpyo
• Format: Paper
• Language: English
• Published: Ithaca Cornell University Library, arXiv.org 08.02.2021
• Subjects: Ablation; Bells; Embedding; Learning; Process parameters; Regularization; Regularization methods

We propose a novel distance-based regularization method for deep metric learning called Multi-level Distance Regularization (MDR). MDR explicitly disturbs a learning procedure by regularizing pairwise distances between embedding vectors into multiple levels that represents a degree of similarity bet...