Proxy Synthesis: Learning With Synthetic Classes For Deep Metric Learning

Abstract

One of the main purposes of deep metric learning is to construct an embedding space that has well-generalized embeddings on both seen (training) classes and unseen (test) classes. Most existing works have tried to achieve this using different types of metric objectives and hard sample mining strategies with given training data. However, learning with only the training data can be overfitted to the seen classes, leading to the lack of generalization capability on unseen classes. To address this problem, we propose a simple regularizer called Proxy Synthesis that exploits synthetic classes for stronger generalization in deep metric learning. The proposed method generates synthetic embeddings and proxies that work as synthetic classes, and they mimic unseen classes when computing proxy-based losses. Proxy Synthesis derives an embedding space considering class relations and smooth decision boundaries for robustness on unseen classes. Our method is applicable to any proxy-based losses, incl

Stats

• citations27
• S2 citations—
• github stars0
• HF likes0
• heat score10.85
• arxiv keygu2021proxy

Related papers

• Symmetrical Synthesis For Deep Metric Learning (2020)8.60
• Hierarchical Proxy-based Loss For Deep Metric Learning (2021)10.85
• Learning To Generate Novel Classes For Deep Metric Learning (2022)0.00
• Informative Sample-aware Proxy For Deep Metric Learning (2022)0.00
• Robust Calibrate Proxy Loss For Deep Metric Learning (2023)0.00
• Deep Metric Learning With Soft Orthogonal Proxies (2023)0.00
• Adaptive Additive Classification-based Loss For Deep Metric Learning (2020)0.00
• Towards Improved Proxy-based Deep Metric Learning Via Data-augmented Domain Adaptation (2024)6.34