Proxynca++: Revisiting And Revitalizing Proxy Neighborhood Component Analysis

Abstract

We consider the problem of distance metric learning (DML), where the task is to learn an effective similarity measure between images. We revisit ProxyNCA and incorporate several enhancements. We find that low temperature scaling is a performance-critical component and explain why it works. Besides, we also discover that Global Max Pooling works better in general when compared to Global Average Pooling. Additionally, our proposed fast moving proxies also addresses small gradient issue of proxies, and this component synergizes well with low temperature scaling and Global Max Pooling. Our enhanced model, called ProxyNCA++, achieves a 22.9 percentage point average improvement of Recall@1 across four different zero-shot retrieval datasets compared to the original ProxyNCA algorithm. Furthermore, we achieve state-of-the-art results on the CUB200, Cars196, Sop, and InShop datasets, achieving Recall@1 scores of 72.2, 90.1, 81.4, and 90.9, respectively.

Stats

• citations90
• S2 citations—
• github stars0
• HF likes0
• heat score14.69
• arxiv keyteh2020proxynca

Related papers

• Informative Sample-aware Proxy For Deep Metric Learning (2022)0.00
• Towards Improved Proxy-based Deep Metric Learning Via Data-augmented Domain Adaptation (2024)6.34
• Hierarchical Proxy-based Loss For Deep Metric Learning (2021)10.85
• Robust Calibrate Proxy Loss For Deep Metric Learning (2023)0.00
• No Fuss Distance Metric Learning Using Proxies (2017)20.03
• Deep Metric Learning With Soft Orthogonal Proxies (2023)0.00
• Procsim: Proxy-based Confidence For Robust Similarity Learning (2023)0.00
• Contextually Affinitive Neighborhood Refinery For Deep Clustering (2023)2.26