Rethinking Preventing Class-collapsing In Metric Learning With Margin-based Losses

Abstract

Metric learning seeks perceptual embeddings where visually similar instances are close and dissimilar instances are apart, but learned representations can be sub-optimal when the distribution of intra-class samples is diverse and distinct sub-clusters are present. Although theoretically with optimal assumptions, margin-based losses such as the triplet loss and margin loss have a diverse family of solutions. We theoretically prove and empirically show that under reasonable noise assumptions, margin-based losses tend to project all samples of a class with various modes onto a single point in the embedding space, resulting in a class collapse that usually renders the space ill-sorted for classification or retrieval. To address this problem, we propose a simple modification to the embedding losses such that each sample selects its nearest same-class counterpart in a batch as the positive element in the tuple. This allows for the presence of multiple sub-clusters within each class. The adap

Stats

• citations16
• S2 citations—
• github stars0
• HF likes0
• heat score9.23
• arxiv keylevi2020rethinking

Related papers

• Adaptive Additive Classification-based Loss For Deep Metric Learning (2020)0.00
• Variance & Greediness: A Comparative Study Of Metric-learning Losses (2026)0.00
• Constellation Loss: Improving The Efficiency Of Deep Metric Learning Loss Functions For Optimal Embedding (2019)0.00
• Anti-collapse Loss For Deep Metric Learning Based On Coding Rate Metric (2024)0.00
• Susana Distancia Is All You Need: Enforcing Class Separability In Metric Learning Via Two Novel Distance-based Loss Functions For Few-shot Image Classification (2023)0.00
• The Group Loss For Deep Metric Learning (2019)11.39
• Shadow Loss: Memory-linear Deep Metric Learning For Efficient Training (2023)0.00
• DAAL: Density-aware Adaptive Line Margin Loss For Multi-modal Deep Metric Learning (2024)0.00