# Comparing Contrastive and Triplet Loss: Variance Analysis and Optimization Behavior

arXiv: https://arxiv.org/abs/2510.02161 · HTML: https://mlsanity.com/paper/2510.02161/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 9 references that are in the map.
- First posted on arXiv 2 Oct 2025.

## Builds on (most cited first, 8 of 8)

- [FaceNet: A Unified Embedding for Face Recognition and Clustering](https://mlsanity.com/paper/1503.03832/) (2015)
- [In Defense of the Triplet Loss for Person Re-Identification](https://mlsanity.com/paper/1703.07737/) (2017)
- [Understanding Dimensional Collapse in Contrastive Self-supervised Learning](https://mlsanity.com/paper/2110.09348/) (2021)
- [Multi-Similarity Loss with General Pair Weighting for Deep Metric Learning](https://mlsanity.com/paper/1904.06627/) (2019)
- [A Metric Learning Reality Check](https://mlsanity.com/paper/2003.08505/) (2020)
- [Sampling Matters in Deep Embedding Learning](https://mlsanity.com/paper/1706.07567/) (2017)
- [Feature Learning based Deep Supervised Hashing with Pairwise Labels](https://mlsanity.com/paper/1511.03855/) (2015)
- [Fisher Discriminant Triplet and Contrastive Losses for Training Siamese Networks](https://mlsanity.com/paper/2004.04674/) (2020)

## Built on by (most cited first, 1 of 1)

- [Taxonomy of the Retrieval System Framework: Pitfalls and Paradigms](https://mlsanity.com/paper/2601.20131/) (2026)

## Abstract (arXiv)

Contrastive loss and triplet loss are widely used objectives in deep metric learning, yet their effects on representation quality remain insufficiently understood. We present a theoretical and empirical comparison of these losses, focusing on intra- and inter-class variance and optimization behavior (e.g., greedy updates). Through task-specific experiments with consistent settings on synthetic data and real datasets-MNIST, CIFAR-10-it is shown that triplet loss preserves greater variance within and across classes, supporting finer-grained distinctions in the learned representations. In contrast, contrastive loss tends to compact intra-class embeddings, which may obscure subtle semantic differences. To better understand their optimization dynamics, By examining loss-decay rate, active ratio, and gradient norm, we find that contrastive loss drives many small updates early on, while triplet loss produces fewer but stronger updates that sustain learning on hard examples. Finally, across both classification and retrieval tasks on MNIST, CIFAR-10, CUB-200, and CARS196 datasets, our results consistently show that triplet loss yields superior performance, which suggests using triplet loss for detail retention and hard-sample focus, and contrastive loss for smoother, broad-based embedding refinement.