mlsanityData to 7 Oct 2026

Negative Margin Matters: Understanding Margin in Few-shot Classification

arXiv 2003.12060 ↗First posted 26 Mar 2020cs.CV cs.LG stat.MLcited by 129

Not on the map. Why?

Abstract (arXiv)

This paper introduces a negative margin loss to metric learning based few-shot learning methods. The negative margin loss significantly outperforms regular softmax loss, and achieves state-of-the-art accuracy on three standard few-shot classification benchmarks with few bells and whistles. These results are contrary to the common practice in the metric learning field, that the margin is zero or positive. To understand why the negative margin loss performs well for the few-shot classification, we analyze the discriminability of learned features w.r.t different margins for training and novel classes, both empirically and theoretically. We find that although negative margin reduces the feature discriminability for training classes, it may also avoid falsely mapping samples of the same novel class to multiple peaks or clusters, and thus benefit the discrimination of novel classes. Code is available at https://github.com/bl0/negative-margin.few-shot.

Key facts. Cited by 129 papers in the map (data through 7 Oct 2026). First posted on arXiv 26 Mar 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 48; 2022: 39; 2023: 23; 2024: 11; 2025: 4; 2026: 4.

129 papers.

Sort:

By area:

  1. 2021Free Lunch for Few-shot Learning: Distribution Calibrationcited by 165
  2. 2021Open-World Semi-Supervised Learningcited by 165
  3. 2022A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunitiescited by 162
  4. 2022Forward Compatible Few-Shot Class-Incremental Learningcited by 125
  5. 2021Simpler is Better: Few-shot Semantic Segmentation with Classifier Weight Transformercited by 116
  6. 2023Neural Collapse Inspired Feature-Classifier Alignment for Few-Shot Class Incremental Learningcited by 112
  7. 2021Beyond Max-Margin: Class Margin Equilibrium for Few-shot Object Detectioncited by 80
  8. 2021Relational Embedding for Few-Shot Classificationcited by 74
  9. 2021Rectifying the Shortcut Learning of Background for Few-Shot Learningcited by 51
  10. 2021Boosting the Generalization Capability in Cross-Domain Few-shot Learning via Noise-enhanced Supervised Autoencodercited by 51
  11. 2022Joint Distribution Matters: Deep Brownian Distance Covariance for Few-Shot Classificationcited by 50
  12. 2022Margin-Based Few-Shot Class-Incremental Learning with Class-Level Overfitting Mitigationcited by 45
38 more
  1. 2021Orthogonal Projection Losscited by 44
  2. 2023Few-Shot Class-Incremental Learning via Training-Free Prototype Calibrationcited by 44
  3. 2021Learning Dynamic Alignment via Meta-filter for Few-shot Learningcited by 41
  4. 2022Generating Representative Samples for Few-Shot Classificationcited by 39
  5. 2021Revisiting the Transferability of Supervised Pretraining: an MLP Perspectivecited by 36
  6. 2022Adversarial Feature Augmentation for Cross-domain Few-shot Classificationcited by 36
  7. 2021PML: Progressive Margin Loss for Long-tailed Age Classificationcited by 32
  8. 2022Boosting Few-shot Fine-grained Recognition with Background Suppression and Foreground Alignmentcited by 29
  9. 2021Few-shot Action Recognition with Prototype-centered Attentive Learningcited by 28
  10. 2021Binocular Mutual Learning for Improving Few-shot Classificationcited by 26
  11. 2022Integrative Few-Shot Learning for Classification and Segmentationcited by 25
  12. 2022Matching Feature Sets for Few-Shot Image Classificationcited by 25
  13. 2022HyperShot: Few-Shot Learning by Kernel HyperNetworkscited by 24
  14. 2023Learngene: Inheriting Condensed Knowledge from the Ancestry Model to Descendant Modelscited by 24
  15. 2023Distilling Self-Supervised Vision Transformers for Weakly-Supervised Few-Shot Classification & Segmentationcited by 24
  16. 2021Deep Metric Learning for Few-Shot Image Classification: A Review of Recent Developmentscited by 23
  17. 2021Anti-aliasing Semantic Reconstruction for Few-Shot Semantic Segmentationcited by 23
  18. 2022Few-shot Learning with Noisy Labelscited by 23
  19. 2021How to Train Your MAML to Excel in Few-Shot Classificationcited by 22
  20. 2022Self-Promoted Supervision for Few-Shot Transformercited by 22
  21. 2022WaveGAN: Frequency-aware GAN for High-Fidelity Few-shot Image Generationcited by 22
  22. 2021LibFewShot: A Comprehensive Library for Few-shot Learningcited by 21
  23. 2022Task Discrepancy Maximization for Fine-grained Few-Shot Classificationcited by 21
  24. 2023DiGeo: Discriminative Geometry-Aware Learning for Generalized Few-Shot Object Detectioncited by 21
  25. 2022Channel Importance Matters in Few-Shot Image Classificationcited by 20
  26. 2022Alleviating the Sample Selection Bias in Few-shot Learning by Removing Projection to the Centroidcited by 20
  27. 2021Meta Navigator: Search for a Good Adaptation Policy for Few-shot Learningcited by 19
  28. 2022Self-Supervision Can Be a Good Few-Shot Learnercited by 19
  29. 2023Multimodal Parameter-Efficient Few-Shot Class Incremental Learningcited by 19
  30. 2023Does Few-shot Learning Suffer from Backdoor Attacks?cited by 19
  31. 2021Reinforced Attention for Few-Shot Learning and Beyondcited by 18
  32. 2021Learning to Affiliate: Mutual Centralized Learning for Few-shot Classificationcited by 18
  33. 2021Few-Shot Learning with a Strong Teachercited by 16
  34. 2022MarginNCE: Robust Sound Localization with a Negative Margincited by 16
  35. 2021Contrastive Prototype Learning with Augmented Embeddings for Few-Shot Learningcited by 15
  36. 2021A Closer Look at Prototype Classifier for Few-shot Image Classificationcited by 15
  37. 2021SEGA: Semantic Guided Attention on Visual Prototype for Few-Shot Learningcited by 15
  38. 2021Adaptive Poincar'e Point to Set Distance for Few-Shot Classificationcited by 15

50 most cited of 129 All 129 (JSON)