mlsanityData to 7 Oct 2026

MHFC: Multi-Head Feature Collaboration for Few-Shot Learning

arXiv 2109.07785 ↗First posted 16 Sep 2021cs.CVcited by 6

Not on the map. Why?

Abstract (arXiv)

Few-shot learning (FSL) aims to address the data-scarce problem. A standard FSL framework is composed of two components: (1) Pre-train. Employ the base data to generate a CNN-based feature extraction model (FEM). (2) Meta-test. Apply the trained FEM to acquire the novel data's features and recognize them. FSL relies heavily on the design of the FEM. However, various FEMs have distinct emphases. For example, several may focus more attention on the contour information, whereas others may lay particular emphasis on the texture information. The single-head feature is only a one-sided representation of the sample. Besides the negative influence of cross-domain (e.g., the trained FEM can not adapt to the novel class flawlessly), the distribution of novel data may have a certain degree of deviation compared with the ground truth distribution, which is dubbed as distribution-shift-problem (DSP). To address the DSP, we propose Multi-Head Feature Collaboration (MHFC) algorithm, which attempts to project the multi-head features (e.g., multiple features extracted from a variety of FEMs) to a unified space and fuse them to capture more discriminative information. Typically, first, we introduce a subspace learning method to transform the multi-head features to aligned low-dimensional representations. It corrects the DSP via learning the feature with more powerful discrimination and overcomes the problem of inconsistent measurement scales from different head features. Then, we design an attention block to update combination weights for each head feature automatically. It comprehensively considers the contribution of various perspectives and further improves the discrimination of features. We evaluate the proposed method on five benchmark datasets (including cross-domain experiments) and achieve significant improvements of 2.1%-7.8% compared with state-of-the-arts.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 37 of its 56 references that are in the map. First posted on arXiv 16 Sep 2021.

Builds on

37 of 56 references.

References · 37

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  3. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  4. 2017Prototypical Networks for Few-shot Learningcited by 3,001
  5. 2016Matching Networks for One Shot Learningcited by 2,246
  6. 2017Learning to Compare: Relation Network for Few-Shot Learningcited by 1,139
  7. 2018On First-Order Meta-Learning Algorithmscited by 1,003
  8. 2019A Closer Look at Few-shot Classificationcited by 679
  9. 2018Meta-Learning with Latent Embedding Optimizationcited by 476
  10. 2018Meta-Learning for Semi-Supervised Few-Shot Classificationcited by 456
  11. 2020Rethinking Few-Shot Image Classification: a Good Embedding Is All You Need?cited by 435
  12. 2018TADAM: Task dependent adaptive metric for improved few-shot learningcited by 408
25 more
  1. 2018Meta-learning with differentiable closed-form solverscited by 395
  2. 2019A Baseline for Few-Shot Image Classificationcited by 298
  3. 2018Learning to Propagate Labels: Transductive Propagation Network for Few-shot Learningcited by 208
  4. 2020Cross-Domain Few-Shot Classification via Learned Feature-Wise Transformationcited by 183
  5. 2019Charting the Right Manifold: Manifold Mixup for Few-shot Learningcited by 135
  6. 2020Negative Margin Matters: Understanding Margin in Few-shot Classificationcited by 129
  7. 2020Interventional Few-Shot Learningcited by 112
  8. 2020Unsupervised Person Re-identification via Multi-label Classificationcited by 107
  9. 2019Finding Task-Relevant Features for Few-Shot Learning by Category Traversalcited by 101
  10. 2019Adaptive Cross-Modal Few-Shot Learningcited by 95
  11. 2019TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learningcited by 90
  12. 2020Laplacian Regularized Few-Shot Learningcited by 90
  13. 2019Learning to Self-Train for Semi-Supervised Few-Shot Classificationcited by 83
  14. 2020Embedding Propagation: Smoother Manifold for Few-Shot Classificationcited by 79
  15. 2020DPGN: Distribution Propagation Graph Network for Few-shot Learningcited by 79
  16. 2019Diversity with Cooperation: Ensemble Methods for Few-Shot Classificationcited by 70
  17. 2019Transductive Episodic-Wise Adaptive Metric for Few-Shot Learningcited by 67
  18. 2020Instance Credibility Inference for Few-Shot Learningcited by 57
  19. 2021Exploring Complementary Strengths of Invariant and Equivariant Representations for Few-Shot Learningcited by 56
  20. 2020Empirical Bayes Transductive Meta-Learning with Synthetic Gradientscited by 48
  21. 2020Transductive Information Maximization For Few-Shot Learningcited by 44
  22. 2021Learning Dynamic Alignment via Meta-filter for Few-shot Learningcited by 41
  23. 2019TransMatch: A Transfer-Learning Scheme for Semi-Supervised Few-Shot Learningcited by 39
  24. 2019Label Embedded Dictionary Learning for Image Classificationcited by 8
  25. 2013Multiview Hessian Discriminative Sparse Coding for Image Annotationcited by 4

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 4; 2023: 1.

6 papers.

  1. 2022Memorizing Complementation Network for Few-Shot Class-Incremental Learningcited by 30
  2. 2022SphereFed: Hyperspherical Federated Learningcited by 19
  3. 2023Semantic-Aware Graph Matching Mechanism for Multi-Label Image Recognitioncited by 7
  4. 2021MDFM: Multi-Decision Fusing Model for Few-Shot Learningcited by 5
  5. 2022GCT: Graph Co-Training for Semi-Supervised Few-Shot Learningcited by 5
  6. 2022Enhancing Few-shot Image Classification with Cosine Transformercited by 4