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

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.

[1]  Yanjiang Wang,et al.  Weighted Spatial Pyramid Matching Collaborative Representation for Remote-Sensing-Image Scene Classification , 2019, Remote. Sens..

[2]  Sung Whan Yoon,et al.  TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learning , 2019, ICML.

[3]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[4]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[5]  Yue Wang,et al.  Rethinking Few-Shot Image Classification: a Good Embedding Is All You Need? , 2020, ECCV.

[6]  Alexandre Drouin,et al.  Embedding Propagation: Smoother Manifold for Few-Shot Classification , 2020, ECCV.

[7]  Michael E. Tipping,et al.  Probabilistic Principal Component Analysis , 1999 .

[8]  Hanwang Zhang,et al.  Interventional Few-Shot Learning , 2020, NeurIPS.

[9]  Abhishek Sinha,et al.  Charting the Right Manifold: Manifold Mixup for Few-shot Learning , 2019, 2020 IEEE Winter Conference on Applications of Computer Vision (WACV).

[10]  Jiebo Luo,et al.  TransMatch: A Transfer-Learning Scheme for Semi-Supervised Few-Shot Learning , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[11]  Oriol Vinyals,et al.  Matching Networks for One Shot Learning , 2016, NIPS.

[12]  Luca Bertinetto,et al.  Meta-learning with differentiable closed-form solvers , 2018, ICLR.

[13]  Gunhee Kim,et al.  Model-Agnostic Boundary-Adversarial Sampling for Test-Time Generalization in Few-Shot Learning , 2020, ECCV.

[14]  Pedro H. O. Pinheiro,et al.  Adaptive Cross-Modal Few-Shot Learning , 2019, NeurIPS.

[15]  Pablo Piantanida,et al.  Transductive Information Maximization For Few-Shot Learning , 2020, ArXiv.

[16]  Yuan Yan Tang,et al.  Multiview Hessian discriminative sparse coding for image annotation , 2013, Comput. Vis. Image Underst..

[17]  Feiyue Huang,et al.  Learning Dynamic Alignment via Meta-filter for Few-shot Learning , 2021, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[18]  Joshua B. Tenenbaum,et al.  Meta-Learning for Semi-Supervised Few-Shot Classification , 2018, ICLR.

[19]  Zheng Zhang,et al.  Negative Margin Matters: Understanding Margin in Few-shot Classification , 2020, ECCV.

[20]  Stefano Soatto,et al.  A Baseline for Few-Shot Image Classification , 2019, ICLR.

[21]  Zhiwu Lu,et al.  MELR: Meta-Learning via Modeling Episode-Level Relationships for Few-Shot Learning , 2021, ICLR.

[22]  Eunho Yang,et al.  Learning to Propagate Labels: Transductive Propagation Network for Few-Shot Learning , 2018, ICLR.

[23]  Mikhail Belkin,et al.  Laplacian Eigenmaps and Spectral Techniques for Embedding and Clustering , 2001, NIPS.

[24]  Yu-Chiang Frank Wang,et al.  A Closer Look at Few-shot Classification , 2019, ICLR.

[25]  Fahad Shahbaz Khan,et al.  Exploring Complementary Strengths of Invariant and Equivariant Representations for Few-Shot Learning , 2021, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[26]  J. Shawe-Taylor,et al.  Multi-View Canonical Correlation Analysis , 2010 .

[27]  Tao Xiang,et al.  Learning to Compare: Relation Network for Few-Shot Learning , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[28]  Shiliang Zhang,et al.  Unsupervised Person Re-Identification via Multi-Label Classification , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[29]  Razvan Pascanu,et al.  Meta-Learning with Latent Embedding Optimization , 2018, ICLR.

[30]  Zhenhua Guo,et al.  Contextual Multi-Scale Feature Learning for Person Re-Identification , 2020, ACM Multimedia.

[31]  Richard S. Zemel,et al.  Prototypical Networks for Few-shot Learning , 2017, NIPS.

[32]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[33]  Alexandre Lacoste,et al.  TADAM: Task dependent adaptive metric for improved few-shot learning , 2018, NeurIPS.

[34]  Neil D. Lawrence,et al.  Empirical Bayes Transductive Meta-Learning with Synthetic Gradients , 2020, ICLR.

[35]  Geoffrey E. Hinton,et al.  Visualizing Data using t-SNE , 2008 .

[36]  Cordelia Schmid,et al.  Diversity With Cooperation: Ensemble Methods for Few-Shot Classification , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[37]  Ioannis Mitliagkas,et al.  Manifold Mixup: Better Representations by Interpolating Hidden States , 2018, ICML.

[38]  Jose Dolz,et al.  Laplacian Regularized Few-Shot Learning , 2020, ICML.

[39]  Yicong Zhou,et al.  Orthogonalization-Guided Feature Fusion Network for Multimodal 2D+3D Facial Expression Recognition , 2020, IEEE Transactions on Multimedia.

[40]  Weifeng Liu,et al.  Label Embedded Dictionary Learning for Image Classification , 2019, Neurocomputing.

[41]  Hung-Yu Tseng,et al.  Cross-Domain Few-Shot Classification via Learned Feature-Wise Transformation , 2020, ICLR.

[42]  Yonghong Tian,et al.  Transductive Episodic-Wise Adaptive Metric for Few-Shot Learning , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[43]  Dinggang Shen,et al.  Multi-Hypergraph Learning for Incomplete Multimodality Data , 2018, IEEE Journal of Biomedical and Health Informatics.

[44]  Yanwei Fu,et al.  Instance Credibility Inference for Few-Shot Learning , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[45]  Sergey Levine,et al.  Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks , 2017, ICML.

[46]  Bernt Schiele,et al.  Learning to Self-Train for Semi-Supervised Few-Shot Classification , 2019, NeurIPS.

[47]  S T Roweis,et al.  Nonlinear dimensionality reduction by locally linear embedding. , 2000, Science.

[48]  Yue Gao,et al.  Inductive Multi-Hypergraph Learning and Its Application on View-Based 3D Object Classification , 2018, IEEE Transactions on Image Processing.

[49]  Ling Yang,et al.  DPGN: Distribution Propagation Graph Network for Few-Shot Learning , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[50]  Julien Mairal,et al.  Selecting Relevant Features from a Multi-domain Representation for Few-Shot Classification , 2020, ECCV.

[51]  Xiaogang Wang,et al.  Finding Task-Relevant Features for Few-Shot Learning by Category Traversal , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[52]  Alex Krizhevsky,et al.  Learning Multiple Layers of Features from Tiny Images , 2009 .

[53]  Yi Yang,et al.  Image Classification by Cross-Media Active Learning With Privileged Information , 2016, IEEE Transactions on Multimedia.

[54]  Rengang Li,et al.  Dense-Scale Feature Learning in Person Re-identification , 2020, ACCV.