Edge-Labeling Graph Neural Network for Few-Shot Learning

By Jongmin Kim, Taesup Kim, Sungwoong Kim, Chang-Dong Yoo. Computer Vision and Pattern Recognition, 2019

529 citations recorded by the source corpus.

In this paper, we propose a novel edge-labeling graph neural network (EGNN), which adapts a deep neural network on the edge-labeling graph, for few-shot learning. The previous graph neural network (GNN) approaches in few-shot learning have been based on the node-labeling framework, which implicitly models the intra-cluster similarity and the inter-cluster dissimilarity. In contrast, the proposed EGNN learns to predict the edge-labels rather than the node-labels on the graph that enables the evolution of an explicit clustering by iteratively updating the edge-labels with direct exploitation of both intra-cluster similarity and the inter-cluster dissimilarity. It is also well suited for performing on various numbers of classes without retraining, and can be easily extended to perform a transductive inference. The parameters of the EGNN are learned by episodic training with an edge-labeling loss to obtain a well-generalizable model for unseen low-data problem. On both of the supervised and semi-supervised few-shot image classification tasks with two benchmark datasets, the proposed EGNN significantly improves the performances over the existing GNNs.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/098b138f58e43338248e3bc35cb36adfed8008d1

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Chunhua Shen
2. Guosheng Lin
3. Chi Zhang
4. Zeynep Akata
5. Bernt Schiele
6. Han-Jia Ye
7. Yu-Wing Tai
8. Yujun Cai

Citations in the corpus

• Improved Few-Shot Visual Classification (2019, unknown): “Relation networks [33] (and recent GCNN variations [13, 29]) extended this by parameterizing and learning the classification metric using a Multi-Layer Per-ceptron (MLP).”
• Aligning Visual Prototypes with BERT Embeddings for Few-Shot Learning (2021, background): “In addition, some works have used Graph Convolutional Networks [18] to exploit the relationship among support and query examples [17, 37].”
• Few-Shot Object Detection With Attention-RPN and Multi-Relation Detector (2019, background): “In [43, 44] the authors introduced Graph Neural Network (GNN) to model relationship between different categories.”
• Variational Few-Shot Learning (2019, methodology): “On this basis, labelpropagation-related approaches [29, 22] are developed to explore intra-class or inter-class relationships in the classification task.”
• Few-Shot Segmentation Without Meta-Learning: A Good Transductive Inference Is All You Need? (2020, background): “Furthermore, in a recent line of works, the transductive setting has been investigated for few-shot classification [6, 2, 14, 16, 20, 24, 31, 44], and yielded performance improvements over inductive inference.”
• Relational Generalized Few-Shot Learning (2019, methodology): “Graph-based methods have been applied to FSL [13, 17] by representing image instances with graph nodes in an N-way K-shot classification setup.”
• DeepEMD: Differentiable Earth Mover's Distance for Few-Shot Learning (2020, background): “classification problem, such as works based on graph theories [66], [67], [68], [69], reinforcement learning [70], differentiable SVM [71], generative models [72], [73], [74], [75], [76], [77], [78], [79], [80], transductive learning [81], [82], [83], [84], [85], recurrent models [86], [87], self-su”
• An Ensemble of Epoch-Wise Empirical Bayes for Few-Shot Learning (2019, unknown): “EGNN [31] ResNet-12 64.”
• Leveraging the Feature Distribution in Transfer-based Few-Shot Learning (2020, methodology): “works [20, 18, 19, 13, 17, 9, 16, 10, 39], where we exploit unlabelled samples during the procedure as well as priors about their relative proportions.”
• DeepEMD: Few-Shot Image Classification With Differentiable Earth Mover’s Distance and Structured Classifiers (2020, unknown): “Apart from the two popular branches, many other promising methods are also proposed to handle the few-shot classification problem, such as works based on graph theories [13, 15, 21], reinforcement learning [6], differentiable”
• Hierarchical Graph Attention Network for Few-shot Visual-Semantic Learning (2021, methodology): “GNN [14] and EGNN [22] are two GNN-based few-shot classification models.”
• A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities (2022, background): “36% ✓ EGNN [194] CVPR’19 Metric GNN ✓ - 76.”
• Graph Information Aggregation Cross-Domain Few-Shot Learning for Hyperspectral Image Classification (2022, methodology): “This training procedure is iteratively carried out episode by episode until convergence [35].”