Abstract:Node classification is an essential problem in graph learning. However, many models typically obtain unsatisfactory performance when applied to few-shot scenarios. Some studies have attempted to combine meta-learning with graph neural networks to solve few-shot node classification on graphs. Despite their promising performance, some limitations remain. First, they employ the node encoding mechanism of homophilic graphs to learn node embeddings, even in heterophilic graphs. Second, existing models based on meta-learning ignore the interference of randomness in the learning process. Third, they are trained using only limited labeled nodes within the specific task, without explicitly utilizing numerous unlabeled nodes. Finally, they treat almost all sampled tasks equally without customizing them for their uniqueness. To address these issues, we propose a novel framework for few-shot node classification called Meta-GPS++. Specifically, we first adopt an efficient method to learn discriminative node representations on homophilic and heterophilic graphs. Then, we leverage a prototype-based approach to initialize parameters and contrastive learning for regularizing the distribution of node embeddings. Moreover, we apply self-training to extract valuable information from unlabeled nodes. Additionally, we adopt S$^2$ (scaling & shifting) transformation to learn transferable knowledge from diverse tasks. The results on real-world datasets show the superiority of Meta-GPS++. Our code is available here.

What problem does this paper attempt to address?

The main problems that this paper attempts to solve are some key limitations of existing meta - learning methods in few - shot node classification tasks. Specifically, these problems include: 1. **Homogeneity Assumption**: Existing graph encoding mechanisms are usually based on the homogeneity assumption (that is, nodes with the same label are more likely to be connected), even in heterogeneous graphs. This leads to poor performance in heterogeneous graphs because these models will learn uninformative node representations. 2. **Randomness Interference**: Existing meta - learning - based models ignore the influence of randomness during the training process, which may cause the model to be difficult to converge to the optimal solution. For example, MAML (Model - Agnostic Meta - Learning) uses random and class - independent initialization parameters, which makes the model vulnerable to data noise in a few - shot environment. 3. **Under - utilization of Unlabeled Nodes**: Existing meta - learning models usually only use a small number of labeled nodes in each task for training, ignoring the rich self - supervised and task - related information contained in a large number of unlabeled nodes. This may lead to over - fitting of the model and affect its generalization ability. 4. **Equal Treatment of Tasks**: Many existing methods treat all tasks equally without considering the complexity and diversity between different tasks. For example, different types of amino acid sequences in protein networks may exhibit different structural patterns, which requires customized parameter adjustment for each task. To solve the above problems, the author proposes a new framework - Meta - GPS++ for few - shot node classification on graphs. Meta - GPS++ contains the following five key components: 1. **Graph Network Encoder**: An efficient method is designed to learn discriminative node representations on homogeneous and heterogeneous graphs. 2. **Prototype - Based Parameter Initialization**: A prototype - based method is introduced to initialize parameters to solve the problem of random initialization of model parameters. 3. **Contrastive Learning to Deal with Task Randomness**: Contrastive learning is used to regularize the learned node embeddings and mitigate the negative impact of task randomness. 4. **Self - Training to Regularize the Model**: Valuable information in unlabeled nodes is extracted through self - training to enhance the generalization ability of the model. 5. **S2 Transformation to Adapt to Different Tasks**: Two networks are used to generate scaling and offset vectors to dynamically adjust the prior parameters of each new task to adapt to the differences between tasks. Through these innovations, Meta - GPS++ has demonstrated superior performance on multiple real - world datasets, especially in few - shot node classification tasks.

Meta-GPS++: Enhancing Graph Meta-Learning with Contrastive Learning and Self-Training

Graph Contrastive Learning Meets Graph Meta Learning: A Unified Method for Few-shot Node Tasks

Homophily-Enhanced Self-Supervision for Graph Structure Learning: Insights and Directions.

Self-Supervised Node Representation Learning Via Node-to-Neighbourhood Alignment.

Meta-GNN: On Few-shot Node Classification in Graph Meta-learning

Node Classification on Graphs with Few-Shot Novel Labels Via Meta Transformed Network Embedding

Robust Graph Meta-learning for Weakly-supervised Few-shot Node Classification

Adaptive-Step Graph Meta-Learner for Few-Shot Graph Classification

Few-shot Graph Classification with Contrastive Loss and Meta-classifier

Task-Equivariant Graph Few-shot Learning

CPT: Competence-progressive Training Strategy for Few-shot Node Classification

Structural Attention Enhanced Continual Meta-Learning for Graph Edge Labeling Based Few-Shot Remote Sensing Scene Classification

Graph-based few-shot learning with transformed feature propagation and optimal class allocation

Multi-level self-adaptive prototypical networks for few-shot node classification on attributed networks

Meta-node: A Concise Approach to Effectively Learn Complex Relationships in Heterogeneous Graphs

Meta-Hybrid: Integrate Meta-Learning to Enhance Class Imbalance Graph Learning

SPContrastNet: A Self-Paced Contrastive Learning Model for Few-Shot Text Classification

Few-shot Heterogeneous Graph Learning via Cross-domain Knowledge Transfer

Few-Shot Graph Classification with Structural-Enhanced Contrastive Learning for Graph Data Copyright Protection

Structure-Enhanced Meta-Learning for Few-Shot Graph Classification

Graph Few-shot Learning with Attribute Matching