Node Classification Tasks Research Articles

Graph Relearn Network: Reducing performance variance and improving prediction accuracy of graph neural networks

Recent studies show that the predictive performance of graph neural networks (GNNs) is inconsistent and varies across different experimental runs, even with identical parameters. The prediction variability limits GNNs’ applicability, and the underlying reasons remain unclear. We identified a key factor contributing to this issue: the oscillation of the predicted classes of some nodes during GNN training. To address this problem, we propose a novel framework, known as Graph Relearn Network (GRN), designed to reduce prediction variance by iteratively refining the predictions of unstable nodes. The GRN framework operates in two phases: pre-predict and relearn. During the pre-predict phase, a graph-dense encoder is trained to pre-predict the node categories. In the relearn phase, the model intensively focuses on the unstable nodes to optimize the predictions. Extensive experiments on ten graph datasets demonstrate that the GRN significantly enhances the performance stability of GNNs (with std. reduced by up to 75%), and achieves state-of-the-art performance in prediction accuracy (increased by up to 11.97%). By mitigating the instability introduced by unstable nodes, GRN enhances both the performance stability and prediction accuracy of GNNs in node classification tasks.

Sparse graphs-based dynamic attention networks

In previous research, the prevailing assumption was that Graph Neural Networks (GNNs) precisely depicted the interconnections among nodes within the graph's architecture. Nonetheless, real-world graph datasets are often rife with noise, elements that can disseminate through the network and ultimately affect the outcome of the downstream tasks. Facing the complex fabric of real-world graphs and the myriad potential disturbances, we introduce the Sparse Graph Dynamic Attention Networks (SDGAT) in this research. SDGAT employs the L0 regularization technique to achieve a sparse representation of the graph structure, which eliminates noise and generates a more concise sparse graph. Building upon this foundation, the model integrates a dynamic attention mechanism, allowing it to selectively focus on key nodes and edges, filter out irrelevant data, and simultaneously facilitate effective feature aggregation with important neighbors. To evaluate the performance of SDGAT, we conducted experiments on three citation datasets and compared its performance against commonly employed models. The outcomes indicate that SDGAT excels in node classification tasks, notably on the Cora dataset, with an accuracy rate of 85.29%, marking a roughly 3% enhancement over the majority of baseline models. The experimental findings provide evidence that SDGAT delivers effective performance on all three citation datasets, underscoring the efficacy of the dynamic attention network built upon a sparse graph.

FICOM: an effective and scalable active learning framework for GNNs on semi-supervised node classification

Active learning for graph neural networks (GNNs) aims to select B nodes to label for the best possible GNN performance. Carefully selected labeled nodes can help improve GNN performance and hence motivates a line of research works. Unfortunately, existing methods still provide inferior GNN performance or cannot scale to large networks.Motivated by these limitations, in this paper, we present FICOM, an effective and scalable GNN active learning framework. Firstly, we formulate the node selection as an optimization problem where we consider the importance of a node from (i) the importance of a node during the feature propagation with a connection to the personalized PageRank (PPR), and (ii) the diversity of a node brings in the embedding space generated by feature propagation. We show that the defined problem is submodular, and a greedy solution can provide a (1-1/e)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$(1-1/e)$$\\end{document}-approximate solution.However, a standard greedy solution requires getting the node with the maximum marginal gain of the objective score in each iteration, which incurs a prohibitive running cost and cannot scale to large datasets. As our main contribution, we present FICOM, an efficient and scalable solution that provides (1-1/e)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$(1-1/e)$$\\end{document}-approximation guarantee and scales to graphs with millions of nodes on a single machine. The main idea is that we adaptively maintain the lower- and upper-bound of the marginal gain for each node v. In each iteration, we can first derive a small subset of candidate nodes and then compute the exact score for this subset of candidate nodes so that we can find the node with the maximum marginal gain efficiently. Extensive experiments on six benchmark datasets using four GNNs, including GCN, SGC, APPNP, and GCNII, show that our FICOM consistently outperforms existing active learning approaches on semi-supervised node classification tasks using different GNNs. Moreover, our solution can finish within 5 h on a million-node graph.

High-frequency and low-frequency dual-channel graph attention network

Most existing graph convolution layers use learnable or fixed weights to sum up neighbor features to aggregate neighbor information. Since the attention values are always positive, these graph convolution layers perform as low-pass filters, which may result in their poor performance on heterophilic graphs. In this paper, two graph convolutional layers are proposed, HLGAT and NGAT. NGAT is a convolution network using only negative attention values, which only make the aggregation of high-frequency information of neighbor nodes. HLGAT makes the aggregation of low-frequency and high-frequency information by two channels, respectively, and fuses two outputs by using a learnable way. On node-classification task, both NGAT and HLGAT offer significant performance improvement compared to existing methods. The results clearly show that: (1) High-frequency information of neighborhoods plays a decisive role in heterophilic graphs. (2) The aggregation of low-frequency and high-frequency information of neighbor nodes can significantly improve the performance on heterophilic graphs.

MDGCL: Graph Contrastive Learning Framework with Multiple Graph Diffusion Methods

In recent years, some classical graph contrastive learning(GCL) frameworks have been proposed to address the problem of sparse labeling of graph data in the real world. However, in node classification tasks, there are two obvious problems with existing GCL frameworks: first, the stochastic augmentation methods they adopt lose a lot of semantic information; second, the local–local contrasting mode selected by most frameworks ignores the global semantic information of the original graph, which limits the node classification performance of these frameworks. To address the above problems, this paper proposes a novel graph contrastive learning framework, MDGCL, which introduces two graph diffusion methods, Markov and PPR, and a deterministic–stochastic data augmentation strategy while retaining the local–local contrasting mode. Specifically, before using the two stochastic augmentation methods (FeatureDrop and EdgeDrop), MDGCL first uses two deterministic augmentation methods (Markov diffusion and PPR diffusion) to perform data augmentation on the original graph to increase the semantic information, this step ensures subsequent stochastic augmentation methods do not lose too much semantic information. Meanwhile, the diffusion matrices carried by the augmented views contain global semantic information of the original graph, allowing the framework to utilize the global semantic information while retaining the local-local contrasting mode, which further enhances the node classification performance of the framework. We conduct extensive comparative experiments on multiple benchmark datasets, and the results show that MDGCL outperforms the representative baseline frameworks on node classification tasks. Among them, compared with COSTA, MDGCL’s node classification accuracy has been improved by 1.07% and 0.41% respectively on two representative datasets, Amazon-Photo and Coauthor-CS. In addition, we also conduct ablation experiments on two datasets, Cora and CiteSeer, to verify the effectiveness of each improvement work of our framework.

Global-local graph attention: unifying global and local attention for node classification

Abstract Graph Neural Networks (GNNs) are deep learning models specifically designed for analyzing graph-structured data, capturing complex relationships and structures to improve analysis and prediction. A common task in GNNs is node classification, where each node in the graph is assigned a predefined category. The Graph Attention Network (GAT) is a popular variant of GNNs known for its ability to capture complex dependencies by assigning importance weights to nodes during information aggregation. However, the GAT’s reliance on local attention mechanisms limits its effectiveness in capturing global information and long-range dependencies. To address this limitation, we propose a new attention mechanism called Global-Local Graph Attention (GLGA). Our mechanism enables the GAT to capture long-range dependencies and global graph structures while maintaining its ability to focus on local interactions. We evaluate our algorithm on three citation datasets (Cora, Citeseer, and Pubmed) using multiple metrics, demonstrating its superiority over other baseline models. The proposed GLGA mechanism has been proven to be an effective solution for improving node classification tasks.

XGBoost-Enhanced Graph Neural Networks: A New Architecture for Heterogeneous Tabular Data

Graph neural networks (GNNs) perform well in text analysis tasks. Their unique structure allows them to capture complex patterns and dependencies in text, making them ideal for processing natural language tasks. At the same time, XGBoost (version 1.6.2.) outperforms other machine learning methods on heterogeneous tabular data. However, traditional graph neural networks mainly study isomorphic and sparse data features. Therefore, when dealing with tabular data, traditional graph neural networks encounter challenges such as data structure mismatch, feature selection, and processing difficulties. To solve these problems, we propose a novel architecture, XGNN, which combines the advantages of XGBoost and GNNs to deal with heterogeneous features and graph structures. In this paper, we use GAT for our graph neural network model. We can train XGBoost and GNN end-to-end to fit and adjust the new tree in XGBoost based on the gradient information from the GNN. Extensive experiments on node prediction and node classification tasks demonstrate that the performance of our proposed new model is significantly improved for both prediction and classification tasks and performs particularly well on heterogeneous tabular data.

Heterophily-aware graph attention network

Graph Neural Networks (GNNs) have shown remarkable success in graph representation learning. Unfortunately, current weight assignment schemes in standard GNNs, such as the calculation based on node degrees or pair-wise representations, can hardly be effective in processing the networks with heterophily, in which the connected nodes usually possess different labels or features. Existing heterophilic GNNs tend to ignore the modeling of heterophily of each edge, which is also a vital part in tackling the heterophily problem. In this paper, we first propose a heterophily-aware attention scheme and reveal the benefits of modeling the edge heterophily, i.e., if a GNN assigns different weights to edges according to different heterophilic types, it can learn effective local attention patterns, enabling nodes to acquire appropriate information from distinct neighbors. Then, we propose a novel Heterophily-Aware Graph Attention Network (HA-GAT) by fully exploring and utilizing the local distribution as the underlying heterophily, to handle the networks with different homophily ratios. To demonstrate the effectiveness of the proposed HA-GAT, we analyze the proposed heterophily-aware attention scheme and local distribution exploration, by seeking an interpretation from their mechanism. Extensive results demonstrate that our HA-GAT achieves state-of-the-art performances on eight datasets with different homophily ratios in both the supervised and semi-supervised node classification tasks.

GNN-MgrPool: Enhanced graph neural networks with multi-granularity pooling for graph classification

Graph neural networks (GNNs) have gained sufficient attention and are applied to various domain tasks. At present, numerous pooling approaches are being proposed to aggregate node features and obtain node embeddings. However, current GNNs are black-box models that typically use a flat or single pooling step to aggregate nodes, which only considers the similarity between nodes within the cluster. These approaches ignore the influence of relationships within and between clusters in the learning process. To address this issue, we propose a novel multi-granular pooling method that aggregates nodes by simultaneously considering density and relationships among nodes and clusters. This method allows us to obtain multi-granular node-embedding clusters from GNN layers. The clusters in the current layer are built upon those in the previous layer, and these clusters change from fine to coarse as the number of clusters decreases during learning, which is achieved by using multi-granular pooling (MgrPool). Additionally, there is an inclusion relation between adjacent layers, and the node representation of each layer is established through the ratio of node distance within clusters to that between clusters. Finally, we conducted several experiments on node and graph classification tasks by combining GNN models with this pooling approach. The results demonstrate that our GNN-MgrPool model outperforms similar state-of-the-art algorithms and largely improves the interpretability of the learning process.

Multi-view Heterogeneous Graph Neural Networks for Node Classification

Recently, with graph neural networks (GNNs) becoming a powerful technique for graph representation, many excellent GNN-based models have been proposed for processing heterogeneous graphs, which are termed Heterogeneous graph neural networks (HGNNs). However, existing HGNNs tend to aggregate information from either direct neighbors or those connected by short metapaths, thereby neglecting the higher-order information and global feature similarity information in heterogeneous graphs. In this paper, we propose a Multi-View Heterogeneous graph neural network (MV-HGNN) to aggregate these information. Firstly, two auxiliary views, specifically a global feature similarity view and a graph diffusion view, are generated from the original heterogeneous graph. Secondly, MV-HGNN performs two message-passing strategies to get the representation of different views. Subsequently, a transformer-based aggregator is used to get the semantic information. Subsequently, the representations of the three views are fused into a final composite representation. We evaluate our method on the node classification task over three commonly used heterogeneous graph datasets, and the results demonstrate that our proposed MV-HGNN significantly outperforms state-of-the-art baselines.

Global-local graph neural networks for node-classification

The task of graph node classification is often approached by utilizing a local Graph Neural Network (GNN), that learns only local information from the node input features and their adjacency. In this paper, we propose to improve the performance of node classification GNNs by utilizing both global and local information, specifically by learning label- and node- features. We therefore call our method Global-Local-GNN (GLGNN). To learn proper label features, for each label, we maximize the similarity between its features and nodes features that belong to the label, while maximizing the distance between nodes that do not belong to the considered label. We then use the learnt label features to predict the node classification map. We demonstrate our GLGNN using three different GNN backbones, and show that our approach improves baseline performance, revealing the importance of global information utilization for node classification.

Research on Recognition Method of Social Robot Based on T-A-GCNIIT in the Metaverse

Social robots are used in intelligent customer service, intelligent chat, intelligent shopping guides, and more because of emotion recognition studies in cognitive psychology. However, determining the user's purpose quickly and precisely has proved challenging. Domestic researchers proposed the A-GCNII model to address missing feature information; however, it needs a lot of math. This research offers a social robot recognition approach using the T-A-GCNIIT model and cognitive psychology to optimize computing complexity and performance. The T-A-GCNIIT algorithm processes social network data, and the Viola–Jones algorithm improves social robot intelligence to represent social robots in the meta-universe. The model performs well in node classification, link prediction, community discovery, and other tasks, with enhanced accuracy, recall, F1 score value, and other metrics. The model can also better comprehend the user's emotional state using cognitive psychology to better recognize their purpose and propose a fresh notion for enhancing social robots' cognitive psychology.

Graph Aggregating-Repelling Network: Do Not Trust All Neighbors in Heterophilic Graphs

Graph neural networks (GNNs) have demonstrated exceptional performance in processing various types of graph data, such as citation networks and social networks, etc. Although many of these GNNs prove their superiority in handling homophilic graphs, they often overlook the other kind of widespread heterophilic graphs, in which adjacent nodes tend to have different classes or dissimilar features. Recent methods attempt to address heterophilic graphs from the graph spatial domain, which try to aggregate more similar nodes or prevent dissimilar nodes with negative weights. However, they may neglect valuable heterophilic information or extract heterophilic information ineffectively, which could cause poor performance of downstream tasks on heterophilic graphs, including node classification and graph classification, etc. Hence, a novel framework named GARN is proposed to effectively extract both homophilic and heterophilic information. First, we analyze the shortcomings of most GNNs in tackling heterophilic graphs from the perspective of graph spectral and spatial theory. Then, motivated by these analyses, a Graph Aggregating-Repelling Convolution (GARC) mechanism is designed with the objective of fusing both low-pass and high-pass graph filters. Technically, it learns positive attention weights as a low-pass filter to aggregate similar adjacent nodes, and learns negative attention weights as a high-pass filter to repel dissimilar adjacent nodes. A learnable integration weight is used to adaptively fuse these two filters and balance the proportion of the learned positive and negative weights, which could control our GARC to evolve into different types of graph filters and prevent it from over-relying on high intra-class similarity. Finally, a framework named GARN is established by simply stacking several layers of GARC to evaluate its graph representation learning ability on both the node classification and image-converted graph classification tasks. Extensive experiments conducted on multiple homophilic and heterophilic graphs and complex real-world image-converted graphs indicate the effectiveness of our proposed framework and mechanism over several representative GNN baselines.

HealthGAT: Node Classifications in Electronic Health Records using Graph Attention Networks.

While electronic health records (EHRs) are widely used across various applications in healthcare, most applications use the EHRs in their raw (tabular) format. Relying on raw or simple data pre-processing can greatly limit the performance or even applicability of downstream tasks using EHRs. To address this challenge, we present HealthGAT, a novel graph attention network framework that utilizes a hierarchical approach to generate embeddings from EHR, surpassing traditional graph-based methods. Our model iteratively refines the embeddings for medical codes, resulting in improved EHR data analysis. We also introduce customized EHR-centric auxiliary pre-training tasks to leverage the rich medical knowledge embedded within the data. This approach provides a comprehensive analysis of complex medical relationships and offers significant advancement over standard data representation techniques. HealthGAT has demonstrated its effectiveness in various healthcare scenarios through comprehensive evaluations against established methodologies. Specifically, our model shows outstanding performance in node classification and downstream tasks such as predicting readmissions and diagnosis classifications.

A Fully Test-time Training Framework for Semi-supervised Node Classification on Out-of-Distribution Graphs

Graph neural networks (GNNs) have shown great potential in representation learning for various graph tasks. However, the distribution shift between the training and test sets poses a challenge to the efficiency of GNNs. To address this challenge, HomoTTT proposes a fully test-time training framework for GNNs to enhance the model’s generalization capabilities for node classification tasks. Specifically, our proposed HomoTTT designs a homophily-based and parameter-free graph contrastive learning task with adaptive augmentation to guide the model’s adaptation during the test-time training, allowing the model to adapt for specific target data. In the inference stage, HomoTTT proposes to integrate the original GNN model and the adapted model after TTT using a homophily-based model selection method, which prevents potential performance degradation caused by unconstrained model adaptation. Extensive experimental results on six benchmark datasets demonstrate the effectiveness of our proposed framework. Additionally, the exploratory study further validates the rationality of the homophily-based graph contrastive learning task with adaptive augmentation and the homophily-based model selection designed in HomoTTT .

Predicting properties of nodes via community-aware features

This paper shows how information about the network’s community structure can be used to define node features with high predictive power for classification tasks. To do so, we define a family of community-aware node features and investigate their properties. Those features are designed to ensure that they can be efficiently computed even for large graphs. We show that community-aware node features contain information that cannot be completely recovered by classical node features or node embeddings (both classical and structural) and bring value in node classification tasks. This is verified for various classification tasks on synthetic and real-life networks.

Graph Contrastive Multi-view Learning: A Pre-training Framework for Graph Classification

Recent advancements in node and graph classification tasks can be attributed to the implementation of contrastive learning and similarity search. Despite considerable progress, these approaches present challenges. The integration of similarity search introduces an additional layer of complexity to the model. At the same time, applying contrastive learning to non-transferable domains or out-of-domain datasets results in less competitive outcomes. In this work, we propose maintaining domain specificity for these tasks, which has demonstrated the potential to improve performance by eliminating the need for additional similarity searches. We adopt a fraction of domain-specific datasets for pre-training purposes, generating augmented pairs that retain structural similarity to the original graph, thereby broadening the number of views. This strategy involves a comprehensive exploration of optimal augmentations to devise multi-view embeddings. An evaluation protocol, which focuses on error minimization, accuracy enhancement, and overfitting prevention, guides this process to learn inherent, transferable structural representations that span diverse datasets. We combine pre-trained embeddings and the source graph as a beneficial input, leveraging local and global graph information to enrich downstream tasks. Furthermore, to maximize the utility of negative samples in contrastive learning, we extend the training mechanism during the pre-training stage. Our method consistently outperforms comparative baseline approaches in comprehensive experiments conducted on benchmark graph datasets of varying sizes and characteristics, establishing new state-of-the-art results.

BP-MoE: Behavior Pattern-aware Mixture-of-Experts for Temporal Graph Representation Learning

Temporal graph representation learning aims to develop low-dimensional embeddings for nodes in a graph that can effectively capture their structural and temporal properties. Prior approaches primarily focus on devising sophisticated neural architectures to encode the historical interactions of nodes, enabling capturing the specific behavior patterns. However, such methods often overlook the fact that different nodes in a graph may exhibit distinct evolutionary preferences, resulting in a failure to adequately model the variations in the node evolution behavior. To address these issues, we propose a novel temporal graph learning framework called Behavior Pattern-aware Mixture-of-Experts (BP-MoE), which leverages multiple expert encoder networks with diverse architectures to comprehensively capture the behavior patterns for nodes. In detail, we first design a multi-perspective graph encoder that employs experts based on the gated recurrent units, graph attention units, and multilayer perceptron units to encode a node’s long-term memory, high-order neighborhood features, and short-term behavior preference, respectively. Subsequently, to achieve the personalized evolution preference learning for each individual node, we incorporate a behavior pattern-aware MoE mechanism to activate the pattern-specific experts for each node via a gating network. Finally, to optimize the training process of expert networks, we introduce two auxiliary losses that prevent the imbalanced training across experts. Extensive experiments conducted on multiple benchmark datasets verify the superiority of our approach over the state-of-the-art baselines. Specifically, BP-MoE consistently outperforms the baseline models for the link prediction and node classification tasks under various experimental settings.

Neighborhood Pattern Is Crucial for Graph Convolutional Networks Performing Node Classification.

Graph convolutional networks (GCNs) are widely believed to perform well in the graph node classification task, and homophily assumption plays a core rule in the design of previous GCNs. However, some recent advances on this area have pointed out that homophily may not be a necessity for GCNs. For deeper analysis of the critical factor affecting the performance of GCNs, we first propose a metric, namely, neighborhood class consistency (NCC), to quantitatively characterize the neighborhood patterns of graph datasets. Experiments surprisingly illustrate that our NCC is a better indicator, in comparison to the widely used homophily metrics, to estimate GCN performance for node classification. Furthermore, we propose a topology augmentation graph convolutional network (TA-GCN) framework under the guidance of the NCC metric, which simultaneously learns an augmented graph topology with higher NCC score and a node classifier based on the augmented graph topology. Extensive experiments on six public benchmarks clearly show that the proposed TA-GCN derives ideal topology with higher NCC score given the original graph topology and raw features, and it achieves excellent performance for semi-supervised node classification in comparison to several state-of-the-art (SOTA) baseline algorithms.

Designs of graph echo state networks for node classification

Among the Graph Neural Network (GNN) models that address the task of node classification, Graph Echo State Networks (GESN) have proved particularly effective in addressing the challenge of heterophily, i.e. the presence of a significant fraction of inter-class edges in the learning task graph. The effectiveness of GESN is paired with its efficiency, owing to the reservoir computing paradigm. While previous literature has analyzed the design of reservoirs for sequence ESN and GESN for graph-level tasks, the problem of providing effective designs of reservoirs for node-level GESN is so far largely unexplored. In this paper we analyze the impact of different reservoir designs on node classification accuracy and on the quality of node embeddings computed by GESN, focusing both on dense and sparse reservoir layouts. As measures of embedding richness, we adopt both graph topology-dependent metrics previously employed in the analysis of embedding smoothing, and topology-independent metrics from the areas of information theory and numerical analysis. In particular, we propose the application of entropy measures for quantifying information in node embeddings.