Sparse Graph Transformer With Contrastive Learning

Article Computer Science, Cybernetics

Multigraph Random Walk for Joint Learning of Multiview Clustering and Semisupervised Classification

Shiping Wang et al.

Summary: Recent research has proposed a multigraph random walk scheme for addressing multiview clustering and semisupervised classification tasks, which integrates random walk with multiview learning. The scheme recursively learns a globally stable probability distribution matrix from multiple views to obtain label indicators for clustering or semisupervised classification. An adaptive weight vector is also learned to incorporate the diversity and complementarity of multiview data. Comprehensive comparative experiments demonstrate the superiority of the proposed method in terms of both clustering and classification performance.

IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS (2022)

Add to Collection

Article Computer Science, Cybernetics

Smoothing Adversarial Training for GNN

Jinyin Chen et al.

Summary: This study introduces a smoothing adversarial training (SAT) method to improve the robustness of Graph Neural Networks (GNNs) by smoothing the gradients of the Graph Convolutional Network (GCN), reducing the amplitude of adversarial gradients for defense against global and target label node attacks. Comprehensive experiments demonstrate that SAT method exhibits state-of-the-art defensibility against various adversarial attacks on node classification and community detection.

IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS (2021)

Add to Collection

Article Computer Science, Artificial Intelligence

Incremental multi-view spectral clustering with sparse and connected graph learning

Hongwei Yin et al.

Summary: This paper proposes an efficient incremental multi-view spectral clustering method SCGL, which only stores one consensus similarity matrix to represent the structural information of all historical views, and integrates sparse graph learning and connected graph learning. Experimental results demonstrate that the method outperforms traditional methods in clustering accuracy.

NEURAL NETWORKS (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Graph InfoClust: Maximizing Coarse-Grain Mutual Information in Graphs

Costas Mavromatis et al.

Summary: This paper introduces a new method for node representation learning that leverages coarse-grain information for clustering and maximizing mutual information among nodes, showing superior performance in tasks such as node classification, link prediction, and node clustering compared to competing methods.

ADVANCES IN KNOWLEDGE DISCOVERY AND DATA MINING, PAKDD 2021, PT I (2021)

Add to Collection

Proceedings Paper Computer Science, Information Systems

Variational Graph Normalized AutoEncoders

Seong Jin Ahn et al.

Summary: Link prediction is a key problem for graph-structured data, with graph autoencoders and variational graph autoencoders proposed to learn graph embeddings. However, existing methods do not perform well in link predictions involving isolated nodes. A new VGNAE model has been introduced, utilizing L-2 normalization for better embeddings of isolated nodes.

PROCEEDINGS OF THE 30TH ACM INTERNATIONAL CONFERENCE ON INFORMATION & KNOWLEDGE MANAGEMENT, CIKM 2021 (2021)

Add to Collection

Article Computer Science, Information Systems