Attentional multilabel learning over graphs: a message passing approach

DOI10.1007/S10994-019-05782-6MaRDI QIDQ2320538zbMATH OpenOpenAlexWikidataFDO

Authors Kien Do, Truyen Tran, Thin Nguyen, Svetha Venkatesh

Publication date 23 August 2019

Published in Machine Learning (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1804.00293

zbMATH Keywords

graph classificationgraph neural networksmessage passingmultilabel learning

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Learning and adaptive systems in artificial intelligence (68T05)

Abstract: We address a largely open problem of multilabel classification over graphs. Unlike traditional vector input, a graph has rich variable-size substructures which are related to the labels in some ways. We believe that uncovering these relations might hold the key to classification performance and explainability. We introduce GAML (Graph Attentional Multi-Label learning), a novel graph neural network that can handle this problem effectively. GAML regards labels as auxiliary nodes and models them in conjunction with the input graph. By applying message passing and attention mechanisms to both the label nodes and the input nodes iteratively, GAML can capture the relations between the labels and the input subgraphs at various resolution scales. Moreover, our model can take advantage of explicit label dependencies. It also scales linearly with the number of labels and graph size thanks to our proposed hierarchical attention. We evaluate GAML on an extensive set of experiments with both graph-structured inputs and classical unstructured inputs. The results show that GAML significantly outperforms other competing methods. Importantly, GAML enables intuitive visualizations for better understanding of the label-substructure relations and explanation of the model behaviors.

Recommendations

• MGAT: multi-view graph attention networks
• Locality preserving dense graph convolutional networks with graph context-aware node representations
• Multi-label learning with missing labels using mixed dependency graphs
• Semisupervised, multilabel, multi-instance learning for structured data
• Multilabel classification through random graph ensembles

Cites work

• scientific article; zbMATH DE number 6253944 (Why is no real title available?)
• MEKA: a multi-label/multi-target extension to WEKA
• ML-KNN: A lazy learning approach to multi-label learning
• Multilabel classification via calibrated label ranking
• Weisfeiler-Lehman graph kernels

Cited in

(7)

• Locality preserving dense graph convolutional networks with graph context-aware node representations
• MTGL-ADMET: a novel multi-task graph learning framework for ADMET prediction enhanced by status-theory and maximum flow
• Neural predictor-based automated graph classifier framework
• Relation-aware Graph Contrastive Learning
• Informative pseudo-labeling for graph neural networks with few labels
• Saliency-aware regularized graph neural network
• Mg-SubAgg: multi-granularity subgraph aggregation with topology for GNN

Describes a project that uses

Uses Software

• MEKA
• Adam
• node2vec
• MULAN
• ML-KNN
• CNN-RNN
• HCP
• HIPPIE
• DeepWalk
• GraphVAE
• graph2vec

This page was built for publication: Attentional multilabel learning over graphs: a message passing approach

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2320538)