Projects per year
Abstract
Graph neural networks (GNNs) are now the mainstream method for mining graphstructured data and learning lowdimensional node and graphlevel embeddings to serve downstream tasks. However, limited by the bottleneck of interpretability that deep neural networks present, existing GNNs have ignored the issue of estimating the appropriate number of dimensions for the embeddings. Hence, we propose a novel framework called Minimum Graph Entropy principleguided Dimension Estimation, i.e. MGEDE, that learns the appropriate embedding dimensions for both node and graph representations. In terms of nodelevel estimation, a minimum entropy function that counts both structure and attribute entropy, appraises the appropriate number of dimensions. In terms of graphlevel estimation, each graph is assigned a customized embedding dimension from a candidate set based on the number of dimensions estimated for the nodelevel embeddings. Comprehensive experiments with node and graph classification tasks and nine benchmark datasets verify the effectiveness and generalizability of MGEDE.
Original language  English 

Title of host publication  Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining (WSDM '23) 
Place of Publication  New York 
Publisher  Association for Computing Machinery (ACM) 
Pages  114122 
Number of pages  9 
Volume  1 
ISBN (Electronic)  9781450394079 
DOIs  
Publication status  Published  2023 
Event  16th ACM International Conference on Web Search and Data Mining, WSDM 2023  Singapore, Singapore Duration: 27 Feb 2023 → 3 Mar 2023 
Conference
Conference  16th ACM International Conference on Web Search and Data Mining, WSDM 2023 

Country/Territory  Singapore 
City  Singapore 
Period  27/02/23 → 3/03/23 
Keywords
 Dimension estimation
 graph neural network
 graph entropy
 node embedding
 graph embe
Fingerprint
Dive into the research topics of 'Minimum entropy principle guided graph neural networks'. Together they form a unique fingerprint.Projects
 1 Active

DP230100899: New Graph Mining Technologies to Enable Timely Exploration of Social Events
1/01/23 → 31/12/25
Project: Research