Graphs, Entities, and Step Mixture for Enriching Graph Representation

Graph neural networks have shown promising results on representing and analyzing diverse graph-structured data, such as social networks, traffic flow, drug discovery, and recommendation systems. Existing approaches for graph neural networks typically suffer from the oversmoothing issue that results...

Full description

Saved in:
Bibliographic Details
Published in:IEEE access 2021, Vol.9, p.144025-144034
Main Authors: Shin, Kyuyong, Shin, Wonyoung, Ha, Jung-Woo, Kwon, Sunyoung
Format: Article
Language:English
Subjects:
Adaptation models
Aggregates
aggregation
Analytical models
attention
Benchmark testing
Cognitive tasks
embedding
graph
Graph neural network
Graph neural networks
Graph representations
Graphical representations
Graphs
mixture
Neural networks
Nodes
oversmoothing
Random walk
Recommender systems
Regularization
representation
Robustness
Social networks
Source code
Structured data
Traffic flow
Training
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Graph neural networks have shown promising results on representing and analyzing diverse graph-structured data, such as social networks, traffic flow, drug discovery, and recommendation systems. Existing approaches for graph neural networks typically suffer from the oversmoothing issue that results in indistinguishable node representation, as recursive and simultaneous neighborhood aggregation deepens. Also, most methods focus on transductive scenarios that are limited to fixed graphs, which do not generalize properly to unseen graphs. To address these issues, we propose a novel graph neural network that considers both edge-based neighborhood relationships and node-based entity features with multiple steps, i.e. G raph E ntities with S tep M ixture via random walk (GESM). GESM employs a mixture of various steps through random walk to alleviate the oversmoothing problem, an attention mechanism to dynamically reflect interrelations depending on node information, and structure-based regularization to enhance embedding representation. With intensive experiments, we show that our proposed GESM achieves state-of-the-art or comparable performances on eight benchmark graph datasets in both transductive and inductive learning tasks. Furthermore, we empirically demonstrate the superiority of our method on the oversmoothing issue with rich graph representations. Our source code is available at https://github.com/ShinKyuY/GESM .
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2021.3121708