DEGNN: Dual Experts Graph Neural Network Handling Both Edge and Node Feature Noise.
Tai HasegawaSukwon YunXin LiuYin Jun PhuaTsuyoshi MurataPublished in: CoRR (2024)
Keyphrases
- neural network
- undirected graph
- edge weights
- weighted graph
- graph structure
- directed graph
- disjoint paths
- edge orientation
- edge information
- noise level
- spanning tree
- shortest path
- random walk
- densely connected
- strongly connected
- edge detection
- noisy data
- finding the shortest path
- graph representation
- image edges
- graph theory
- connected components
- bipartite graph
- image features
- artificial neural networks
- genetic algorithm
- vector representation
- preprocessing stage
- vertex set
- nodes of a graph
- graph model
- noise reduction
- signal to noise ratio
- self organizing maps
- missing data
- back propagation
- graphical models
- feature vectors
- directed acyclic graph
- approximation algorithms
- structured data
- overlapping communities
- color images
- train a neural network