SimCalib: Graph Neural Network Calibration Based on Similarity between Nodes.
Boshi TangZhiyong WuXixin WuQiaochu HuangJun ChenShun LeiHelen MengPublished in: AAAI (2024)
Keyphrases
- neural network
- graph structure
- directed graph
- weighted graph
- adjacency matrix
- undirected graph
- graph structures
- spanning tree
- similarity measure
- nodes of a graph
- fully connected
- edge weights
- connected graphs
- random walk
- camera calibration
- similarity graph
- root node
- graph theory
- planar graphs
- graph representation
- strongly connected
- proximity measures
- small world
- connected components
- inexact graph matching
- artificial neural networks
- path length
- directed acyclic graph
- neural network model
- average degree
- betweenness centrality
- dominating set
- shortest path
- normalized cut
- back propagation
- graph model
- citation graph
- finding the shortest path
- graphical models
- attributed graphs
- self organizing maps
- euclidean distance
- random graphs
- graph theoretic
- distance function
- similarity metric
- genetic algorithm
- directed edges
- similarity function
- overlapping communities
- graph matching
- edit distance
- structured data
- focal length
- graph databases