Heterogeneous Graph Neural Network
A variant of graph neural networks in which the graph contains multiple types of nodes and edges, with different types having distinct feature spaces and semantic relations. It models heterogeneous systems more naturally.
Timeline (last 90 days)
The framework embeds impairment information into the edge features of a heterogeneous graph neural network, where nodes represent access points, users and sensing targets, and edges carry the CSI error level and clutter covariance.
A heterogeneous graph neural network is a variant of graph neural networks whose graph contains multiple types of nodes and edges, with different types of nodes and edges having different feature spaces and semantic relations.
All Facts (2)
The framework embeds impairment information into the edge features of a heterogeneous graph neural network, where nodes represent access points, users and sensing targets, and edges carry the CSI error level and clutter covariance.
50%UnverifiedA heterogeneous graph neural network is a variant of graph neural networks whose graph contains multiple types of nodes and edges, with different types of nodes and edges having different feature spaces and semantic relations.
50%