McGcn: Learning Continuous Graph Dynamics for Multi-Channel Fusion
A novel framework for multi-channel data fusion that integrates a multi-channel encoding module with continuous feature dynamics to initialize and evolve the latent node representations over static graph topologies, and effectively mitigates the performance degradation typically associated with deep graph architectures...