The paper claims a k-layer message-passing network is equivalent to a single layer on the k-th power of the adjacency matrix, with deep-network failures on sparse graphs blamed on gradients.
Understanding the Representation Power of Graph Neural Networks in Learning Graph Topology
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Demystifying MPNNs: Message Passing as Merely Efficient Matrix Multiplication
The paper claims a k-layer message-passing network is equivalent to a single layer on the k-th power of the adjacency matrix, with deep-network failures on sparse graphs blamed on gradients.