Theoretically Expressive and Edge-aware Graph Learning
This work addresses a foundational limitation in graph learning for researchers and practitioners, though it appears incremental as it builds on recent advancements.
The paper tackles the problem of limited expressiveness in Graph Neural Networks by proposing a new model that is theoretically proven to be more general than existing ones like Graph Isomorphism Network and Gated Graph Neural Network, as it can handle arbitrary edge values and allow unchanged node information flow.
We propose a new Graph Neural Network that combines recent advancements in the field. We give theoretical contributions by proving that the model is strictly more general than the Graph Isomorphism Network and the Gated Graph Neural Network, as it can approximate the same functions and deal with arbitrary edge values. Then, we show how a single node information can flow through the graph unchanged.