What graph neural networks cannot learn: depth vs width

This paper studies the expressive power of graph neural networks falling within the message-passing framework (GNNmp). Two results are presented. First, GNNmp are shown to be Turing universal under sufficient conditions on their depth, width, node attributes, and layer expressiveness. Second, it is discovered that GNNmp can lose a significant portion of their power when their depth and width is restricted. The proposed impossibility statements stem from a new technique that enables the repurposing of seminal results from distributed computing and leads to lower bounds for an array of decision, optimization, and estimation problems involving graphs. Strikingly, several of these problems are deemed impossible unless the product of a GNNmp's depth and width exceeds a polynomial of the graph size; this dependence remains significant even for tasks that appear simple or when considering approximation.

Presented at:
International Conference on Learning Representations, Addis Ababa, Ethiopia, April 26-30, 2020
Additional link:

Note: The status of this file is: Anyone

 Record created 2020-01-28, last modified 2020-10-29

Download fulltext

Rate this document:

Rate this document:
(Not yet reviewed)