Representation Power of Graph Convolutions : Neural Tangent Kernel Analysis

Mahalakshmi Sabanayagam, Pascal Esser, Debarghya Ghoshdastidar

Keywords: Graph Neural Networks, Neural Tangent Kernels, Node classification, Stochastic Block Model

TL;DR: Graph NTK shows that row normalized graph convolution preserves the underlying class structure, and skip connections retain the class structure at infinite depth.

Abstract: The fundamental principle of Graph Neural Networks (GNNs) is to exploit the structural information of the data by aggregating the neighboring nodes using a `graph convolution’. Therefore, understanding its influence on the network performance is crucial. Convolutions based on graph Laplacian have emerged as the dominant choice with the symmetric normalization of the adjacency matrix $A$, defined as $D^{-1/2}AD^{−1/2}$, being the most widely adopted one, where $D$ is the degree matrix. However, some empirical studies show that row normalization $D^{−1}A$ outperforms it in node classification. Despite the widespread use of GNNs, there is no rigorous theoretical study on the representation power of these convolution operators, that could explain this behavior. In this work, we analyze the influence of the graph convolutions theoretically using Graph Neural Tangent Kernel in a semi-supervised node classification setting. Under a Degree Corrected Stochastic Block Model, we prove that: (i) row normalization preserves the underlying class structure better than other graph convolutions; (ii) performance degrades with network depth due to over-smoothing, but the loss in class information is the slowest in row normalization; (iii) skip connections retain the class information even at infinite depth, thereby eliminating over-smoothing. We finally validate our theoretical findings on real datasets.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Theory (eg, control theory, learning theory, algorithmic game theory)

Community Implementations: [ 3 code implementations](https://www.catalyzex.com/paper/representation-power-of-graph-convolutions/code)