Understanding Sparse Neural Networks from their Topology via Multipartite Graph Representations

arXiv:2305.16886v2 Announce Type: replace
Abstract: Pruning-at-Initialization (PaI) algorithms provide Sparse Neural Networks (SNNs) which are computationally more efficient than their dense counterparts, and try to avoid performance degradation. While much emphasis has been directed towards \emph{how} to prune, we still do not know \emph{what topological metrics} of the SNNs characterize \emph{good performance}. From prior work, we have layer-wise topological metrics by which SNN performance can be predicted: the Ramanujan-based metrics. To exploit these metrics, proper ways to represent network layers …

arxiv cs.ai cs.lg graph networks neural networks topology type understanding via