<p>Techniques for representing graphs and preserving different graph features in low-dimensional embeddings have garnered much popularity and application over the years. The resulting embeddings can be applied to a wide variety of downstream machine learning tasks such as node classification and community detection. Despite the successes of Graph Neural Networks (GNN) in graph learning, this classical architecture is plagued with limitations including but not limited to issues of oversmoothing from message passing at increasing layer depth, dependence on scarce labeled data, long-range dependencies, and a large execution time resulting from long graph updates. By way of an alternative solution to the highlighted limitations, we propose an improved graph representation technique using a mapping bijection function with log smoothing to aggregate neighborhood properties and update node states in a message passing process, akin to that of a classical GNN. Similarly, to instantiate each node’s class value, we adopt the unsupervised affinity propagation algorithm; with these class values an important component of the bijection function. The resulting state for each node at every update level is a simple encoded integer that captures in itself the properties of the neighbor nodes. Because the final node states are high-dimensional, they are transformed into low-latent vector representation, while preserving rich and meaningful features of the input graph. Through experiments with real-life datasets, we show that our proposed model outperforms existing models on node classification and community detection tasks.</p>