🤖 AI Summary
This work addresses the lack of interpretability and structural awareness in knowledge graph node embeddings by proposing a plug-and-play, general-purpose graph node representation method. The approach jointly models local proximity and long-range structural correlations through a multi-source subspace comprising hop-based topology, label overlap, Markov transition probabilities, and Recursive Spectral Bisection (RSB) clustering indices. It introduces, for the first time, a ground-truth-guided loss function that jointly estimates Jaccard similarity and label overlap. A multivariate stochastic gradient descent framework is designed to jointly optimize subspace weights. Experiments on multiple benchmark datasets demonstrate significant improvements in node similarity retrieval accuracy. The resulting embeddings are directly applicable to downstream tasks without fine-tuning, and each subspace’s contribution is both interpretable and ablatable.
📝 Abstract
This paper discusses how to generate general graph node embeddings from knowledge graph representations. The embedded space is composed of a number of sub-features to mimic both local affinity and remote structural relevance. These sub-feature dimensions are defined by several indicators that we speculate to catch nodal similarities, such as hop-based topological patterns, the number of overlapping labels, the transitional probabilities (markov-chain probabilities), and the cluster indices computed by our recursive spectral bisection (RSB) algorithm. These measures are flattened over the one dimensional vector space into their respective sub-component ranges such that the entire set of vector similarity functions could be used for finding similar nodes. The error is defined by the sum of pairwise square differences across a randomly selected sample of graph nodes between the assumed embeddings and the ground truth estimates as our novel loss function. The ground truth is estimated to be a combination of pairwise Jaccard similarity and the number of overlapping labels. Finally, we demonstrate a multi-variate stochastic gradient descent (SGD) algorithm to compute the weighing factors among sub-vector spaces to minimize the average error using a random sampling logic.