Learning Graph Embeddings from WordNet-based Similarity Measures [article]

Andrey Kutuzov, Mohammad Dorgham, Oleksiy Oliynyk, Chris Biemann, Alexander Panchenko
2019 arXiv   pre-print
We present path2vec, a new approach for learning graph embeddings that relies on structural measures of pairwise node similarities. The model learns representations for nodes in a dense space that approximate a given user-defined graph distance measure, such as e.g. the shortest path distance or distance measures that take information beyond the graph structure into account. Evaluation of the proposed model on semantic similarity and word sense disambiguation tasks, using various WordNet-based
more » ... imilarity measures, show that our approach yields competitive results, outperforming strong graph embedding baselines. The model is computationally efficient, being orders of magnitude faster than the direct computation of graph-based distances.
arXiv:1808.05611v4 fatcat:mrb36kdyabapzgnl4r3wtmcmt4