Knowledge graph embedding for data mining vs. knowledge graph embedding for link prediction – two sides of the same coin?

Knowledge Graph Embeddings, i.e., projections of entities and relations to lower dimensional spaces, have been proposed for two purposes: (1) providing an encoding data mining tasks, (2) predicting links in a knowledge graph. Both lines research pursued rather isolation from each other so far, with their own benchmarks evaluation methodologies. In this paper, we argue that both tasks are actually related, show the first family approaches can also be used second task vice versa. series experiments, provide comparison families on which, best our knowledge, has not done far. Furthermore, discuss differences similarity functions evoked by different embedding approaches.