Transformation of Node to Knowledge Graph Embeddings for Faster Link\n Prediction in Social Networks

Recent advances in neural networks have solved common graph problems such as\nlink prediction, node classification, node clustering, node recommendation by\ndeveloping embeddings of entities and relations into vector spaces. Graph\nembeddings encode the structural information present in a graph. The encoded\nembeddings then can be used to predict the missing links in a graph. However,\nobtaining the optimal embeddings for a graph can be a computationally\nchallenging task specially in an embedded system. Two techniques which we focus\non in this work are 1) node embeddings from random walk based methods and 2)\nknowledge graph embeddings. Random walk based embeddings are computationally\ninexpensive to obtain but are sub-optimal whereas knowledge graph embeddings\nperform better but are computationally expensive. In this work, we investigate\na transformation model which converts node embeddings obtained from random walk\nbased methods to embeddings obtained from knowledge graph methods directly\nwithout an increase in the computational cost. Extensive experimentation shows\nthat the proposed transformation model can be used for solving link prediction\nin real-time.\n