Knowledge Graph Embedding for Link Prediction

Knowledge Graphs (KGs) have found many applications in industrial and academic settings, which turn, motivated considerable research efforts towards large-scale information extraction from a variety of sources. Despite such efforts, it is well known that even the largest KGs suffer incompleteness; Link Prediction (LP) techniques address this issue by identifying missing facts among entities already KG. Among recent LP techniques, those based on KG embeddings achieved very promising performance some benchmarks. fast-growing literature subject, insufficient attention has been paid to effect design choices methods. Moreover, standard practice area report accuracy aggregating over large number test are vastly more represented than others; allows methods exhibit good results just attending structural properties include entities, while ignoring remaining majority This analysis provides comprehensive comparison embedding-based methods, extending dimensions beyond what commonly available literature. We experimentally compare effectiveness efficiency 18 state-of-the-art consider rule-based baseline, detailed most popular benchmarks