Knowledge Representation in Graphs using Convolutional Neural Networks
This work addresses visualization and completion challenges in biomedical KGs, but it is incremental as it builds on existing methods.
The paper tackles the problem of missing edges and difficult visualization in sparse Knowledge Graphs (KG) by applying a compositional model for graph completion and a CNN-based visualization tool, achieving performance comparable to structural models on a biomedical dataset.
Knowledge Graphs (KG) constitute a flexible representation of complex relationships between entities particularly useful for biomedical data. These KG, however, are very sparse with many missing edges (facts) and the visualisation of the mesh of interactions nontrivial. Here we apply a compositional model to embed nodes and relationships into a vectorised semantic space to perform graph completion. A visualisation tool based on Convolutional Neural Networks and Self-Organised Maps (SOM) is proposed to extract high-level insights from the KG. We apply this technique to a subset of CTD, containing interactions of compounds with human genes / proteins and show that the performance is comparable to the one obtained by structural models.