Understanding Graph Embeddings

In the last year, graph embeddings have become increasingly important in Enterprise Knowledge Graph (EKG) strategy. Graph embeddings will soon become the de facto way to quickly find similar items in large billion-vertex EKGs. And as we have discussed in our prior articles, real-time similarity calculations are critical to many areas such as recommendation, next best action, and cohort building. The goal of this article is to give you an intuitive feeling for what graph embeddings are and how they are used so you can decide if these are right for your EKG project. For those of you with a bit of data science background, we will also touch a bit on how they are calculated. For the most part, we will be using storytelling and metaphors to explain these concepts.