Connecting the Dots: Harnessing the Power of Graphs & ML by Ebru Cucen

The world is becoming more complex, and more so than ever informed decision making has become critical for managing the uncertainty this creates.

James Fowler notes that “Relationships are the strongest predictors of behaviour” in his book “Connected”. Graphs provide a very natural way to model complexity, including constantly changing and evolving relationships. Graphs (in the mathematical sense) are extremely good at encapsulating and being able to embody and convey crucial data (context) about how and why entities are related to one another. Such information can often be THE crucial piece of a puzzle needed to answer complex business questions, or indeed to serve as the secret ingredient to unlock, optimise and improve predictive ML models to help us do this.

Whilst ML is already helping us make better decisions, if predictive information like relationships can be even more deeply embedded and taken into account, the potential for better answers can be significantly enhanced.

This talk aims to provide tour of, and insights into how graphs can help improve upon the predictions and answers arrived at by using traditional ML algorithms and approaches alone. We show how you are able to combine the power of both graphs and ML (in a variety of different ways) to help you arrive at better answers compared to using standard ML approaches alone. This includes diving into some demos of key approaches to illustrate and explain the ideas along the way.