Graph Neural Networks (GNNs) are gaining tons of recognition in the machine learning community due to their potential for solving complex tasks in social networks, drug discovery, recommendation systems, and more.
Unlike traditional neural networks that operate on fixed-size, ordered data, GNNs can operate on interconnected data with varying size, patterns, and complexity, which are ubiquitous in many real-world applications. This allows GNNs to capture and learn the relationships between entities in a graph and perform tasks that require reasoning about network structure. This is particularly useful for dealing with complex systems that have interconnected components, and makes GNNs a considerable candidate for modeling many of these messy real-world problems.
In this blog series, I will help you understand more about GNNs, separate some of the reality from the hype, and learn how to practically apply GNNs and related Graph ML with coded examples. This requires some level-setting, so I’ve split this blog into five parts for easy reading and learning, which I’ll try to release each week over the next five weeks.
- Graph Machine Learning Overview
- [Coming Soon!] GNNs: What They Are and Why They Matter
- The Promises and Pitfalls of GNNs
- Discover What’s Possible: Neo4j Graph Data Science and GNNs
- Discover What’s Possible: Graph ML Ecosystem & Alternatives to GNNs
However, if you want a crash course in all of the above, register for my webinar by the same title, Demystifying Graph Neural Networks, available now!