Machine learning on structured/tabular data can be transformed the same way computer vision and NLP were, by having a neural network learn directly on raw multi-table relational data represented as a graph, instead of manually feature-engineering a single flattened table.
Jure explains the core idea behind relational deep learning: treat a database's tables and foreign-key links as a graph of entities and relationships, then apply graph neural networks directly to the raw data instead of hand-engineering features into a single table. ✦ AI generated
Jure Leskovec · The TWIML AI Podcast · 2026-05-21 · original ↗
starts at this moment · 14:16
“You talk a little bit about how the these two areas of of research and practice relate to one another.”
So our question was how could just how could I just learn directly with a neural network over this multitabular data? Um and the answer is you know kind of surprisingly simple is to say just think of the database think of these tables as a graph of relationships between the entities in the database.
verbatim transcript · starts at 14:16
14:16complexity. So our question was how could just how could I just learn directly with a neural network over this multitabular data? Um and the answer is you know kind of surprisingly simple is to say just think of the database think of these tables as a graph of relationships between the entities in the database. So this would mean in some in in my you know I I'm a graph person
14:43so I like to think in terms of graphs right so graphs are composed of vertices the nodes this would be pe my users would be my um would be my products would be my transactions and so on. So this would denow the nodes and then the connections are just saying this user ID was part of this transaction that was part of that product and now we have a
15:05path from a user to the transaction to the product and then you know another user or another transaction is another path in this very simplistic graph. Um and now that we have a graph, we can basically apply graph deep learning like graph neural networks. Um which is a way to generalize deep learning to graph structured data and just train over that to get to get an accurate prediction.
15:30And what happens is two things happen. The first thing that happens is you don't have to do manual feature engineering any right. So it's much faster. It requires much less effort to train these models. And the second thing that happens is your models are more accurate. And then you say why can my models be more accurate? And the answer is very similar to what happens in computer vision. Right? If you are
15:56saying I am a human, I know what a car is. So I will I will build perfect features that detect whether there is a car on the image or not. I know cars. I drive them. I'm such a car expert. I can build the best features for detecting cars. Nobody in the right mind claims that, right? But you know in in machine learning, data science prediction, people are still saying, you know, I'm
- ·Treat database tables as a graph of entities and relationships
- ·Apply GNNs directly to raw multi-table data
- ·No manual feature engineering or table flattening needed
- ·Parallel to how CV and NLP were transformed by end-to-end learning
- ·Foreign keys become edges between entity nodes
- ·Neural network learns directly on the relational graph
- ·Surprisingly simple: just think of tables as a graph
- ·Eliminates hand-crafted feature extraction pipelines