Scalable Graph Learning for Enterprise Data

Explore the cutting-edge world of Relational Deep Learning (RDL) in this 35-minute invited speaker talk at KDD2024. Dive into Kumo AI's innovative approach to machine learning on relational databases, presented by Hema Raghavan. Learn how RDL views relational databases as temporal, heterogeneous graphs, enabling automatic learning across interconnected tables without manual feature engineering. Discover the power of predictive SQL (pSQL) and its applications in various enterprise scenarios. Gain insights into scaling graph engines to handle billions of entities, ensuring security, privacy, and explainability in business contexts. Explore open-source initiatives like PyG.org and RelBench, which are advancing research in graph learning. Understand how this technology can revolutionize machine learning tasks in enterprises, from recommender systems to fraud detection.