AutoML in the Age of Structured Foundation Models

In this 43-minute AutoML Seminar, Nick Erickson explores the evolving relationship between AutoML systems and structured foundation models. Learn how recent models like TabPFNv2 and Chronos have achieved breakthrough performance on tabular and time series data, often surpassing traditional methods on small to medium datasets while requiring significantly less training time. Rather than viewing these developments as a threat to AutoML, discover why automated machine learning frameworks may become even more crucial in this new foundation model-driven landscape. The presentation offers valuable insights for data scientists and ML practitioners interested in the future direction of automated machine learning systems.