CORTEXA
← Browse

Alexander Pfefferle

2 papers indexed

arxivcs.LGcs.AI2026-06-29

Beyond IID: How General Are Tabular Foundation Models, Really?

Lennart Purucker, Andrej Tschalzev, Nick Erickson, Gioia Blayer, David Holzmüller, Alan Arazi, et al.

Foundation models for predictive machine learning on tabular data have recently gained significant traction in academia and industry. Research communities across disciplines are increasingly evaluating tabular foundation models on diverse datasets and tasks. However, these task-…

View free PDFSource page
arxivcs.LG2026-06-28

Towards Evaluating Data Priors for Tabular Foundation Models

Zeynep Türkmen, Kürşat Kaya, Alexander Pfefferle, Frank Hutter

Data-generating priors are a central component of tabular foundation models because they define the task distribution used during pretraining. However, priors are rarely evaluated as independent components, making it difficult to understand how much they affect downstream model b…

View free PDFSource page