Trends in tabular foundation research
Based on 151 papers from ICML workshop "Foundation Models for Structured Data"
Mindful Modeler
Tabular foundation models, ML interpretability, and beyond by a statistician turned machine learner.
Live Last read · last published · next check
Latest posts
Time for a change
After 4 years of writing, I'm exploring what's next for me in ML / data science / AI (or whatever you want to call it)
TabFM minus the hype
top performance on TabArena; large model; slow inference; non-commercial license
Which ML models produce the best quantile estimates?
Results from ScoringBench
When trees still beat tabular foundation models
As you might have noticed, I’m rather optimistic about tabular foundation models.
What is TabPFN's Thinking mode?
When I read the TabPFN-3 Technical Report, the benchmark for “TabPFN-3-Thinking” stood out: it appeared at the top of the TabArena benchmark for large datasets.
How TabICL and TabPFN handle missing values
You can give training or test data with missing values to the tabular foundation models TabPFN and TabICL, and the prediction will “just work”.
TabPFN and TabICL are not everything
An overview of other families for tabular foundation models
I’m writing a book on Tabular Foundation Models
tl;dr: In-progress book here: tabularfoundationmodels.com
Tabular ML is entering a new benchmark era
From static and narrow benchmarks to live, capability-driven evaluation