Evaluating TabPFN for Mild Cognitive Impairment to Alzheimer's Disease Conversion in Data Limited Settings
The study evaluates TabPFN, a tabular foundation model, for predicting the conversion from Mild Cognitive Impairment (MCI) to Alzheimer's Disease (AD) using limited data from the TADPOLE dataset. TabPFN achieved high performance with an AUC of 0.892, outperforming traditional models like LightGBM, especially in low-data settings with as few as 50 training samples. The results suggest that foundation models like TabPFN show promise for disease prediction in scenarios with limited longitudinal data.
- ▪TabPFN achieved an AUC of 0.892 in predicting 3-year MCI to AD conversion.
- ▪TabPFN outperformed traditional models including XGBoost, Random Forest, LightGBM, and Logistic Regression.
- ▪In settings with only 50 training samples, TabPFN maintained strong performance while traditional models struggled.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2604.27195 (cs) [Submitted on 29 Apr 2026] Title:Evaluating TabPFN for Mild Cognitive Impairment to Alzheimer's Disease Conversion in Data Limited Settings Authors:Brad Ye, Bulent Soykan, Gulsah Hancerliogullari Koksalmis, Hsin-Hsiung Huang, Laura J. Brattain View a PDF of the paper titled Evaluating TabPFN for Mild Cognitive Impairment to Alzheimer's Disease Conversion in Data Limited Settings, by Brad Ye and 4 other authors View PDF HTML (experimental) Abstract:Accurate prediction of conversion from Mild Cognitive Impairment (MCI) to Alzheimers Diseases (AD) is essential for early intervention, however, developing reliable conversion predictive models is difficult to develop due to limited longitudinal data availability We evaluate TabPFN…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv.org.
Discussion
0 commentsMore from arXiv.org
