Hugging Face Papers ↗ · 2026-11-18
TabBioMed: A Large-Scale Benchmark for Biomedical Tabular Learning
Media reports Hugging Face Papers Source language: Chinese
TabBioMed provides 94 curated biomedical tabular datasets for classification, regression and multi‑target tasks; foundation models lead overall, while tree‑based models and tuned MLPs remain competitive under cost constraints.
Full text
A large-scale benchmark of 94 curated biomedical tabular datasets standardizes evaluation across classification, regression, and multi-target tasks and finds that foundation models lead overall while tree-based models and tuned multilayer perceptrons remain competitive under cost constraints.