Reproduce “TabBioMed: A Large-Scale Benchmark for Biomedical Tabular Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Pau Mateo-Bernadó, Saivenkata Nagavyjayanthi Polapragada, David Bonet, Pol Arbiol Rakuljic, Laia Mogas Pladevall, Marçal Comajoan Cara, Jesus Gonzalez-Ferrer, Jordi Abante and 3 more
TabBioMed benchmarks 94 biomedical datasets across eight domains with 23 models from four families and dataset-specific splits. In the aggregate comparison, TabDPT has the highest mean normalized score, 0.865, while TabICLv2 has the best mean rank. The leading models are not statistically separated, and LightGBM offers near-frontier performance with much lower inference latency. The benchmark supports comparison, not clinical deployment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.