Reproduce “One Table, Many Queries: Benchmarking Generalist Tabular Models Across Generation, Imputation, and Prediction” 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
Yuanrui Wang, Kaifang Li, Xuanyue Li, Ziyun Li, Yan Lu, Xingxuan Zhang, Peng Cui
OTBench evaluates tabular models on generating rows, imputing missing features, and predicting targets, then measures which capabilities align across tasks. It compares 56 datasets, 46 methods, and 20 capability axes under shared splits. Rank associations link imputation and prediction on numerical recovery and utility, while generation fidelity need not track predictive usefulness; these results describe alignment, not causal transfer.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.