Reproduce “Toward Scalable and Valid Conditional Independence Testing with Spectral Representations” 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
Alek Fröhlich, Vladimir R. Kostic, Karim Lounici, Daniel Perazzo, Daniel Tiezzi, Massimiliano Pontil
SpectralCIT learns features to test whether X and Y remain dependent given Z. A bi-level contrastive procedure learns features for X, (Y,Z), and Z from the partial covariance operator, then evaluates a statistic on held-out data against a chi-squared threshold. In one breast-cancer analysis, it returned p < 10⁻³; its validity and power depend on feature quality and stated assumptions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.