Reproduce “The Blessings of Multiple Causes” 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

Yixin Wang, David M. Blei

The deconfounder models dependence among multiple causes with a latent-variable factor model, infers substitute confounders, and uses them in outcome models. It checks the assignment model by predicting held-out causes before estimating effects. In a semi-simulated smoking study, bias fell from 20.02 to 11.55 and MSE from 24.80 to 17.51 (×10⁻²), but validity still requires single ignorability and additional conditions.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.