Reproduce “Differentiable Causal Backdoor Discovery” 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
Limor Gultchin, Matt J. Kusner, Varun Kanade, Ricardo Silva
The method learns a sparse covariate adjustment for estimating causal effects from observational data. It uses an observed auxiliary variable associated with treatment to select covariates through gradient-based optimization rather than searching subsets. In an NHS-survey-derived evaluation under a fitted linear-Gaussian model, its absolute ATE error was 0.163, versus 0.324 for all covariates; this does not establish real-world effects.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.