Reproduce “Exponential Family Synthetic Controls” 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
Hector Rodriguez-Deniz, David M. Blei
EFSC estimates counterfactual distributions in panel studies where each unit-time cell contains a dataset. It models cell distributions with exponential-family natural parameters and uses shared latent factors, learned from untreated cells, to reconstruct treated cells’ missing untreated distributions. In a sparse Gaussian-panel test, EFSC-PMF’s KL-recovery MAE was 0.061 versus 0.812 for cell-wise MLE; this gap narrows with larger cells.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.