Reproduce “Local Causal Discovery with Linear non-Gaussian Cyclic Models” 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

Haoyue Dai, Ignavier Ng, Yujia Zheng, Zhengqing Gao, Kun Zhang

Local ISA-LiNG identifies local directed, weighted causal structure in linear models with independent non-Gaussian noise. It applies independent subspace analysis to a target and its Markov blanket, allowing signals to remain grouped when they cannot be separated. For PIP2 in a protein dataset, it reports SHD 1, versus 1.5 for GSBN and 3.6 for Local A*, under the correct-blanket assumption.

Summary by alphaXiv.

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