Reproduce “A Skewness-Based Criterion for Addressing Heteroscedastic Noise in Causal 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
Yingyu Lin, Yuxing Huang, Wenqin Liu, Haoran Deng, Ignavier Ng, Kun Zhang, Mingming Gong, Yi-An Ma and 1 more
SkewScore uses score-component skewness to infer direction when noise spread varies with the cause. It compares score components without extracting residual noise. Under its assumptions, the effect component is unskewed, while the cause component is nonzero under an additional condition. In one synthetic setting, it achieves 98% accuracy versus HOST’s 89%, a result limited to that setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.