AdaPS-LiNGAM: Adaptive Predecessor Selection for Linear Non-Gaussian Acyclic Models under Small-Sample Settings

Key points

• Causal discovery becomes particularly challenging when the available sample size is small relative to the number of variables.1
• We establish a structural limitation of this procedure: when the number of variables exceeds the sample size, repeated residualization necessarily becomes degenerate before the full causal order can be determined.1
• Our analysis further reveals that each residual can be reconstructed using only a graph-determined subset of variables already placed earlier in the causal order, termed the active boundary.1
• Experiments on synthetic data demonstrate that AdaPS-LiNGAM provides accurate causal-structure recovery in sample-limited settings and degrades more gradually as the sample size decreases.1

Sources (1)

• [1]AdaPS-LiNGAM: Adaptive Predecessor Selection for Linear Non-Gaussian Acyclic Models under Small-Sample Settings
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 10:01 AM

  This challenge also arises in the linear non-Gaussian acyclic model (LiNGAM), an identifiable framework for causal discovery from observational data.

  Causal discovery becomes particularly challenging when the available sample size is small relative to the number of variables.

Extractive summary: sentences quoted from the sources.