Research map / Probabilistic and scientific modeling / Causal discovery
Skewness-Robust Causal Discovery in Location-Scale Noise Models
Daniel Klippert, Alexander Marx
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use independence testing instead of likelihood scoring | SkewD · 13 established benchmarks | mean accuracy (%): 77.43 -> 85.02 | Improved |
| use independence testing instead of likelihood scoring | LOCI · 13 established benchmarks | mean accuracy (%): 77.08 -> 84.09 | Improved |
| model skew-normal noise instead of Gaussian noise (SkewD-LL vs LOCI-LL) | likelihood scoring · 13 established benchmarks | mean accuracy (%): 77.08 -> 77.43 | No clear effect |
| model skew-normal noise instead of Gaussian noise (SkewD-IT vs LOCI-IT) | independence testing · 13 established benchmarks | mean accuracy (%): 84.09 -> 85.02 | Improved |
| use SkewD-IT instead of ROCHE | cause-effect inference · 13 established benchmarks | mean accuracy (%): 87.69 -> 85.02 | Hurt |
| increase noise skewness to LSs(1.750) | ROCHE · LSs skew-noise datasets | accuracy: dropping to 81% | Hurt |
| increase noise skewness across the three levels | CAREFL · skew-noise LSNM datasets | accuracy: 0.90, 0.71, 0.66 | Hurt |
| use oracle MEC instead of PC-estimated MEC (ablation) | SkewD-MV · synthetic LSNM graphs d=6 | SHD: 2.27 -> 0.14 | Improved |
| use oracle MEC instead of PC-estimated MEC (ablation) | SkewD-MV · synthetic LSNM graphs d=10 | SHD: 4.72 -> 0.04 | Improved |
| use SkewD-MV instead of HOST | causal discovery · synthetic LSNM graphs d=10 | SHD: 21.21 -> 4.72 | Improved |
| use SkewD-MV instead of CAM (SID) | causal discovery · synthetic LSNM graphs d=6 | SID: 9.23 -> 4.35 | Improved |
| use SkewD-LL instead of next best method | cause-effect inference · SIMG | accuracy: 88% accuracy, at least 5% above all others | Improved |
Closest papers by meaning
- A Skewness-Based Criterion for Addressing Heteroscedastic Noise in Causal Discovery (ICLR 2025 · similarity 0.93)
- Causal Discovery via Transformed Low-Rank Quantile Surfaces (NeurIPS 2026 · similarity 0.88)
- Bivariate Causal Discovery using Bayesian Model Selection (ICML 2024 · similarity 0.88)
- Hybrid Top-Down Global Causal Discovery with Local Search for Linear and Nonlinear Additive Noise Models (NeurIPS 2024 · similarity 0.87)
- Strong and Weak Identifiability of Optimization-based Causal Discovery in Non-linear Additive Noise Models (ICML 2025 · similarity 0.87)
- When Additive Noise Meets Unobserved Mediators: Bivariate Denoising Diffusion for Causal Discovery (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
