Research map / Probabilistic and scientific modeling / Causal discovery
Multi-View Causal Discovery without Non-Gaussianity: Identifiability and Algorithms
Ambroise Heurtebise, Omar Chehab, Pierre Ablin, Alexandre Gramfort, Aapo Hyvarinen
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 PairwiseLiMVAM and DirectLiMVAM (second-order statistics, multi-view) instead of non-Gaussian methods on Gaussian disturbances | multi-view causal discovery · synthetic, shared Gaussian disturbances, 50 seeds | l2 error of B and wrong causal order rate | Improved |
| apply non-Gaussian methods (MultiGroupDirectLiNGAM, ICA-LiNGAM) in Gaussian setting | baselines · synthetic, Gaussian disturbances | estimation error | Hurt |
| use ICA-LiMVAM (Shared ICA) instead of PairwiseLiMVAM/DirectLiMVAM | multi-view causal discovery · synthetic, shared Gaussian disturbances | estimation error | Hurt |
| apply PairwiseLiMVAM/DirectLiMVAM when equal-variance Gaussian disturbances break other methods' assumptions | PairwiseLiMVAM, DirectLiMVAM · synthetic, m=6 views, p=5, n=1000, 200 runs | adjacency estimation error vs fitting time | Improved |
| apply ICA-LiNGAM separately to each view rather than exploiting shared ordering | ICA-LiNGAM · synthetic, shared disturbances | causal order error rate | Hurt |
| compare DirectLiMVAM against PairwiseLiMVAM | LiMVAM algorithms · appendix experiments | estimation error | Mixed |
| rerun PairwiseLiMVAM on random 30-participant subsets of MEG data (ablation) | PairwiseLiMVAM · Cam-CAN MEG | correlation of median causal matrices: average correlation: 0.67 | No clear effect |
Closest papers by meaning
- Causal Discovery with Unobserved Confounding and Non-Gaussian Data (JMLR 2023 · similarity 0.90)
- Differentiable Causal Discovery for Latent Hierarchical Causal Models (ICLR 2025 · similarity 0.89)
- Strong and Weak Identifiability of Optimization-based Causal Discovery in Non-linear Additive Noise Models (ICML 2025 · similarity 0.88)
- Coarsening Linear Non-Gaussian Causal Models with Cycles (NeurIPS 2026 · similarity 0.88)
- Generating Hypotheses of Dynamic Causal Graphs in Neuroscience: Leveraging Generative Factor Models of Observed Time Series (ICML 2025 · similarity 0.88)
- Since Faithfulness Fails: The Performance Limits of Neural Causal Discovery (ICML 2025 · similarity 0.88)
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.
