Research map / Graph learning / Graph algorithms and clustering
Iterative Causal Discovery in the Possible Presence of Latent Confounders and Selection Bias
Raanan Yehezkel Rohekar, Shami Nisimov, Yaniv Gurwicz, Gal Novik
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 |
|---|---|---|---|
| Replace FCI with ICD (conditioning-set size tied to distance from tested nodes) | FCI · random DAGs, 15-35 nodes, CI oracle | number of CI tests: fewer for all 100 DAGs | Improved |
| Replace FCI with ICD | FCI · random DAGs 15-35 nodes | runtime ratio FCI/ICD: 1.3, 1.8, 2.9, 5.6 | Improved |
| Replace FCI+ / RFCI with ICD (all ICD-Sep conditions) | FCI+, RFCI · random DAGs 15-35 nodes, CI oracle | number of CI tests | Improved |
| Replace FCI with ICD at 3000 samples | FCI · 15-node DAGs, 3000 samples | CI tests ratio: 1.0 -> 0.08 | Improved |
| Replace FCI with ICD at 100 samples | FCI · 15-node DAGs, 100 samples | CI tests ratio: 1.0 -> 0.5 | Improved |
| Replace FCI with ICD at 3000 samples | FCI · 15-node DAGs, 3000 samples | skeleton F1 score: 0.69 -> 0.77 | Improved |
| Replace FCI with ICD at 3000 samples | FCI · 15-node DAGs, 3000 samples | orientation accuracy: 0.38 -> 0.46 | Improved |
| Replace FCI with ICD at 3000 samples | FCI · 15-node DAGs, 3000 samples | false-negative ratio: 0.447 -> 0.209 | Improved |
| Replace FCI with ICD at 3000 samples | FCI · 15-node DAGs, 3000 samples | false-positive ratio: 0.005 -> 0.111 | Hurt |
| Replace FCI+ with ICD at 1000 samples | FCI+ · 15-node DAGs, 1000 samples | skeleton F1 score: 0.70 -> 0.73 | Improved |
| Replace RFCI with FCI+ | RFCI · random DAGs 15-35 nodes | number of CI tests | No clear effect |
| Increase graph size | ICD vs FCI · random DAGs | CI test advantage | Improved |
Closest papers by meaning
- Causal Discovery with Fewer Conditional Independence Tests (ICML 2024 · similarity 0.92)
- Adaptive Online Experimental Design for Causal Discovery (ICML 2024 · similarity 0.91)
- Foundations of Testing for Finite-Sample Causal Discovery (ICML 2024 · similarity 0.90)
- Differentiable Cyclic Causal Discovery Under Unmeasured Confounders (NeurIPS 2025 · similarity 0.90)
- Stable Differentiable Causal Discovery (ICML 2024 · similarity 0.90)
- THE ROBUSTNESS OF DIFFERENTIABLE CAUSAL DISCOVERY IN MISSPECIFIED SCENARIOS (ICLR 2025 · similarity 0.90)
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.
