Research map / Probabilistic and scientific modeling / Causal discovery
Boosting Causal Additive Models
Maximilian Kertel, Nadja Klein
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 |
|---|---|---|---|
| increase sample size N from 10 to 200 | boosting causal-order score · Synthetic additive SEs, p=5 | transposition distance: 3.29 -> 0.03 | Improved |
| increase sample size N from 10 to 200 with non-additive SEs | boosting causal-order score · Synthetic non-additive SEs, p=5 | transposition distance: 3.26 -> 0.23 | Improved |
| DAGBoost versus DAGSNOTEARS | DAGBoost · Synthetic additive SF, p=100 | SHD: 176.96 -> 30.08 | Improved |
| DAGBoost versus CAM on non-additive SF graphs | DAGBoost · Synthetic non-additive SF, p=100 | SHD: 77.63 -> 67.47 | Improved |
| DAGBoost versus CAM on additive ER graphs | DAGBoost · Synthetic additive ER, p=100 | SHD: 12.07 -> 14.90 | Hurt |
| DAGBoost versus CAM on additive SF graphs | DAGBoost · Synthetic additive SF, p=100 | SHD: 30.18 -> 30.08 | No clear effect |
| DAGBoost versus CAM, precision | DAGBoost · Synthetic non-additive SF, p=100 | precision: 0.676 -> 0.930 | Improved |
| DAGBoost versus CAM, recall | DAGBoost · Synthetic additive SF, p=100 | recall: 0.817 -> 0.712 | Hurt |
| increase boosting step size from 0.3 to 0.9 (ablation) | DAGBoost · Synthetic ER additive, p=100 | runtime (s): 58.47 -> 20.17 | Improved |
| increase boosting step size from 0.3 to 0.9 (ablation) | DAGBoost · Synthetic ER additive, p=100 | SHD: 14.90 -> 13.17 | No clear effect |
| vary penalty lambda from 0.001 to 0.1 (ablation) | DAGBoost · Synthetic ER additive, p=100 | SHD: 15.58 -> 14.90 | No clear effect |
| increase penalty lambda from 0.001 to 0.1 (ablation) | DAGBoost · Synthetic ER additive, p=100 | runtime (s): 43.54 -> 204.87 | Hurt |
Closest papers by meaning
- Learning Structural Causal Models from Ordering: Identifiable Flow Models (AAAI 2025 · similarity 0.87)
- Scores for Learning Discrete Causal Graphs with Unobserved Confounders (AAAI 2024 · similarity 0.86)
- Standardizing Structural Causal Models (ICLR 2025 · similarity 0.85)
- Data-Driven Selection of Instrumental Variables for Additive Nonlinear, Constant Effects Models (ICML 2025 · similarity 0.85)
- Score-informed Neural Operator for Enhancing Ordering-based Causal Discovery (NeurIPS 2025 · similarity 0.85)
- Empirical Analysis of Model Selection for Heterogeneous Causal Effect Estimation (ICLR 2024 · similarity 0.85)
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.
