Research map / Probabilistic and scientific modeling / Causal discovery
Identifiability of Causal Graphs under Non-Additive Conditionally Parametric Causal Models
Juraj Bodik, Valérie Chavez-Demoulin
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 a one-parameter family F matching exponential support and tail (Pareto shifted) instead of Gaussian fixed sigma (ablation) | CPCM · Misspecified F, random theta | direction accuracy: 0.01 -> 0.99 | Improved |
| use a two-parameter Gamma instead of the one-parameter oracle exponential (ablation) | CPCM · Misspecified F, theta(x)=x^2+1 | direction accuracy: 0.99 -> 0.64 | Hurt |
| use a three-parameter family (ablation) | CPCM · Misspecified F | direction accuracy: about 50% | Hurt |
| use two-parameter Gaussian instead of fixed sigma (ablation) | CPCM · Misspecified F, random theta | direction accuracy: 0.01 -> 0.69 | Improved |
| CPCM instead of LOCI on Pareto pairs | CPCM · Bivariate PARg | accuracy (%): 0 -> 90 | Improved |
| CPCM instead of bQCD on Pareto pairs | CPCM · Bivariate PARg | accuracy (%): 34 -> 90 | Improved |
| CPCM instead of LOCI on Poisson pairs | CPCM · Bivariate POISg | accuracy (%): 79 -> 94 | Improved |
| CPCM instead of LOCI on multiplicative noise | CPCM · Bivariate MNs | accuracy (%): 99 -> 88 | Hurt |
| CPCM instead of LINGAM on linear non-Gaussian data | CPCM · Multivariate d=5, Linear | SID: 0.12 -> 0.22 | Hurt |
| CPCM instead of ANM RESIT on CPCM(Exp) data | CPCM · Multivariate d=5, CPCM(Exp) | SID: 12.21 -> 2.16 | Improved |
| use RESIT-greedy instead of RESIT (ablation) | ANM · Multivariate d=5, Linear | SID: 1.10 -> 0.12 | Improved |
| increase gamma from 0 to positive in Pareto CPCM (ablation) | CPCM conservative algorithm · Pareto model, n=500 | direction recovery | Improved |
Closest papers by meaning
- On the Parameter Identifiability of Partially Observed Linear Causal Models (NeurIPS 2024 · similarity 0.90)
- Identifying Causal Effects Under Functional Dependencies (NeurIPS 2024 · similarity 0.90)
- Bivariate Causal Discovery using Bayesian Model Selection (ICML 2024 · similarity 0.90)
- Testing Causal Models with Hidden Variables in Polynomial Delay via Conditional Independencies (AAAI 2025 · similarity 0.90)
- Identifiable Latent Polynomial Causal Models through the Lens of Change (ICLR 2024 · similarity 0.89)
- Causal Effect Identification in LiNGAM Models with Latent Confounders (ICML 2024 · similarity 0.89)
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.
