Research map / Probabilistic and scientific modeling / Causal discovery
Directed Cyclic Graphs for Simultaneous Discovery of Time-Lagged and Instantaneous Causality from Longitudinal Data Using Instrumental Variables
Wei Jin, Yang Ni, Amanda B. Spence, Leah H. Rubin, Yanxun Xu
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 from 1,000 to 5,000 | Proposed Bayesian DCG model · Scenario I, edge Yj1-Yj4, instrument Xj1 | edge detection probability: 23% -> 98% | Improved |
| increase sample size from 1,000 to 5,000 | LiNG-D · Scenario I, edge Yj1-Yj4, instrument Xj1 | edge detection probability: 37% -> 96% | Improved |
| increase sample size from 1,000 to 5,000 | PCMCI+ · Scenario I, edge Yj1-Yj4, instrument Xj1 | edge detection probability: 47% -> 100% | Improved |
| increase sample size from 1,000 to 5,000 | Proposed Bayesian DCG model · Scenario I, edge Yj1-Yj4, instrument Yj-1,1 | edge detection probability: 27% -> 95% | Improved |
| use PCMCI+ instead of the proposed model with lagged outcome as instrument | Scenario I edge detection · Scenario I, edge Yj1-Yj4, instrument Yj-1,1, n=1000 | edge detection probability: 27% -> 52% | Improved |
| use Yj-1,1 instead of Xj1 as instrument for Yj1 (proposed model, n=1000) (ablation) | Proposed Bayesian DCG model · Scenario I, edge Yj1-Yj4 | edge detection probability: 23% -> 27% | Mixed |
| reduce causal effect size from 1 to 0.5 (ablation) | LiNG-D · Scenario II, edge Yj1-Yj3 | edge detection probability: almost 100% -> 3% | Hurt |
| reduce causal effect size from 1 to 0.5 (ablation) | Proposed Bayesian DCG model · Scenario II, edge Yj1-Yj3 | edge detection probability: almost 100% -> 70% | Hurt |
| reduce causal effect size from 1 to 0.5 (ablation) | PCMCI+ · Scenario II, spurious edge Yj-1,2 to Yj3 | false positive probability: 48% -> 1% | Improved |
| reduce causal effect size from 1 to 0.5 (ablation) | VAR-LiNGAM · Scenario II, spurious edge Yj-1,3 to Yj1 | false positive probability: 84% -> 0% | Improved |
Closest papers by meaning
- Discovering Mixtures of Structural Causal Models from Time Series Data (ICML 2024 · similarity 0.90)
- Coarsening Linear Non-Gaussian Causal Models with Cycles (NeurIPS 2026 · similarity 0.89)
- Learning Mixtures of Unknown Causal Interventions (NeurIPS 2024 · similarity 0.89)
- CausalTab: Pretraining Across Causal Environments for Tabular Causal Discovery (NeurIPS 2026 · similarity 0.89)
- Differentiable Causal Discovery for Latent Hierarchical Causal Models (ICLR 2025 · similarity 0.89)
- Sample Efficient Bayesian Learning of Causal Graphs from Interventions (NeurIPS 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.
