Research map / Probabilistic and scientific modeling / Causal discovery
DeepMed: Semiparametric Causal Mediation Analysis with Debiased Deep Learning
Siqi Xu, Lin Liu, Zhonghua Liu
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 |
|---|---|---|---|
| add input-layer L1 regularization (ablation) | DeepMed · Case 1, p=100, tau_tot | RMSE: 0.177 -> 0.047 | Improved |
| add input-layer L1 regularization (ablation) | DeepMed · Case 2, p=100, tau_NDE(1) | RMSE: 0.270 -> 0.044 | Improved |
| add input-layer L1 regularization (ablation) | DeepMed · Case 3, p=100, tau_tot | RMSE: 0.231 -> 0.047 | Improved |
| add input-layer L1 regularization (ablation) | DeepMed · Case 2, p=100, tau_tot | RMSE: 0.048 -> 0.038 | Improved |
| add input-layer L1 regularization (ablation) | DeepMed · Case 1, p=100, tau_NDE(0) | RMSE: 0.049 -> 0.040 | No clear effect |
| replace DNN nuisance estimators with Lasso (ablation) | DeepMed · Case 2, p=5, tau_NDE(1) | RMSE: 0.024 -> 0.380 | Hurt |
| replace DNN nuisance estimators with random forest (ablation) | DeepMed · Case 2, p=5, tau_NDE(1) | RMSE: 0.024 -> 0.193 | Hurt |
| replace DNN nuisance estimators with GBM (ablation) | DeepMed · Case 1, p=5, tau_tot | RMSE: 0.032 -> 0.039 | Hurt |
| replace DNN nuisance estimators with random forest (ablation) | DeepMed · Case 3, p=5, tau_tot | RMSE: 0.038 -> 0.043 | No clear effect |
| use true nuisance functions (Oracle) instead of DNN estimates (ablation) | DeepMed · Case 1, p=5, tau_tot | RMSE: 0.032 -> 0.029 | Improved |
| replace DNN with Lasso (ablation) | DeepMed · COMPAS tau_NDE(1) | estimate: 0.564 -> 0.575 | No clear effect |
| replace DNN with GBM (ablation) | DeepMed · COMPAS tau_NIE(1) | estimate: 0.612 -> 0.654 | No clear effect |
Closest papers by meaning
- A Neural Framework for Generalized Causal Sensitivity Analysis (ICLR 2024 · similarity 0.84)
- Causal Fairness under Unobserved Confounding: A Neural Sensitivity Framework (ICLR 2024 · similarity 0.83)
- Causal Bias Detection in Generative Artifical Intelligence (NeurIPS 2026 · similarity 0.83)
- When Additive Noise Meets Unobserved Mediators: Bivariate Denoising Diffusion for Causal Discovery (NeurIPS 2025 · similarity 0.83)
- Modular Learning of Deep Causal Generative Models for High-dimensional Causal Inference (ICML 2024 · similarity 0.83)
- Quantifying Aleatoric Uncertainty of the Treatment Effect: A Novel Orthogonal Learner (NeurIPS 2024 · similarity 0.83)
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.
