Research map / Probabilistic and scientific modeling / Treatment effects
A Relative Error-Based Evaluation Framework of Heterogeneous Treatment Effect Estimators
Jiayi Guo, Haoxuan Li, YE TIAN, Peng Wu
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 proposed network with new loss and balance regularizers vs best baseline | HTE estimation · IHDP in-sample | sqrt(PEHE): 0.741 -> 0.638 | Improved |
| proposed method vs best baseline | HTE estimation · Twins in-sample | sqrt(PEHE): 0.290 -> 0.284 | Improved |
| proposed method vs LinDML on Twins ATE | HTE estimation · Twins in-sample | ATE error: 0.013 -> 0.009 | Improved |
| use proposed nuisance network instead of linear regression | relative error estimation · IHDP | selection accuracy: 0.44 -> 0.80 | Improved |
| use proposed nuisance network instead of gradient boosting | relative error estimation · Twins | selection accuracy: 0.86 -> 0.94 | Improved |
| conventional nuisance estimators | relative error estimation · IHDP/Twins | coverage: 0.94 vs 0.96 | No clear effect |
| add constraint loss L_const (ablation) | proposed network · IHDP in-sample | sqrt(PEHE): 3.495 -> 0.638 | Improved |
| add cross-entropy loss L_ce (ablation) | proposed network · IHDP in-sample | sqrt(PEHE): 0.725 -> 0.638 | Improved |
| add L_ce on Twins (ablation) | proposed network · Twins in-sample | sqrt(PEHE): 0.284 -> 0.284 | No clear effect |
| increase lambda_2 from 0.01 to 1 (ablation) | proposed network · IHDP selection | selection accuracy: 0.50 -> 0.80 | Improved |
| vary lambda_2 between 0.5 and 5 (ablation) | proposed network · IHDP | sqrt(PEHE) in: 0.714 / 0.638 / 0.715 | No clear effect |
Closest papers by meaning
- A Non-parametric Direct Learning Approach to Heterogeneous Treatment Effect Estimation under Unmeasured Confounding (NeurIPS 2024 · similarity 0.89)
- Treatment Effect Estimation for Optimal Decision-Making (NeurIPS 2025 · similarity 0.87)
- PairNet: Training with Observed Pairs to Estimate Individual Treatment Effect (ICML 2024 · similarity 0.86)
- AutoCATE: End-to-End, Automated Treatment Effect Estimation (ICML 2025 · similarity 0.86)
- Unveiling the Potential of Robustness in Selecting Conditional Average Treatment Effect Estimators (NeurIPS 2024 · similarity 0.86)
- Orthogonal Survival Learners for Estimating Heterogeneous Treatment Effects from Time-to-Event Data (NeurIPS 2025 · similarity 0.86)
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.
