Research map / Probabilistic and scientific modeling / Treatment effects
An Efficient Doubly-Robust Test for the Kernel Treatment Effect
Diego Martinez-Taboada, Aaditya Ramdas, Edward Kennedy
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 |
|---|---|---|---|
| doubly robust AIPW estimator instead of IPW (beta=0) (ablation) | AIPW-xKTE · synthetic Scenarios II-IV, experimental setting | power | Improved |
| replace permutation test (KTE) with permutation-free cross U-statistic AIPW-xKTE | AIPW-xKTE · synthetic experimental setting, n=350 | runtime (ms): 89.875 -> 4.940 | Improved |
| replace permutation test (KTE) with AIPW-xKTE | AIPW-xKTE · synthetic Scenarios II-IV, experimental setting | power | No clear effect |
| IPW-xKTE versus permutation-based KTE | IPW-xKTE · synthetic experimental setting | power | Hurt |
| distributional kernel test versus mean-based tests (Causal Forests, BART, Baseline-AIPW) | AIPW-xKTE · IHDP Scenario II (mean shift) | true positive rate: 0.95 -> 0.44 | Hurt |
| distributional kernel test versus Baseline-AIPW | AIPW-xKTE · IHDP Scenario III | true positive rate: 0.00 -> 0.34 | Improved |
| distributional kernel test versus Baseline-AIPW | AIPW-xKTE · IHDP Scenario IV | true positive rate: 0.00 -> 0.53 | Improved |
| distributional kernel test versus BART | AIPW-xKTE · IHDP Scenario V (IQ score) | true positive rate: 0.10 -> 0.99 | Improved |
| distributional kernel test versus Causal Forests | AIPW-xKTE · IHDP Scenario III | true positive rate: 0.04 -> 0.34 | Improved |
| distributional kernel test versus Causal Forests | AIPW-xKTE · IHDP Scenario II | true positive rate: 1.00 -> 0.44 | Hurt |
| AIPW-xKTE under zero mean effect (Scenario VI) | AIPW-xKTE · IHDP Scenario VI | rejection rate: 0.03 | No clear effect |
| non-linear outcome model cos(beta^T X) instead of linear | AIPW-xKTE · synthetic Scenario II | power | Improved |
Closest papers by meaning
- Doubly Robust Proximal Causal Learning for Continuous Treatments (ICLR 2024 · similarity 0.88)
- Semiparametric Efficient Tests for Interpretable Distributional Treatment Effects (NeurIPS 2026 · similarity 0.87)
- Density Ratio-Free Doubly Robust Proxy Causal Learning (NeurIPS 2025 · similarity 0.86)
- Doubly-Robust Estimation of Counterfactual Policy Mean Embeddings (NeurIPS 2025 · similarity 0.85)
- Doubly robust identification of treatment effects from multiple environments (ICLR 2025 · similarity 0.85)
- Counterfactual Density Estimation using Kernel Stein Discrepancies (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.
