Research map / Probabilistic and scientific modeling / Treatment effects
Rank-Learner: Orthogonal Ranking of Treatment Effects
Henri Arno, Dennis Frauen, Emil Javurek, Thomas Demeester, Stefan Feuerriegel
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 orthogonal pseudo labels instead of plug-in soft targets (ablation) | ranking learner · Synthetic benchmark (n=100) | AUTOC: 0.69 -> 1.00 | Improved |
| add orthogonal pseudo labels instead of plug-in soft targets (ablation) | ranking learner · Synthetic benchmark (n=250) | AUTOC: 0.95 -> 1.28 | Improved |
| add orthogonal pseudo labels instead of plug-in soft targets (ablation) | ranking learner · Synthetic benchmark (n=2000) | AUTOC: 1.36 -> 1.37 | No clear effect |
| target ranking directly instead of T-learner CATE estimation | Rank-learner · Synthetic benchmark (n=100) | AUTOC: 0.88 -> 1.00 | Improved |
| target ranking directly instead of DR-learner CATE estimation | Rank-learner · Synthetic benchmark (n=500) | AUTOC: 1.28 -> 1.31 | Improved |
| target ranking directly instead of DR-learner | Rank-learner · Semi-synthetic CPS (n=1000) | AUTOC: 0.92 -> 0.95 | Improved |
| add orthogonal ranking instead of plug-in ranker (ablation) | ranking learner · Semi-synthetic MIMIC-III (n=1000) | AUTOC: 1.11 -> 1.18 | Improved |
| add orthogonal ranking instead of plug-in ranker (ablation) | ranking learner · Semi-synthetic MovieLens (n=1000) | AUTOC: 1.30 -> 1.35 | Improved |
| use Rank-learner instead of DR-learner | uplift ranking · Criteo uplift (50k test) | AUUC x10^3: 4.44 -> 5.19 | Improved |
| use Rank-learner instead of DR-learner | uplift ranking · Criteo uplift (1M test) | AUUC x10^3: 5.17 -> 5.90 | Improved |
| subsample training pairs per epoch (ablation) | Rank-learner · Synthetic benchmark (pair subsampling) | AUTOC | No clear effect |
| decrease overlap in propensity mechanism | Rank-learner · Synthetic benchmark (overlap sensitivity) | AUTOC | Hurt |
Closest papers by meaning
- A Meta-learner for Heterogeneous Effects in Difference-in-Differences (ICML 2025 · similarity 0.87)
- PairNet: Training with Observed Pairs to Estimate Individual Treatment Effect (ICML 2024 · similarity 0.87)
- Quantifying Aleatoric Uncertainty of the Treatment Effect: A Novel Orthogonal Learner (NeurIPS 2024 · similarity 0.86)
- AutoCATE: End-to-End, Automated Treatment Effect Estimation (ICML 2025 · similarity 0.86)
- Treatment Effect Estimation for Optimal Decision-Making (NeurIPS 2025 · 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.
