Research map / Probabilistic and scientific modeling / Treatment effects
The Adaptive Doubly Robust Estimator and a Paradox Concerning Logging Policy
Masahiro Kato, Kenichiro McAlinn, Shota Yasui
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 ADR estimator instead of AIPW with true logging policy | ADR · Section 6.1 LinUCB, T=250 | RMSE: 0.158 -> 0.061 | Improved |
| use ADR instead of AIPW | ADR · Section 6.1 LinUCB, T=750 | RMSE: 0.088 -> 0.033 | Improved |
| use ADR instead of AIPW | ADR · Section 6.1 LinTS, T=750 | RMSE: 0.060 -> 0.030 | Improved |
| use ADR instead of IPW with true logging policy | ADR · Section 6.1 LinUCB, T=250 | RMSE: 0.205 -> 0.061 | Improved |
| use ADR instead of EIPW (estimated logging policy) | ADR · Section 6.1 LinUCB, T=750 | RMSE: 0.134 -> 0.033 | Improved |
| use ADR instead of DM | ADR · Section 6.1 LinUCB, T=250 | RMSE: 0.073 -> 0.061 | Improved |
| replace AIPW with AW-AIPW (StableVar weight) | AIPW · Section 6.1 LinUCB, T=250 | RMSE: 0.158 -> 0.179 | No clear effect |
| use ADR instead of AIPW | ADR · mnist LinUCB, T=1000, alpha=0.4 | RMSE: 0.112 -> 0.028 | Improved |
| use ADR instead of AIPW | ADR · satimage LinUCB, T=1000, alpha=0.4 | RMSE: 0.019 -> 0.022 | Mixed |
| use ADR instead of AIPW at alpha=0.1 | ADR · mnist LinUCB, T=1000, alpha=0.1 | RMSE: 0.085 -> 0.086 | Hurt |
| increase sample size T from 250 to 750 | ADR · Section 6.1 LinUCB | RMSE: 0.061 -> 0.033 | Improved |
| increase sample size T from 250 to 750 | EIPW · Section 6.1 LinUCB | coverage ratio (CR): 0.81 -> 0.13 | Hurt |
Closest papers by meaning
- Optimistic Algorithms for Adaptive Estimation of the Average Treatment Effect (ICML 2025 · similarity 0.88)
- Off-policy estimation with adaptively collected data: the power of online learning (NeurIPS 2024 · similarity 0.86)
- Efficient Adaptive Experimentation with Noncompliance (NeurIPS 2025 · similarity 0.86)
- Doubly Robust Proximal Causal Learning for Continuous Treatments (ICLR 2024 · similarity 0.86)
- Simulation-Based Inference for Adaptive Experiments (NeurIPS 2025 · similarity 0.85)
- Asymptotically Log-Optimal Bayes-Assisted Confidence Sequences for Bounded Means (NeurIPS 2026 · 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.
