Research map / Probabilistic and scientific modeling / Treatment effects
Breaking the Curse of Nonregularity with Subagging --- Inference of the Mean Outcome under Optimal Treatment Regimes
Chengchun Shi, Wenbin Lu, Rui Song
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 subagging CI instead of online one-step CI | optimal value inference · Scenario (E), n=500 | empirical coverage probability (%): 88.2 -> 94.4 | Improved |
| use subagging CI instead of online one-step CI | optimal value inference · Scenario (F), n=500 | empirical coverage probability (%): 87.5 -> 92.9 | Improved |
| use subagging CI instead of single sample-splitting CI | optimal value inference · Scenario (A), n=500 | average CI length x100: 17.1 -> 11.2 | Improved |
| use subagging CI instead of oracle CI | optimal value inference · Scenario (A), n=500 | average CI length x100: 13.1 -> 11.2 | Improved |
| use subagging CI instead of online one-step CI in a regular case | optimal value inference · Scenario (D), n=1000 | average CI length x100: 27.1 -> 25.7 | Improved |
| use subagging CI instead of online one-step CI (l_n=200) | two-stage optimal value inference · Multi-stage Setting (G), n=600 | empirical coverage probability (%): 82.1 -> 93.6 | Improved |
| use subagging CI instead of online one-step CI (l_n=200) | two-stage optimal value inference · Multi-stage Setting (H), n=600 | average CI length x100: 45.3 -> 37.3 | Improved |
| vary tuning constant K0 from 1 to 6 in s_n = K0 n/log n (ablation) | subagging CI · Scenarios (A)-(F), n=1000 | empirical coverage and length | No clear effect |
| vary K0 among 3, 3.5, 4 (ablation) | subagging CI · ACTG175 | 95% CI length: 23.4 -> 23.6 -> 23.8 | No clear effect |
| increase online one-step l_n from 50 to 100 to 200 (ablation) | online one-step CI · ACTG175 | 95% CI length: 27.1 -> 27.7 -> 28.9 | Hurt |
Closest papers by meaning
- Assessing the robustness of heterogeneous treatment effects in survival analysis under informative censoring (NeurIPS 2026 · similarity 0.85)
- An effective framework for estimating individualized treatment rules (NeurIPS 2024 · similarity 0.85)
- Doubly Protected Estimation for Survival Outcomes Utilizing External Controls for Randomized Clinical Trials (ICML 2025 · similarity 0.85)
- Heterogeneous Treatment Effect in Time-to-Event Outcomes: Harnessing Censored Data with Recursively Imputed Trees (ICML 2025 · similarity 0.84)
- Doubly robust identification of treatment effects from multiple environments (ICLR 2025 · similarity 0.83)
- Optimistic Algorithms for Adaptive Estimation of the Average Treatment Effect (ICML 2025 · 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.
