Research map / Probabilistic and scientific modeling / Treatment effects
Estimation of Treatment Effects Under Nonstationarity via the Truncated Policy Gradient Estimator
Ramesh Johari, Tianyi Peng, Wenqian Xing
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 |
|---|---|---|---|
| accumulate k=1 step rewards instead of instantaneous outcome (DM) | DM estimator · Two-state nonstationary MDP | mean absolute error (%): 48.71% -> 27.11% | Improved |
| increase truncation size from k=1 to k=3 | TPG estimator · Two-state nonstationary MDP | mean absolute error (%): 27.11% -> 23.17% | Improved |
| use untruncated PG (k=T) instead of truncated (ablation) | TPG estimator · Two-state nonstationary MDP | absolute error and variance | Hurt |
| use TPG with k=4 or 5 instead of stationary DQ or DM | DM estimator · Case Study I queueing simulator | bias | Improved |
| increase the alpha threshold for k selection from 1.036 to 1.960 (ablation) | heuristic k selection · Case Study I | RMSE: 0.043 -> 0.050 | No clear effect |
| use TPG instead of DM across switchback intervals | DM estimator · NYC ride-sharing simulator | bias | Improved |
| increase truncation size k (ablation) | TPG estimator · NYC ride-sharing simulator | variance | Hurt |
Closest papers by meaning
- Do Contemporary Causal Inference Models Capture Real-World Heterogeneity? Findings from a Large-Scale Benchmark (ICLR 2025 · similarity 0.86)
- A Non-parametric Direct Learning Approach to Heterogeneous Treatment Effect Estimation under Unmeasured Confounding (NeurIPS 2024 · similarity 0.86)
- Optimistic Algorithms for Adaptive Estimation of the Average Treatment Effect (ICML 2025 · similarity 0.86)
- Off-policy estimation with adaptively collected data: the power of online learning (NeurIPS 2024 · similarity 0.85)
- Higher-Order Causal Message Passing for Experimentation with Complex Interference (NeurIPS 2024 · similarity 0.85)
- Beyond the Average: Distributional Causal Inference under Imperfect Compliance (NeurIPS 2025 · 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.
