Research map / Reinforcement learning / Policy gradient and actor-critic
Accountable Off-Policy Evaluation With Kernel Bellman Statistics
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Increase the number of transitions | Kernel Bellman confidence bounds · Inverted-Pendulum | confidence interval width | Helped |
| Increase the failure probability | Kernel Bellman confidence bounds · Inverted-Pendulum | confidence interval width | Helped |
| Vary the behavior-policy temperature | Kernel Bellman confidence bounds · Inverted-Pendulum | confidence bounds | No clear effect |
| Use fewer features to constrain the Q-function to a lower-dimensional function space | Kernel Bellman confidence bounds · Inverted-Pendulum | confidence interval width | Helped |
| Use fewer features | Kernel Bellman confidence bounds · Inverted-Pendulum | risk of model misspecification | Hurt |
| Vary the number of features | Kernel Bellman confidence bounds · Puck-Mountain | confidence bounds | Mixed |
| Apply post-hoc diagnosis and correction to an existing Q-estimator that falls outside the confidence interval | Existing Q-estimators · Inverted-Pendulum | corrected estimate validity | Helped |
| Apply post-hoc diagnosis to an existing Q-estimator already inside the confidence interval | Existing Q-estimators · Inverted-Pendulum | debias-term norm | No clear effect |
| Increase the number of transitions | Post-hoc correction · Puck-Mountain | debias-term norm | Mixed |
Closest papers by meaning
- Non-asymptotic Confidence Intervals of Off-policy Evaluation: Primal and Dual Bounds (ICLR 2021 · similarity 0.91)
- Benchmarks for Deep Off-Policy Evaluation (ICLR 2021 · similarity 0.91)
- OPERA: Automatic Offline Policy Evaluation with Re-weighted Aggregates of Multiple Estimators (NeurIPS 2024 · similarity 0.89)
- Policy-Adaptive Estimator Selection for Off-Policy Evaluation (AAAI 2023 · similarity 0.88)
- Off-Policy Evaluation under Nonignorable Missing Data (ICML 2025 · similarity 0.88)
- Doubly-Robust Estimation of Counterfactual Policy Mean Embeddings (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.