Research map / Probabilistic and scientific modeling / Treatment effects
Marginal Density Ratio for Off-Policy Evaluation in Contextual Bandits
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 |
|---|---|---|---|
| Use MR instead of the strongest listed baseline | contextual-bandit OPE · Digits | MSE: 0.0847$\pm$0.0025 -> 0.0034$\pm$0.0001 | Helped |
| Use MR instead of the strongest listed baseline | contextual-bandit OPE · Letter | MSE: 0.0886$\pm$0.0026 -> 0.0018$\pm$0.0004 | Helped |
| Use MR instead of the strongest listed baseline | contextual-bandit OPE · OptDigits | MSE: 0.0384$\pm$0.0025 -> 0.0006$\pm$0.0002 | Helped |
| Use MR instead of the strongest listed baseline | contextual-bandit OPE · PenDigits | MSE: 0.0138$\pm$0.0029 -> 0.0008$\pm$0.0002 | Helped |
| Use MR instead of the strongest listed baseline | contextual-bandit OPE · SatImage | MSE: 0.0078$\pm$0.0008 -> 0.0016$\pm$0.0003 | Helped |
| Use MR instead of the strongest listed baseline | contextual-bandit OPE · Mnist | MSE: 0.1109$\pm$0.0014 -> 0.0121$\pm$0.0009 | Helped |
| Use MR instead of the strongest listed baseline | contextual-bandit OPE · CIFAR-100 | MSE: 0.0020$\pm$0.0001 -> 0.0007$\pm$0.0002 | Helped |
| Use MR instead of DM | ATE estimation · Twins (n=50) | absolute ATE error: 0.092$\pm$0.003 -> 0.062$\pm$0.007 | Helped |
| Use MR instead of DR | ATE estimation · Twins (n=200) | absolute ATE error: 0.065$\pm$0.009 -> 0.065$\pm$0.007 | No clear effect |
| Use MR instead of IPW | ATE estimation · Twins (n=1600) | absolute ATE error: 0.067$\pm$0.007 -> 0.061$\pm$0.005 | Helped |
| Use MR instead of IPW | ATE estimation · Twins (n=3200) | absolute ATE error: 0.067$\pm$0.007 -> 0.061$\pm$0.006 | Helped |
| Increase policy shift by increasing $\alpha^*$ (ablation) | MR · synthetic contextual-bandit data | MSE | Mixed |
Closest papers by meaning
- Optimal and Adaptive Off-policy Evaluation in Contextual Bandits (ICML 2017 · similarity 0.93)
- Cross-Domain Off-Policy Evaluation and Learning for Contextual Bandits (ICLR 2025 · similarity 0.90)
- Log-Sum-Exponential Estimator for Off-Policy Evaluation and Learning (ICML 2025 · similarity 0.90)
- Logarithmic Smoothing for Pessimistic Off-Policy Evaluation, Selection and Learning (NeurIPS 2024 · similarity 0.89)
- Off-policy estimation with adaptively collected data: the power of online learning (NeurIPS 2024 · similarity 0.88)
- Demystifying the Paradox of Importance Sampling with an Estimated History-Dependent Behavior Policy in Off-Policy Evaluation (ICML 2025 · similarity 0.88)
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.