Research map / Reinforcement learning / Policy gradient and actor-critic
Off-Policy Evaluation for Large Action Spaces via Embeddings
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 action-space size from 10 to 5000 | IPS · Synthetic OPE | variance and MSE: over 300 | Hurt |
| Use marginalized importance weights over action embeddings instead of action-level weights | IPS and DR OPE · Synthetic OPE, |A|=10 to 5000 | MSE: 1.38; 12.38 | Helped |
| Use marginalized importance weights over action embeddings instead of action-level weights | IPS and DR OPE · Synthetic OPE, |A|=1000 | MSE: 9.10; 4.87 | Helped |
| Use true marginal importance weights rather than estimated marginal weights (ablation) | MIPS · Synthetic OPE with large action sets | MSE | Helped |
| Remove some action-embedding dimensions (ablation) | MIPS · Synthetic OPE, 20 embedding dimensions | MSE | Mixed |
| Select embedding dimensions with SLOPE (ablation) | MIPS · Synthetic OPE, |A|=1000 | MSE | Helped |
| Select embedding dimensions with SLOPE (ablation) | MIPS · Open Bandit Dataset, ALL campaign | fraction of bootstrap runs with lower squared error than IPS: about 80% | Helped |
| Use MIPS without action-embedding selection (ablation) | MIPS · Open Bandit Dataset, ALL campaign | squared error relative to IPS | No clear effect |
Closest papers by meaning
- Off-Policy Learning in Large Action Spaces: Optimization Matters More Than Estimation (ICML 2026 · similarity 0.91)
- Cross-Domain Off-Policy Evaluation and Learning for Contextual Bandits (ICLR 2025 · similarity 0.89)
- Log-Sum-Exponential Estimator for Off-Policy Evaluation and Learning (ICML 2025 · similarity 0.88)
- OPERA: Automatic Offline Policy Evaluation with Re-weighted Aggregates of Multiple Estimators (NeurIPS 2024 · similarity 0.87)
- Distributional Off-Policy Evaluation for Slate Recommendations (AAAI 2024 · similarity 0.87)
- Kernel Metric Learning for In-Sample Off-Policy Evaluation of Deterministic RL Policies (ICLR 2024 · similarity 0.87)
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.