Research map / Bandits and online learning / Multi-armed bandits
Identifying Best Interventions through Online Importance Sampling
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 SRISv2 with adaptively clipped importance sampling and information sharing across interventions | Best soft-intervention identification · Flow Cytometry | performance: more than 65% improvement | Helped |
| Use SRISv2 instead of SRISv1, reusing samples from earlier phases (ablation) | Best soft-intervention identification · Flow Cytometry | performance | Helped |
| Use SRISv2 rather than SR, which does not use information leakage | Best soft-intervention identification · Synthetic experiments, low divergence and high gap | simple regret | Helped |
| Use SRISv2 rather than CR with a static clipper | Best soft-intervention identification · Synthetic experiments, high divergence and high gap | simple regret | Helped |
| Replace adaptive clipping with a static clipper (ablation) | Clipped importance sampling · Synthetic experiments, high divergence and high gap | probability of error | Hurt |
| Use information leakage from other interventions rather than treating arms independently | Best soft-intervention identification · Synthetic experiments, low divergence and high gap | probability of error | Helped |
| Use SRISv2 rather than CR with a static clipper | Best soft-intervention identification · Synthetic experiments, low divergence and high gap | probability of error | No clear effect |
| Reuse samples from earlier phases, clipping them using the current phase criterion (ablation) | SRISv2 · Best soft-intervention identification | error and simple regret guarantees | No clear effect |
Closest papers by meaning
- Balancing Performance and Costs in Best Arm Identification (NeurIPS 2025 · similarity 0.88)
- Linear Causal Bandits: Unknown Graph and Soft Interventions (NeurIPS 2024 · similarity 0.88)
- Optimal Multi-Fidelity Best-Arm Identification (NeurIPS 2024 · similarity 0.86)
- Near Optimal Best Arm Identification for Clustered Bandits (ICML 2025 · similarity 0.86)
- Best Arm Identification in Generalized Linear Bandits via Hybrid Feedback (NeurIPS 2026 · similarity 0.86)
- Adaptive Online Experimental Design for Causal Discovery (ICML 2024 · 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.