Research map / Bandits and online learning / Online convex optimization and games
No-regret Algorithms for Capturing Events in Poisson Point Processes
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 |
|---|---|---|---|
| Adding RKHS structure (CAPTURE-UCB style estimators using the RKHS model) versus algorithms that do not exploit it (Thompson sampling of Grant et al. 2019) | adaptive sensing of Poisson processes · taxis, crime, environmental monitoring benchmarks | captured events / count-regret | Helped |
| Use Laplace-approximation (quadratic likelihood-ratio) confidence sets instead of the heteroscedastic regression confidence sets | UCB algorithm (UCB-LAP) · benchmarks with low-intensity regions | captured events / regret | Hurt |
| Increase confidence parameter beta to 4 and tweak past variances with new rate estimates (practical deviation from theory) (ablation) | CAPTURE-UCB · toy problem lambda(x)=4exp(-(x+1))sin(2xπ)^2 | finite-time performance | Helped |
| Use penalized likelihood estimator with the heteroscedastic regression confidence sets (UCB-L) instead of the Laplace-based sets | UCB-L · benchmarks | captured events | No clear effect |
| Switch feedback from count-record to histogram (or cost model from uniform to concave/one-off) | CAPTURE-UCB · benchmarks | performance | No clear effect |
Closest papers by meaning
- Online Change Point Detection for Multivariate Inhomogeneous Poisson Processes Time Series (ICML 2026 · similarity 0.86)
- Adaptive Sensor Placement for Continuous Spaces (ICML 2019 · similarity 0.86)
- Learning to Select Exogenous Events for Marked Temporal Point Process (NeurIPS 2021 · similarity 0.85)
- Markov-modulated Marked Poisson Processes for Check-in Data (ICML 2016 · similarity 0.85)
- Arbitrarily Conditioned Hierarchical Flows for Spatiotemporal Events (NeurIPS 2026 · similarity 0.83)
- State-of-art minibatches via novel DPP kernels: discretization, wavelets, and rough objectives (NeurIPS 2026 · similarity 0.83)
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.