Research map / Probabilistic and scientific modeling / Molecules and proteins
Sequential Changepoint Detection via Backward Confidence Sequences
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 |
|---|---|---|---|
| Add backward confidence sequences to the forward confidence sequence detector | SCD · No-change stream | average run length (ARL): \mathbb{E}_{\infty}[\tau] \geq \frac{1}{2\alpha} - \frac{3}{2}; FCS-Detector has infinite ARL | Helped |
| Add backward confidence sequences to the forward confidence sequence detector | SCD · After a change | expected detection delay: FCS delay increases approximately as \sqrt{T}; BCS bound is independent of T | Helped |
| Use a single forward confidence sequence instead of forward and backward confidence sequences (ablation) | BCS-Detector · No-change stream | average run length (ARL): infinite | No clear effect |
| Use a single forward confidence sequence instead of forward and backward confidence sequences (ablation) | BCS-Detector · After a change | detection delay: approximately \sqrt{T} | Hurt |
| Use Gaussian mean confidence sequences | BCS-Detector · Gaussian mean change detection | expected detection delay: \mc{O}\left( \frac{ \log \log(\frac{1}{\dkl(P_{\theta_1}, P_{\theta_0})}) + \log(\frac{1}{\alpha})}{\dkl(P_{\theta_1}, P | Helped |
| Use empirical Bernstein confidence sequences | BCS-Detector · Bounded nonparametric mean change detection | expected detection delay: \mc{O}\left( \sigma_1^2\, \frac{\log(1/\Delta) + \log(1/\alpha)}{\Delta^2} \right) | Helped |
| Use a Kolmogorov-Smirnov confidence sequence | BCS-Detector · CDF change detection | expected detection delay: \mc{O} \left( \frac{ \log \log(1/\Delta) + \log(1/\alpha)}{\Delta^2}\right) | Helped |
| Use kernel-MMD confidence sequences | BCS-Detector · Change in homogeneity of two streams | expected detection delay: \mc{O} \left( \frac{ \log(1/alpha) + \log \log(1/\Delta)}{\Delta^2} \right) | Helped |
| Vary the change magnitude Δ | BCS-Detector · Gaussian mean change detection | average detection delay: 1/\Delta^2 | Mixed |
Closest papers by meaning
- Reducing sequential change detection to sequential estimation (ICML 2024 · similarity 0.93)
- Fixed-Confidence Multiple Change Point Identification under Bandit Feedback (ICML 2025 · similarity 0.85)
- The Sample Complexity of Multiple Change Point Identification under Bandit Feedback (NeurIPS 2026 · similarity 0.85)
- Optimal Online Change Detection via Random Fourier Features (NeurIPS 2025 · similarity 0.85)
- Conjugate Bayesian Two-step Change Point Detection for Hawkes Process (NeurIPS 2024 · similarity 0.84)
- Adaptive Estimation and Learning under Temporal Distribution Shift (ICML 2025 · 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.