Research map / Federated learning and privacy / Differential privacy
Single and Multiple Change-Point Detection with Differential Privacy
Wanrong Zhang, Sara Krehbiel, Rui Tuo, Yajun Mei, Rachel Cummings
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| weaken privacy by increasing epsilon (0.1 to 0.5, 1, infinity) (ablation) | OfflinePCPD / OfflinePTCPD · Offline Monte Carlo, Bernoulli/Gaussian/Gamma, n=200, k*=100 | Pr[|k-hat - k*| > alpha] | Improved |
| increase distance between P0 and P1 (large change vs small change) (ablation) | OfflinePCPD / OfflinePTCPD · Offline Monte Carlo, n=200, k*=100 | Pr[|k-hat - k*| > alpha] | Improved |
| supply underspecified hypotheses (small-change values) when true change is large (ablation) | OfflinePCPD / OfflinePTCPD · Offline Monte Carlo, underspecified change | Pr[|k-hat - k*| > alpha] | No clear effect |
| use a small truncation parameter A instead of a large one (ablation) | OfflinePTCPD · Offline Monte Carlo, Gaussian/Gamma, A from 0 to 4 | Pr[|k-hat - k*| > alpha] | Improved |
| use online sliding-window algorithm with window n=700 and numerically chosen threshold T | OnlinePCPD · Online Monte Carlo, Bernoulli/Gaussian large mean change, k*=5000 | probability of inaccurate estimate | Improved |
Closest papers by meaning
- Differentially Private Change-Point Detection (NeurIPS 2018 · similarity 0.99)
- Privately detecting changes in unknown distributions (ICML 2020 · similarity 0.96)
- Monitoring Violations of Differential Privacy over Time (NeurIPS 2026 · similarity 0.85)
- Making Old Things New: A Unified Algorithm for Differentially Private Clustering (ICML 2024 · similarity 0.85)
- Mean Estimation in the Add-Remove Model of Differential Privacy (ICML 2024 · similarity 0.85)
- Privacy Guarantees in Posterior Sampling under Contamination (NeurIPS 2026 · similarity 0.84)
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.
