Research map / Federated learning and privacy / Differential privacy
Network change point localisation under local differential privacy
Mengchu Li, Thomas Berrett, Yi Yu
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 |
|---|---|---|---|
| Apply edge randomized response before NBS | Dynamic IBN change point localisation · Dynamic IBN, edge LDP | signal-to-noise ratio condition: \kappa_0^2 \rho^2 n \Delta \alpha^2 \geq c_0 \log^{2+\xi}(T); infeasibility when \kappa_0^2 \rho^2 n \Delta \alpha^2 \le | Improved |
| Apply node LDP to bipartite network rows | Dynamic bipartite IBN change point localisation · Dynamic bipartite IBN, node LDP | signal-to-noise ratio condition: \kappa_0^2 \rho^2 n_1^{1/2} \Delta \alpha^2 \lesssim 1 (lower bound); upper-bound condition: \kappa^2_0 \rho^2 \min \{\s | Mixed |
| Impose edge LDP rather than use raw data | Dynamic IBN change point localisation · Dynamic IBN | infeasibility boundary: \kappa_0^2 \rho n \Delta \lesssim 1 -> \kappa_0^2 \rho^2 n \Delta \alpha^2 \lesssim 1 | Hurt |
| Protect rows with node LDP rather than edges with edge LDP | Dynamic network change point localisation · Bipartite IBN with n1 = n2 = n | infeasibility boundary: edge: \kappa_0^2 \rho^2 n \Delta \alpha^2 \lesssim 1; node: \kappa_0^2 \rho^2 n^{1/2} \Delta \alpha^2 \lesssim 1 | Hurt |
| Decrease the privacy budget α under edge LDP | NBS with edge randomized response · Dynamic IBN simulations | scaled Hausdorff distance | Hurt |
| Use node LDP instead of edge LDP | NBS in dynamic network simulations · Dynamic bipartite IBN simulations | scaled Hausdorff distance | Hurt |
| Use raw data instead of privatized data | Simplified binary segmentation · Dynamic IBN simulations | minimum spacing for detection: 7 | Improved |
Closest papers by meaning
- Minimax Private Estimation of Smooth Optimal-Transport Maps (NeurIPS 2026 · similarity 0.85)
- Locally Optimal Private Sampling: Beyond the Global Minimax (NeurIPS 2025 · similarity 0.84)
- Minimax Optimal Two-Sample Testing under Local Differential Privacy (NeurIPS 2026 · similarity 0.84)
- Convergent Privacy Loss of Noisy-SGD without Convexity and Smoothness (ICLR 2025 · similarity 0.83)
- The Privacy Power of Correlated Noise in Decentralized Learning (ICML 2024 · similarity 0.83)
- Shifted Interpolation for Differential Privacy (ICML 2024 · 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.
