Research map / Federated learning and privacy / Differential privacy
Differentially Private Learning of Undirected Graphical Models Using Collective Graphical Models
Garrett Bernstein, Ryan McKenna, Tao Sun, Daniel Sheldon, Michael Hay, Gerome Miklau
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 |
|---|---|---|---|
| Use CGM inference and EM instead of Naive MLE on noisy sufficient statistics | Differentially private graphical-model learning · Synthetic third-order chains and Erdős–Rényi graphs | KL divergence | Improved |
| Use CGM inference and EM instead of Naive MLE on noisy sufficient statistics | Differentially private graphical-model learning · Wifi data, time-homogeneous chain | holdout log-likelihood | Improved |
| Use CGM instead of PSGD | Differentially private graphical-model learning · Synthetic third-order chains and Erdős–Rényi graphs | KL divergence | Improved |
| Use CGM instead of PSGD | Differentially private graphical-model learning · Wifi data, time-homogeneous chain | holdout log-likelihood | Improved |
| Project each noisy clique pseudo-marginal onto the simplex before MLE (ablation) | Naive MLE with noisy sufficient statistics · Synthetic Erdős–Rényi graph | KL divergence | Improved |
| Vary regularization strength when learning from noisy sufficient statistics (ablation) | Naive MLE · Synthetic chain | KL divergence | Mixed |
| Increase regularization strength (ablation) | Naive MLE with noisy sufficient statistics · Synthetic chain | KL divergence | Mixed |
Closest papers by meaning
- Private Adaptive Covariance Estimation via Gaussian Graphical Models (NeurIPS 2026 · similarity 0.87)
- Differentially Private Sum-Product Networks (ICML 2024 · similarity 0.87)
- Certification for Differentially Private Prediction in Gradient-Based Training (ICML 2025 · similarity 0.87)
- It's My Data Too: Private ML for Datasets with Multi-User Training Examples (ICML 2025 · similarity 0.87)
- Exactly Minimax-Optimal Locally Differentially Private Sampling (NeurIPS 2024 · similarity 0.87)
- Optimal Differentially Private Model Training with Public Data (ICML 2024 · similarity 0.87)
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.
