Research map / Federated learning and privacy / Differential privacy
A Joint Exponential Mechanism For Differentially Private Top-$k$
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 |
|---|---|---|---|
| Replace peeling (PNF-Peel pure DP, CDP-Peel approx DP) with a single joint exponential mechanism over length-k sequences | top-k selection · Books, Movies, News, Tweets | l_infty error | Helped |
| Joint exponential mechanism versus PNF-Peel on datasets with small count gaps | top-k selection · Foods, Games | l_infty error | Helped |
| Replace permute-and-flip with exponential mechanism in joint sampler (PNFJoint variant) (ablation) | Joint mechanism · low-k datasets | l_infty error | No clear effect |
| Joint sampler using ordered sorted t-values and k-way merging (O(dk log k + d log d) time) versus naive approach | Joint mechanism · Foods runtime | runtime | Helped |
| Joint mechanism versus CDP-Peel on l1 error with small datasets | top-k selection · small datasets, l1 error | l1 error | Mixed |
| Joint mechanism on k-relative error (small-scale datasets) | top-k selection · k-relative error | k-relative error | Mixed |
Closest papers by meaning
- Faster Differentially Private Top-$k$ Selection: A Joint Exponential Mechanism with Pruning (NeurIPS 2024 · similarity 0.93)
- Practical Differentially Private Top-k Selection with Pay-what-you-get Composition (NeurIPS 2019 · similarity 0.91)
- Privately Counting Partially Ordered Data (ICLR 2025 · similarity 0.90)
- Approximate Differential Privacy of the $\ell_2$ Mechanism (ICML 2025 · similarity 0.88)
- Differentially Private Set Representations (NeurIPS 2024 · similarity 0.87)
- Sort, Partition, Randomize: Optimal Binary Hypothesis Testing under Local Differential Privacy (NeurIPS 2026 · 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.