Research map / Federated learning and privacy / Differential privacy
Unbounded Differentially Private Quantile and Maximum Estimation
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 |
|---|---|---|---|
| Use UQE instead of EMQ for quantile-based clipping in private sum estimation | Private sum estimation · Ratings data, ε = 1 | mean absolute error: 5.73 -> 4.78 | Helped |
| Use UQE instead of EMQ for quantile-based clipping in private sum estimation | Private sum estimation · Pages data, ε = 1 | mean absolute error: 4324.38 -> 4385.23 | No clear effect |
| Use UQE instead of EMQ for quantile-based clipping in private sum estimation | Private sum estimation · Ages data, ε = 1 | mean absolute error: 187.06 -> 103.05 | Helped |
| Use UQE instead of EMQ for quantile-based clipping in private sum estimation | Private sum estimation · Work hours data, ε = 1 | mean absolute error: ${339.06}_{75.82}$ -> $180.48_{44.92}$ | Helped |
| Use UQE instead of AT for quantile-based clipping in private sum estimation | Private sum estimation · Ratings data, ε = 1 | mean absolute error: ${8.75}_{0.31}$ -> $4.78_{0.21}$ | Helped |
| Use UQE instead of AT for quantile-based clipping in private sum estimation | Private sum estimation · Pages data, ε = 1 | mean absolute error: $4377.45_{2340.93}$ -> $4385.23_{2077.16}$ | No clear effect |
| Use UQE instead of AT for quantile-based clipping in private sum estimation | Private sum estimation · Ages data, ε = 1 | mean absolute error: ${293.11}_{117.32}$ -> $103.05_{16.04}$ | Helped |
| Use UQE instead of AT for quantile-based clipping in private sum estimation | Private sum estimation · Work hours data, ε = 1 | mean absolute error: $471.98_{174.07}$ -> $180.48_{44.92}$ | Helped |
| Use UQE instead of EMQ for quantile-based clipping | Private sum estimation · Ratings data, ε = 0.5 | mean absolute error: $6906.13_{7123.36}$ -> $9.22_{0.31}$ | Helped |
| Use UQE instead of EMQ for quantile-based clipping | Private sum estimation · Pages data, ε = 0.5 | mean absolute error: $7601.47_{3963.52}$ -> $7102.34_{3093.13}$ | Helped |
| Use UQE instead of EMQ for quantile-based clipping | Private sum estimation · Ages data, ε = 0.1 | mean absolute error: $50558.32_{28843.72}$ -> $821.77_{157.68}$ | Helped |
| Use UQE instead of EMQ for quantile estimation | Quantile estimation · Synthetic datasets | mean absolute error | Hurt |
Closest papers by meaning
- Differentially Private Quantiles (ICML 2021 · similarity 0.92)
- Differentially Private Quantiles with Smaller Error (NeurIPS 2025 · similarity 0.91)
- Lightweight Protocols for Distributed Private Quantile Estimation (ICML 2025 · similarity 0.91)
- Archimedes Meets Privacy: On Privately Estimating Quantiles in High Dimensions Under Minimal Assumptions (NeurIPS 2022 · similarity 0.90)
- Online Local Differential Private Quantile Inference via Self-normalization (ICML 2023 · similarity 0.90)
- Differentially Private Approximate Quantiles (ICML 2022 · similarity 0.90)
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.