Research map / Federated learning and privacy / Differential privacy
Local Differential Privacy for Evolving Data
Matthew Joseph, Aaron Roth, Jonathan Ullman, Bo Waggoner
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 Thresh to update the estimate only when private votes indicate it may be inaccurate, rather than recomputing every epoch | Local tracking of evolving Bernoulli means · Evolving Bernoulli means | dependence of privacy cost or error on the number of epochs: error ... scales sublinearly in 1/ℓ and only polylogarithmically with the total number of epochs T; [naive] scales linea | Improved |
| Use HeavyThresh with hashed histogram estimates to track distribution changes | Local heavy-hitter tracking · Evolving heavy hitters | frequency-oracle error dependence on epochs and distribution changes: accuracy guarantees that degrade with the number of times the distribution ... changes, and only polylogarithmically wit | Improved |
| Batch rounds into epochs and adjust epoch length | Thresh and HeavyThresh · Evolving data | tradeoff between report frequency, privacy, and accuracy | Mixed |
Closest papers by meaning
- Setting $\varepsilon$ is not the Issue in Differential Privacy (NeurIPS 2025 · similarity 0.80)
- Monitoring Violations of Differential Privacy over Time (NeurIPS 2026 · similarity 0.80)
- A Simple and Practical Method for Reducing the Disparate Impact of Differential Privacy (AAAI 2024 · similarity 0.79)
- Fairness Issues and Mitigations in (Differentially Private) Socio-Demographic Data Processes (AAAI 2025 · similarity 0.78)
- Privacy Guarantees in Posterior Sampling under Contamination (NeurIPS 2026 · similarity 0.78)
- Position: Considerations for Differentially Private Learning with Large-Scale Public Pretraining (ICML 2024 · similarity 0.77)
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.
