Research map / Federated learning and privacy / Differential privacy
Locally Private Causal Inference for Randomized Experiments
Yuki Ohnishi, Jordan Awan
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 Custom (IPW) privacy scenario instead of Joint | frequentist estimator · Simulation eps_tot=0.3 | MSE: 0.7875 -> 0.0091 | Improved |
| use Custom (IPW) privacy scenario instead of Joint | frequentist estimator · Simulation eps_tot=0.1 | interval width: 1.889 -> 1.091 | Improved |
| use Custom (IPW) instead of Joint | frequentist estimator · Simulation eps_tot=10.0 | MSE: 0.0001 -> 0.0001 | No clear effect |
| use Custom (DM) instead of Joint | frequentist estimator · Simulation eps_tot=0.1 | coverage: 94.55% -> 99.8% | Mixed |
| use Bayesian estimator instead of frequentist | Joint scenario · Simulation eps_tot=0.1 | MSE: 0.9872 -> 0.0099 | Improved |
| use Bayesian estimator instead of frequentist | Custom (IPW) · Simulation eps_tot=0.1 | interval width: 1.091 -> 0.319 | Improved |
| use Bayesian estimator instead of frequentist | Custom (DM) · Simulation eps_tot=0.1 | MSE: 0.7608 -> 0.0099 | Improved |
| use Bayesian estimator instead of frequentist | Custom (IPW) · Simulation eps_tot=10.0 | MSE: 0.0001 -> 0.0 | No clear effect |
| increase total privacy budget from 0.1 to 10.0 | frequentist Joint estimator · Simulation N=10000 | MSE: 0.9872 -> 0.0001 | Improved |
| increase total privacy budget from 0.1 to 10.0 | frequentist Custom (IPW) · Simulation N=10000 | MSE: 0.0803 -> 0.0001 | Improved |
| use Custom (IPW) private scenario with Bayes at eps_tot=0.1 | Bayesian estimator · Colombia cash transfer | interval 2.5%-97.5%: -1.0 to 1.0 -> -0.135 to 0.244 | Improved |
| use Bayesian instead of frequentist Joint | Joint scenario · Colombia cash transfer eps_tot=1.0 | interval 2.5%-97.5%: -0.898 to 0.559 -> -0.022 to 0.111 | Improved |
Closest papers by meaning
- Improving the Variance of Differentially Private Randomized Experiments through Clustering (ICML 2025 · similarity 0.92)
- Privacy Guarantees in Posterior Sampling under Contamination (NeurIPS 2026 · similarity 0.90)
- Locally Private Bayesian Inference for Count Models (ICML 2019 · similarity 0.89)
- Comparing Population Means Under Local Differential Privacy: With Significance and Power (AAAI 2018 · similarity 0.89)
- A Simple and Practical Method for Reducing the Disparate Impact of Differential Privacy (AAAI 2024 · similarity 0.88)
- Locally Optimal Private Sampling: Beyond the Global Minimax (NeurIPS 2025 · 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.
