Research map / Federated learning and privacy / Personalized and heterogeneous FL
Learning privately from multiparty data
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 soft-label weighted ERM (class-probability-weighted loss) instead of majority-vote labels for auxiliary data | private multiparty ensemble · Activity recognition (UCI HAR) | test accuracy: vote 0.79 vs soft 0.76 at 1/eps=0, but vote becomes useless at 1/eps=0.1 while soft stays above indiv until 1/eps=1 | Helped |
| Use soft-label weighted ERM instead of majority-vote labels under output perturbation | private multiparty ensemble · Network intrusion detection and malicious URL | test accuracy under privacy | Helped |
| Increase number of parties M (5K to 20K) for soft-label method | private multiparty ensemble · Network intrusion detection (KDD-99) | test accuracy under privacy | Helped |
| Parameter averaging instead of ensemble-based soft labels | private multiparty aggregation · Activity recognition (UCI HAR) | test accuracy at 1/eps=0: 0.67 vs 0.76 (soft) | Hurt |
| Use a larger number of parties with few samples each (Malicious URL, 9 samples/party) vs fewer samples | private ensemble · Malicious URL detection | gap between batch and private methods | Hurt |
Closest papers by meaning
- Learning from End User Data with Shuffled Differential Privacy over Kernel Densities (ICLR 2025 · similarity 0.87)
- Oracle-Efficient Differentially Private Learning with Public Data (NeurIPS 2024 · similarity 0.85)
- Differential Privacy Under Class Imbalance: Methods and Empirical Insights (ICML 2025 · similarity 0.84)
- Private Truly-Everlasting Robust-Prediction (ICML 2024 · similarity 0.84)
- Certification for Differentially Private Prediction in Gradient-Based Training (ICML 2025 · similarity 0.84)
- Differentially Private Sum-Product Networks (ICML 2024 · similarity 0.84)
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.