Research map / Federated learning and privacy / Privacy attacks and secure learning
Robust Federated Inference
Akash Dhasade, Sadegh Farhadkhani, Rachid Guerraoui, Nirupam Gupta, Maxime Jacovella, Anne-Marie Kermarrec, Rafael Pinot
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 adversarially trained DeepSet with test-time trimmed mean (DeepSet-TM) instead of best static robust aggregator | DeepSet-TM · CIFAR-10, CIFAR-100, AG-News | worst-case accuracy: +4.7 to +22.2 points | Improved |
| use DeepSet-TM instead of static aggregation under PGD-cw | DeepSet-TM · AG-News | accuracy: 55.3 -> 83.2 | Improved |
| use DeepSet-TM instead of CWTM under SIA white-box | DeepSet-TM · CIFAR-10 | accuracy: 44.8 -> 51.4 | Improved |
| use DeepSet-TM instead of Mean under logit flipping | DeepSet-TM · CIFAR-100 | accuracy: 78.8 -> 78.0 | No clear effect |
| scale clients to n=25, f=5 | DeepSet-TM · CIFAR-100 (alpha=0.3) | worst-case accuracy: 30.4 -> 50.7 (vs CWTM) | Improved |
| use DeepSet-TM instead of CoPur under logit flipping with larger amplification or more adversaries | DeepSet-TM · CIFAR-10 | accuracy | Improved |
| add adversarial training to DeepSet aggregator (ablation) | DeepSet-TM · no-adversary case (f=0) | clean accuracy | Hurt |
| use CWMed instead of Mean | CWMed · CIFAR-100 (alpha=0.5) | accuracy under logit flipping: 78.8 -> 65.8 | Hurt |
Closest papers by meaning
- Byzantine-Robust Federated Learning: Impact of Client Subsampling and Local Updates (ICML 2024 · similarity 0.88)
- Federated Generalised Variational Inference: A Robust Probabilistic Federated Learning Framework (ICML 2025 · similarity 0.87)
- Adaptive Gradient Clipping for Robust Federated Learning (ICLR 2025 · similarity 0.87)
- FedLPA: One-shot Federated Learning with Layer-Wise Posterior Aggregation (NeurIPS 2024 · similarity 0.86)
- Certifiably Robust Model Evaluation in Federated Learning under Meta-Distributional Shifts (ICML 2025 · similarity 0.86)
- Efficient Federated Learning against Byzantine Attacks and Data Heterogeneity via Aggregating Normalized Gradients (NeurIPS 2025 · similarity 0.85)
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.
