Research map / Federated learning and privacy / Privacy attacks and secure learning
Byzantine-tolerant federated Gaussian process regression for streaming data
Xu Zhang, Zhenyuan Yuan, Minghui Zhu
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 Byzantine-tolerant PoE aggregation instead of standard PoE under same-value attack | Federated GPR · Synthetic, alpha=15% | MSE: 26.5339 ± 0.019 × 10−3 -> 0.0236 ± 0.172 × 10−3 | Improved |
| remove Byzantine attack (attack-free) vs Byzantine-tolerant PoE | Byzantine-tolerant PoE · Synthetic, 50 functions | MSE: 0.0049 ± 0.007 × 10−3 vs 0.0236 ± 0.172 × 10−3 | No clear effect |
| increase trimming ratio beta above alpha (ablation) | Byzantine-tolerant PoE · Synthetic | MSE | Hurt |
| decrease Byzantine ratio alpha (ablation) | Byzantine-tolerant PoE · Synthetic | MSE | Improved |
| add agent-based fused GPR to local GPR (agent 1, 6000 points) (ablation) | Local GPR · Synthetic, agent 1 | MSE: 0.0721 -> 0.0096 | Improved |
| add agent-based fused GPR to local GPR (agent 1, 10000 points) (ablation) | Local GPR · Synthetic, agent 1 | MSE: 0.0681 -> 0.0681 | No clear effect |
| add agent-based fused GPR to local GPR (agent 1, 150000 points) (ablation) | Local GPR · Synthetic, agent 1 | MSE: 0.0527 -> 0.0537 | Mixed |
| add agent-based fused GPR to local GPR (agent 6, 6000 points) (ablation) | Local GPR · Synthetic, agent 6 | MSE: 0.0727 -> 0.0727 | No clear effect |
| add agent-based fused GPR to local GPR (agent 6, 10000 points) (ablation) | Local GPR · Synthetic, agent 6 | MSE: 0.0566 -> 0.0086 | Improved |
| add agent-based fused GPR to local GPR (agent 6, 150000 points) (ablation) | Local GPR · Synthetic, agent 6 | MSE: 0.0555 -> 0.0076 | Improved |
| increase attack magnitude from 1 to 100 under standard PoE | Attacked standard PoE · SARCOS and Kin40k | prediction error | Hurt |
| increase attack magnitude from 1 to 100 under Byzantine-tolerant PoE | Byzantine-tolerant PoE · SARCOS and Kin40k | prediction error | No clear effect |
Closest papers by meaning
- Efficient Federated Learning against Byzantine Attacks and Data Heterogeneity via Aggregating Normalized Gradients (NeurIPS 2025 · similarity 0.87)
- Self-Driven Entropy Aggregation for Byzantine-Robust Heterogeneous Federated Learning (ICML 2024 · similarity 0.85)
- A Huber Loss Minimization Approach to Byzantine Robust Federated Learning (AAAI 2024 · similarity 0.85)
- From Inexact Gradients to Byzantine Robustness: Acceleration and Optimization under Similarity (NeurIPS 2026 · similarity 0.85)
- Fault Tolerant ML: Efficient Meta-Aggregation and Synchronous Training (ICML 2024 · similarity 0.85)
- Sharp Gaussian approximations for Decentralized Federated Learning (NeurIPS 2025 · 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.
