Research map / Federated learning and privacy / Communication-efficient FL
Tackling the Objective Inconsistency Problem in Heterogeneous Federated Optimization
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 FedNova normalized aggregation instead of FedAvg with vanilla local SGD | VGG-11 · CIFAR-10 test, E_i=2 | test accuracy (%): 60.68 ± 1.05 -> 66.31 ± 0.86 | Helped |
| Use FedNova normalized aggregation instead of FedAvg with local momentum | VGG-11 · CIFAR-10 test, E_i=2 | test accuracy (%): 65.26 ± 2.42 -> 73.32 ± 0.29 | Helped |
| Use FedNova normalized aggregation instead of FedAvg with proximal local SGD | VGG-11 · CIFAR-10 test, E_i=2 | test accuracy (%): 60.44 ± 1.21 -> 69.92 ± 0.34 | Helped |
| Use FedNova normalized aggregation instead of FedAvg with vanilla local SGD | VGG-11 · CIFAR-10 test, E_i^(t) ~ U(2,5) | test accuracy (%): 64.22 ± 1.06 -> 73.22 ± 0.32 | Helped |
| Use FedNova normalized aggregation instead of FedAvg with local momentum | VGG-11 · CIFAR-10 test, E_i^(t) ~ U(2,5) | test accuracy (%): 70.44 ± 2.99 -> 77.07 ± 0.12 | Helped |
| Use FedNova normalized aggregation instead of FedAvg with proximal local SGD | VGG-11 · CIFAR-10 test, E_i^(t) ~ U(2,5) | test accuracy (%): 63.74 ± 1.44 -> 73.41 ± 0.45 | Helped |
| Use FedNova normalized aggregation with variance reduction instead of FedAvg with variance reduction | VGG-11 · CIFAR-10 test, E_i^(t) ~ U(2,5) | test accuracy (%): 74.72 ± 0.34 -> 74.72 ± 0.19 | No clear effect |
| Combine local momentum and variance reduction | FedNova · CIFAR-10 test, E_i^(t) ~ U(2,5) | test accuracy (%) | Helped |
| Use local momentum instead of vanilla SGD | FedNova · CIFAR-10 test | test accuracy (%): 3-7% | Helped |
| Combine local momentum with server momentum instead of using server momentum alone | Federated training · CIFAR-10 test, E_i(t) ~ U(2,5) | test accuracy (%): 77.49 ± 0.25% -> 81.15 ± 0.38% | Helped |
Closest papers by meaning
- A New Theoretical Perspective on Data Heterogeneity in Federated Optimization (ICML 2024 · similarity 0.91)
- FedRSPO+: A Heterogeneity-aware Algorithm for Decision-focused Federated Learning (NeurIPS 2026 · similarity 0.90)
- Exact and Linear Convergence for Federated Learning under Arbitrary Client Participation is Attainable (NeurIPS 2025 · similarity 0.89)
- FedGPS: Statistical Rectification Against Data Heterogeneity in Federated Learning (NeurIPS 2025 · similarity 0.89)
- Decentralized Sporadic Federated Learning: A Unified Algorithmic Framework with Convergence Guarantees (ICLR 2025 · similarity 0.88)
- Federated Optimization with Doubly Regularized Drift Correction (ICML 2024 · similarity 0.88)
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.