Research map / Federated learning and privacy / Communication-efficient FL
A Unified Convergence Analysis for Semi-Decentralized Learning: Sampled-to-Sampled vs. Sampled-to-All Communication
Angelo Rodio, Giovanni Neglia, Zheng Chen, Erik G. Larsson
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 S2S instead of S2A when inter-component data is non-IID (K/n=0.2) | semi-decentralized FL · CIFAR-10 inter non-IID | accuracy: +7 p.p. | Improved |
| use S2S instead of S2A when inter-component data is non-IID (K/n=0.2) | semi-decentralized FL · MNIST inter non-IID | accuracy: +2.4 p.p. | Improved |
| use S2S instead of S2A with intra non-IID, inter IID, ring, K/n=0.2 | semi-decentralized FL · CIFAR-10 ring | accuracy: +8.4 p.p. | Improved |
| use S2A instead of S2S with intra non-IID, inter IID at higher K/n or H | semi-decentralized FL · CIFAR-10 ring | accuracy: +1.2 p.p. | Improved |
| use S2A instead of S2S when both intra and inter data are IID | semi-decentralized FL · MNIST/CIFAR-10 intra IID inter IID | accuracy: up to 1 p.p. | Improved |
| use S2S instead of S2A with intra non-IID, short server period (H=5, ring) | semi-decentralized FL · ring intra non-IID inter IID | accuracy: +8.5 p.p. | Improved |
| use S2A instead of S2S with intra non-IID, H=10, complete graph | semi-decentralized FL · complete graph intra non-IID inter IID | accuracy: +4 p.p. | Improved |
| increase sampling rate K/n from 0.2 to 1 | S2S and S2A · MNIST/CIFAR-10 | accuracy: +2 p.p. | Improved |
| increase server period H from 5 to 20 | S2S and S2A · MNIST/CIFAR-10 | accuracy: -2.4 p.p. | Hurt |
| use S2A broadcast to all devices in inter non-IID regime | S2A · MNIST/CIFAR-10 inter non-IID | test accuracy over rounds: early-round accuracy higher, later periodic drops | Mixed |
Closest papers by meaning
- Decentralized Sporadic Federated Learning: A Unified Algorithmic Framework with Convergence Guarantees (ICLR 2025 · similarity 0.89)
- SemiDFL: A Semi-Supervised Paradigm for Decentralized Federated Learning (AAAI 2025 · similarity 0.89)
- Federated Learning under Periodic Client Participation and Heterogeneous Data: A New Communication-Efficient Algorithm and Analysis (NeurIPS 2024 · similarity 0.87)
- Debiasing Federated Learning with Correlated Client Participation (ICLR 2025 · similarity 0.87)
- TRAIL: Trust-Aware Client Scheduling for Semi-Decentralized Federated Learning (AAAI 2025 · similarity 0.86)
- Heterogeneity-Guided Client Sampling: Towards Fast and Efficient Non-IID Federated Learning (NeurIPS 2024 · similarity 0.86)
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.
