Research map / Optimization / Distributed and compressed
Collaborative Deep Learning in Fixed Topology Networks
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 decentralized consensus-based momentum SGD instead of centralized-parameter-server FedAvg | CDMSGD · CIFAR-10 | validation accuracy | Helped |
| Use decentralized consensus-based momentum SGD instead of centralized-parameter-server FedAvg | CDMSGD · CIFAR-100 | validation accuracy | Helped |
| Use decentralized consensus-based momentum SGD instead of centralized-parameter-server FedAvg | CDMSGD · CIFAR-10 | convergence speed | Hurt |
| Use decentralized consensus-based SGD instead of centralized SGD | CDSGD · CIFAR-10 | convergence speed | Hurt |
| Use decentralized consensus-based SGD instead of centralized SGD | CDSGD · CIFAR-10 | validation accuracy | Helped |
| Increase the number of agents from 2 to 8 to 16 (ablation) | CDMSGD · CIFAR-10 | convergence speed | Hurt |
| Increase the number of agents from 2 to 8 to 16 (ablation) | CDMSGD · CIFAR-10 | accuracy | No clear effect |
| Use a sparser network topology (ablation) | CDMSGD · CIFAR-10 | convergence speed | Helped |
| Use a denser network topology (ablation) | CDMSGD · CIFAR-10 | consensus stability | Helped |
| Use a diminishing step size rather than a fixed step size (ablation) | CDSGD · Strongly convex objective | objective convergence | Helped |
Closest papers by meaning
- DSGD-CECA: Decentralized SGD with Communication-Optimal Exact Consensus Algorithm (ICML 2023 · similarity 0.89)
- Topology-aware Generalization of Decentralized SGD (ICML 2022 · similarity 0.88)
- Consensus Control for Decentralized Deep Learning (ICML 2021 · similarity 0.88)
- Exponential Graph is Provably Efficient for Decentralized Deep Training (NeurIPS 2021 · similarity 0.87)
- Scalable Decentralized Learning with Teleportation (ICLR 2025 · similarity 0.86)
- A Quadratic Synchronization Rule for Distributed Deep Learning (ICLR 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.