Research map / Optimization / Nonconvex and smooth optimization
Communication Acceleration of Local Gradient Methods via an Accelerated Primal-Dual Algorithm with an Inexact Prox
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 FGD+GD instead of GD as the inexact-prox inner method (ablation) | APDA with Inexact Prox · Centralized federated optimization | local steps per round: \widetilde{\mathcal{O}}\left(\sqrt{\kappa}\right) -> \widetilde{\mathcal{O}}\left(\sqrt[3]{\kappa}\right) | Helped |
| Use FGD+FSFOM instead of GD as the inexact-prox inner method (ablation) | APDA with Inexact Prox · Centralized federated optimization | local steps per round: \widetilde{\mathcal{O}}\left(\sqrt{\kappa}\right) -> \widetilde{\mathcal{O}}\left(\sqrt[4]{\kappa}\right) | Helped |
| Use FGD+GD instead of GD as the inexact-prox inner method (ablation) | APDA with Inexact Prox · Centralized federated optimization | gradient computations: \widetilde{\mathcal{O}}\left(\kappa\right) -> \widetilde{\mathcal{O}}\left(\kappa^{\frac{5}{6}}\right) | Helped |
| Use FGD+FSFOM instead of GD as the inexact-prox inner method (ablation) | APDA with Inexact Prox · Centralized federated optimization | gradient computations: \widetilde{\mathcal{O}}\left(\kappa\right) -> \widetilde{\mathcal{O}}\left(\kappa^{\frac{3}{4}}\right) | Helped |
| Add accelerated gossip (ablation) | APDA with Inexact Prox · Decentralized optimization | communication rounds: \widetilde{\mathcal{O}}\left(\sqrt{\kappa\chi} \vee \chi\right) -> \widetilde{\mathcal{O}}\left(\sqrt{\kappa\chi}\right) | Helped |
| Use FGD+GD instead of GD as the inexact-prox inner method (ablation) | APDA with Inexact Prox and Accelerated Gossip · Decentralized optimization | local steps per round: \widetilde{\mathcal{O}}\left(\sqrt{\kappa}\right) -> \widetilde{\mathcal{O}}\left(\sqrt[3]{\kappa}\right) | Helped |
| Use FGD+FSFOM instead of GD as the inexact-prox inner method (ablation) | APDA with Inexact Prox and Accelerated Gossip · Decentralized optimization | local steps per round: \widetilde{\mathcal{O}}\left(\sqrt{\kappa}\right) -> \widetilde{\mathcal{O}}\left(\sqrt[4]{\kappa}\right) | Helped |
Closest papers by meaning
- ProxSkip: Yes! Local Gradient Steps Provably Lead to Communication Acceleration! Finally! (ICML 2022 · similarity 0.94)
- Variance Reduced ProxSkip: Algorithm, Theory and Application to Federated Learning (NeurIPS 2022 · similarity 0.90)
- Prox-PDA: The Proximal Primal-Dual Algorithm for Fast Distributed Nonconvex Optimization and Learning Over Networks (ICML 2017 · similarity 0.89)
- Stabilized Proximal-Point Methods for Federated Optimization (NeurIPS 2024 · similarity 0.89)
- Communication-Efficient Gradient Descent-Accent Methods for Distributed Variational Inequalities: Unified Analysis and Local Updates (ICLR 2024 · similarity 0.87)
- Towards Faster Decentralized Stochastic Optimization with Communication Compression (ICLR 2025 · 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.