Research map / Optimization / Distributed and compressed
On the Convergence of Decentralized Stochastic Minimax Optimization Algorithm with Compressed Communication
Yihan Zhang, Xinghua Shi, Meikang Qiu, Yu Wang, Hongchang Gao
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 |
|---|---|---|---|
| add error-feedback Top-10% compression to decentralized minimax SGDA with momentum | CE-DSGDAR · AUC maximization (CIFAR10, CIFAR100, CATvsDOG, STL10) | test AUC vs epochs | No clear effect |
| add error-feedback Top-10% compression | CE-DSGDAR · AUC maximization, 4 datasets | communication cost (MB) | Improved |
| add error-feedback compression | CE-DSGDAR · Fair classification (CIFAR10, CIFAR100, STL10) | test accuracy | No clear effect |
| add error-feedback compression | CE-DSGDAR · Fair classification | communication cost | Improved |
| vary topology among ring, torus, random (ablation) | CE-DSGDAR · CIFAR10 AUC | test AUC | No clear effect |
| compress with Top-5% instead of Top-10% (ablation) | CE-DSGDAR · CIFAR10 AUC | test AUC | No clear effect |
| vary balance coefficients alpha_x=alpha_y in {0.1..0.9} (ablation) | CE-DSGDAR · CIFAR10 AUC | test AUC | No clear effect |
| vary balance coefficients alpha_u=alpha_v in {0.01..0.1} (ablation) | CE-DSGDAR · CIFAR10 AUC | test AUC | No clear effect |
| use the same scale for alpha_x,alpha_y and alpha_u,alpha_v (ablation) | CE-DSGDAR · CIFAR10 AUC | convergence | Hurt |
Closest papers by meaning
- Decentralized Stochastic Optimization and Gossip Algorithms with Compressed Communication (ICML 2019 · similarity 0.91)
- Jointly Improving the Sample and Communication Complexities in Decentralized Stochastic Minimax Optimization (AAAI 2024 · similarity 0.91)
- Towards Faster Decentralized Stochastic Optimization with Communication Compression (ICLR 2025 · similarity 0.91)
- A Faster Decentralized Algorithm for Nonconvex Minimax Problems (NeurIPS 2021 · similarity 0.90)
- Towards Stability and Generalization Bounds in Decentralized Minibatch Stochastic Gradient Descent (AAAI 2024 · similarity 0.89)
- Scalable Distributed Stochastic Optimization via Bidirectional Compression: Beyond Pessimistic Limits (NeurIPS 2026 · 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.
