Research map / Optimization / Distributed and compressed
Distributed Proximal Gradient Algorithm for Partially Asynchronous Computer Clusters
Yi Zhou, Yingbin Liang, Yaoliang Yu, Wei Dai, Eric P. Xing
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 |
|---|---|---|---|
| allow larger staleness (s=10 to 30) with a staleness-penalized step size eta(alpha_s) (ablation) | m-PAPG · Non-convex group Lasso, n=1000, d=2000 | objective value | No clear effect |
| use the large synchronous step size eta(0) when staleness is nonzero (ablation) | m-PAPG · Non-convex group Lasso, n=1000, d=2000 | convergence | Hurt |
| add a staleness-dependent penalty alpha_s on the learning rate (eta*(alpha_s)) (ablation) | m-PAPG · Non-convex group Lasso, n=1000, d=2000 | stable convergence | Improved |
| increase maximum staleness s from 0 to 1,3,5,7 (ablation) | m-PAPG on Petuum · Large Lasso, n=1M, d=100M, 100 nodes | per-iteration progress | No clear effect |
| increase maximum staleness s from 0 to 1,3,5,7 (ablation) | m-PAPG on Petuum · Large Lasso, n=1M, d=100M, 100 nodes | system throughput (objective vs seconds) | Improved |
Closest papers by meaning
- A Delay-tolerant Proximal-Gradient Algorithm for Distributed Learning (ICML 2018 · similarity 0.90)
- An Accelerated Decentralized Stochastic Proximal Algorithm for Finite Sums (NeurIPS 2019 · similarity 0.87)
- Asynchronous Iterations in Optimization: New Sequence Results and Sharper Algorithmic Guarantees (JMLR 2023 · similarity 0.87)
- Improved Asynchronous Parallel Optimization Analysis for Stochastic Incremental Methods (JMLR 2018 · similarity 0.86)
- Communication Acceleration of Local Gradient Methods via an Accelerated Primal-Dual Algorithm with an Inexact Prox (NeurIPS 2022 · similarity 0.86)
- Stabilized Proximal-Point Methods for Federated Optimization (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.
