Research map / Optimization / Implicit bias and generalization
Distributed Coordinate Descent Method for Learning with Big Data
Peter Richtárik, Martin Takáč
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 ASL protocol with fully parallel organization instead of RA-PS (ablation) | Hydra · LASSO 3TB, tau=10 | time per iteration (s): 0.040 -> 0.025 | Improved |
| use ASL protocol with fully parallel organization instead of RA-PS (ablation) | Hydra · LASSO 3TB, tau=10^2 | time per iteration (s): 0.100 -> 0.032 | Improved |
| use ASL protocol with fully parallel organization instead of RA-PS (ablation) | Hydra · LASSO 3TB, tau=10^3 | time per iteration (s): 0.321 -> 0.249 | Improved |
| use fully parallel instead of parallel-serial organization (RA protocol) (ablation) | Hydra RA · LASSO 3TB, tau=10^2 | time per iteration (s): 0.100 -> 0.077 | Improved |
| use ASL instead of RA with fully parallel organization (ablation) | Hydra FP · LASSO 3TB, tau=10^2 | time per iteration (s): 0.077 -> 0.032 | Improved |
| increase tau from 128 to 256 coordinates updated per node | Hydra · astro-ph SVM, 32 nodes | duality gap vs iterations | No clear effect |
| increase tau (coordinates updated per node per iteration) | Hydra · astro-ph SVM, 32 nodes | epochs to given accuracy | Hurt |
| increase tau (coordinates updated per node per iteration) | Hydra · astro-ph SVM, 32 nodes | iterations to given accuracy | Improved |
| use ASL-FP instead of RA-FP (ablation) | Hydra · LASSO big-data, block angular | L(x_k)-L* | Improved |
| use a stepsize beta hundred times smaller than theory | Hydra · astro-ph experiment | duality gap | Improved |
Closest papers by meaning
- A General Distributed Dual Coordinate Optimization Framework for Regularized Loss Minimization (JMLR 2017 · similarity 0.89)
- Coordinate Descent Methods for DC Minimization: Optimality Conditions and Global Convergence (AAAI 2023 · similarity 0.87)
- DSCOVR: Randomized Primal-Dual Block Coordinate Algorithms for Asynchronous Distributed Optimization (JMLR 2019 · similarity 0.87)
- Dykstra's Algorithm, ADMM, and Coordinate Descent: Connections, Insights, and Extensions (NeurIPS 2017 · similarity 0.86)
- Accelerated Methods with Compressed Communications for Distributed Optimization Problems Under Data Similarity (AAAI 2025 · similarity 0.84)
- Rescaled Asynchronous SGD: Optimal Distributed Optimization under Data and System Heterogeneity (NeurIPS 2026 · similarity 0.84)
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.
