Research map / Optimization / Optimizers
The benefits of full data shuffle, now with optimal I/O cost: $k$-wise independence and matrix transposition to the rescue
Peyman Afshani, Rezaul Chowdhury, Mayank Goswami, Jens Kristian Refsgaard Schou, Francesco Silvestri, Mariafiore Tognon
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 two rounds of IO Shuffle instead of one (ablation) | IO Shuffle · Block-level clustering, N=16,000,057, B=400 | median preserved pairs: 60 591 -> 79 598 | Improved |
| IO Shuffle (2 rounds) versus CorgiPile (1 round) | IO Shuffle · Block-level clustering, N=16,000,057, B=400 | median preserved pairs: 8 017 871 -> 79 598 | Improved |
| Gen-2-Wise-Ind-Perm versus uniform shuffle | Gen-2-Wise-Ind-Perm · Block-level clustering, N=16,000,057, B=400 | median preserved pairs: 0 vs 79 600 | Improved |
Closest papers by meaning
- Improved Last-Iterate Convergence of Shuffling Gradient Methods for Nonsmooth Convex Optimization (ICML 2025 · similarity 0.87)
- On the Last-Iterate Convergence of Shuffling Gradient Methods (ICML 2024 · similarity 0.87)
- Shuffling Gradient-Based Methods for Nonconvex-Concave Minimax Optimization (NeurIPS 2024 · similarity 0.86)
- A Unified Analysis of Stochastic Gradient Descent with Arbitrary Data Permutations and Beyond (NeurIPS 2025 · similarity 0.86)
- Tighter Convergence Bounds for Shuffled SGD via Primal-Dual Perspective (NeurIPS 2024 · similarity 0.86)
- Revisiting Convergence: Shuffling Complexity Beyond Lipschitz Smoothness (ICML 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.
