Reservoir Zero-Coordinatewise Projected Subspace Search for Minimization Over Sparse Symmetric Sets in Machine Learning
Morteza Kimiaei, Shima Shabani, Michael Breuß · 2026
Read Reservoir Zero-Coordinatewise Projected Subspace Search for Minimization Over Sparse Symmetric Sets in Machine Learning by Morteza Kimiaei, Shima Shabani, Michael Breuß (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-28124.txt.
Description
Sparse learning problems carry size limits that make them NP-hard. Morteza Kimiaei, Shima Shabani, and Michael Breuß introduce an algorithm in 2026 for minimising over sparse symmetric sets. Their method blends several moves. Coordinate search and symmetry-aware swap updates keep points feasible. Random subspace search escapes poor supports. A reservoir of past feasible points injects fresh directions.
Sparse projection is built into its setup, keeping geometric diversity. Convergence guarantees accompany the method. Under stated assumptions, every full-support accumulation point is zero-coordinatewise stationary, almost surely. A safeguard can extend that to all limit points. After the support stabilises, a local linear rate holds. Experiments on synthetic sparse learning problems follow. The method improves robustness and solution quality over three named competitors. It stays computationally competitive with them.
Catalog record
| CREATOR | Morteza Kimiaei, Shima Shabani, Michael Breuß |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 119 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.28124