The Ball-Proximal (=“Broximal”) Point Method: a New Algorithm, Convergence Theory, and Applications
Kaja Gruntkowska, Hanmin Li, Aadi Rane ⋆, Peter Richtárik · 2025
Read The Ball-Proximal (=“Broximal”) Point Method: a New Algorithm, Convergence Theory, and Applications by Kaja Gruntkowska, Hanmin Li, Aadi Rane ⋆, Peter Richtárik (2025) 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-2502-02002.txt.
Description
Kaja Gruntkowska, Hanmin Li, Aadi Rane, and Peter Richtárik present the Broximal Point Method in two thousand twenty five. Non smooth, non convex optimization defeats plain gradient methods. The classic proximal point method inspires their new frame. A ball constraint replaces the quadratic distance penalty inside.
The swap yields linear, finite step convergence in the nonsmooth convex case. That beats the sublinear rate of the classic method. A new ball convexity notion widens the guarantees further. Acceleration, smoothing, and trust region ideas all connect back. The method is offered as a blueprint for practical algorithms.
Catalog record
| CREATOR | Kaja Gruntkowska, Hanmin Li, Aadi Rane ⋆, Peter Richtárik |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 130 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-02002