Learning with the p p -adics
André F. T. Martins 1,2 · 2025
Read Learning with the p p -adics by André F. T. Martins 1,2 (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-2512-22692.txt.
Description
André Martins poses a basic question for machine learning: must learned representations live in ordinary real-number spaces? Nearly all current methods use real vector spaces, where distance, projection, and gradient tools feel natural. Martins studies the p-adic numbers instead, a number system where closeness is measured by divisibility by a prime, giving a branching, tree-like geometry well suited to hierarchies.
The paper builds first building blocks for classification, regression, and representation learning over the p-adics, with concrete models and algorithms. A small semantic network of the kind proposed by Quillian fits neatly into a compact p-adic linear network, a trick the reals cannot match. The author closes by listing open problems and arguing that this strange geometry could become a useful home for hierarchical data.
Catalog record
| CREATOR | André F. T. Martins 1,2 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 88 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-22692