Convergence rates of Newton’s method for strongly self-concordant minimization
Nick Tsipinakis, Panos Parpas · 2025
Read Convergence rates of Newton’s method for strongly self-concordant minimization by Nick Tsipinakis, Panos Parpas (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-2507-23558.txt.
Description
Tsipinakis and Parpas study Newton's method for a subclass of well behaved functions in twenty twenty-five. Newton's method is a classic algorithm for minimization. Its theory is well developed for self concordant functions. A local analysis for strongly self concordant functions, a narrower class, was missing.
The known quadratic convergence for the wider class applies here as well. The authors ask whether the narrower class enjoys stronger guarantees. They show its quadratic rate differs from the general case. For a wide range of objectives, convergence is provably faster. The region where local convergence holds is also larger. The results close a gap in the theory of Newton's method. They clarify what extra structure buys in practice.
Catalog record
| CREATOR | Nick Tsipinakis, Panos Parpas |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 24 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-23558