On Universality of Non-Separable Approximate Message Passing Algorithms
Max Lovig ∗ , Tianhao Wang † , Zhou Fan ∗ · 2025
Read On Universality of Non-Separable Approximate Message Passing Algorithms by Max Lovig ∗ , Tianhao Wang † , Zhou Fan ∗ (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-2506-23010.txt.
Introduction
Max Lovig, Tianhao Wang, and Zhou Fan ask when a widely used iterative method behaves the same no matter how the data is spread, in a paper published in two thousand twenty-five. The method is approximate message passing, used to study learning dynamics in statistics.
For methods whose steps split cleanly by coordinate, such sameness was already known. Steps that mix coordinates were understood only for tidy data, such as Gaussian or rotation-friendly matrices. The authors name a condition on the method's building blocks, the Bounded Composition Property. Under it, for steps built from polynomials, the same step-by-step forecast holds for matrices with non-Gaussian entries. The work opens a study of sameness for this harder class of methods.
Catalog record
| CREATOR | Max Lovig ∗ , Tianhao Wang † , Zhou Fan ∗ |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 339 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-23010