Tree-Adaptive Multiscale Kernel Lasso in Samplet Coordinates
Sara Avesani ∗ , Gaia Fumagalli ∗ , Michael Multerer ∗ , and Chiara Segala ∗ · 2026
Read Tree-Adaptive Multiscale Kernel Lasso in Samplet Coordinates by Sara Avesani ∗ , Gaia Fumagalli ∗ , Michael Multerer ∗ , and Chiara Segala ∗ (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-2604-01794.txt.
Description
Avesani and colleagues develop a framework for sparse multiscale kernel approximation of scattered data problems using samplets. Samplets analyze localized discrete signed measures, enabling quasi-sparse kernel matrices for smooth kernels and smoothness detection. They select sites using the kernel Hilbert space norm of samplet coefficients, yielding few sites that greatly shrink the problem. On the reduced subspace they solve an l1 least-squares problem with a semismooth Newton method, stabilized by online low-rank SVD against ill-conditioning. Tests in two and three dimensions show sparser reconstructions.
Catalog record
| CREATOR | Sara Avesani ∗ , Gaia Fumagalli ∗ , Michael Multerer ∗ , and Chiara Segala ∗ |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 69947 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.01794