Interpretable Operator Learning for Inverse Problems via Adaptive Spectral Filtering: Convergence and Discretization Invariance
Hang-Cheng Dong, Pengcheng Cheng \comma \corrauth, Shuhuan Li · 2026
Read Interpretable Operator Learning for Inverse Problems via Adaptive Spectral Filtering: Convergence and Discretization Invariance by Hang-Cheng Dong, Pengcheng Cheng \comma \corrauth, Shuhuan Li (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-2603-20602.txt.
Description
Dong and co-authors propose SC-Net, a spectral method for hard inverse problems. Noise makes such problems unstable, so they need strong regularization. SC-Net works in the spectrum of the forward operator. It learns a filter that weights each part by signal to noise ratio. They prove it tracks the true inverse operator at any grid size. On 1D integral tests it hit the best known convergence rate. It also beat oracle Tikhonov filters and held up on finer grids.
Catalog record
| CREATOR | Hang-Cheng Dong, Pengcheng Cheng \comma \corrauth, Shuhuan Li |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39721 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.20602