Comprehensive Examination of Unrolled Networks for Solving Linear Inverse Problems
Eric Chen, Xi Chen, Arian Maleki, Shirin Jalali · 2025
Read Comprehensive Examination of Unrolled Networks for Solving Linear Inverse Problems by Eric Chen, Xi Chen, Arian Maleki, Shirin Jalali (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-2501-04608.txt.
Description
Eric Chen, Xi Chen, Arian Maleki, and Shirin Jalali survey in a 2025 paper design choices in unrolled networks for inverse problems. These networks unroll an optimizer into layers and shine on imaging tasks. Moving them to new tasks forces many design picks, each needing costly training runs to judge.
Picks span the base algorithm, the loss shape, and layer counts among many more. The paper unifies scattered ideas to shrink the choice set for new users. A broad ablation then measures how each choice moves final performance. Practical recipes distill from those results for scientists and engineers. Problem spotting inside a misbehaving build also gets guidance in the study.
Catalog record
| CREATOR | Eric Chen, Xi Chen, Arian Maleki, Shirin Jalali |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 69 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-04608