English

PtyGenography: using generative models for regularization of the phase retrieval problem

Machine Learning 2025-02-04 v1 Information Theory Machine Learning Functional Analysis math.IT Optimization and Control

Abstract

In phase retrieval and similar inverse problems, the stability of solutions across different noise levels is crucial for applications. One approach to promote it is using signal priors in a form of a generative model as a regularization, at the expense of introducing a bias in the reconstruction. In this paper, we explore and compare the reconstruction properties of classical and generative inverse problem formulations. We propose a new unified reconstruction approach that mitigates overfitting to the generative model for varying noise levels.

Keywords

Cite

@article{arxiv.2502.01338,
  title  = {PtyGenography: using generative models for regularization of the phase retrieval problem},
  author = {Selin Aslan and Tristan van Leeuwen and Allard Mosk and Palina Salanevich},
  journal= {arXiv preprint arXiv:2502.01338},
  year   = {2025}
}
R2 v1 2026-06-28T21:30:34.853Z