ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1807.04261
14
188

Phase Retrieval Under a Generative Prior

11 July 2018
Paul Hand
Oscar Leong
V. Voroninski
ArXivPDFHTML
Abstract

The phase retrieval problem asks to recover a natural signal y0∈Rny_0 \in \mathbb{R}^ny0​∈Rn from mmm quadratic observations, where mmm is to be minimized. As is common in many imaging problems, natural signals are considered sparse with respect to a known basis, and the generic sparsity prior is enforced via ℓ1\ell_1ℓ1​ regularization. While successful in the realm of linear inverse problems, such ℓ1\ell_1ℓ1​ methods have encountered possibly fundamental limitations, as no computationally efficient algorithm for phase retrieval of a kkk-sparse signal has been proven to succeed with fewer than O(k2log⁡n)O(k^2\log n)O(k2logn) generic measurements, exceeding the theoretical optimum of O(klog⁡n)O(k \log n)O(klogn). In this paper, we propose a novel framework for phase retrieval by 1) modeling natural signals as being in the range of a deep generative neural network G:Rk→RnG : \mathbb{R}^k \rightarrow \mathbb{R}^nG:Rk→Rn and 2) enforcing this prior directly by optimizing an empirical risk objective over the domain of the generator. Our formulation has provably favorable global geometry for gradient methods, as soon as m=O(kd2log⁡n)m = O(kd^2\log n)m=O(kd2logn), where ddd is the depth of the network. Specifically, when suitable deterministic conditions on the generator and measurement matrix are met, we construct a descent direction for any point outside of a small neighborhood around the unique global minimizer and its negative multiple, and show that such conditions hold with high probability under Gaussian ensembles of multilayer fully-connected generator networks and measurement matrices. This formulation for structured phase retrieval thus has two advantages over sparsity based methods: 1) deep generative priors can more tightly represent natural signals and 2) information theoretically optimal sample complexity. We corroborate these results with experiments showing that exploiting generative models in phase retrieval tasks outperforms sparse phase retrieval methods.

View on arXiv
Comments on this paper