Matches in SemOpenAlex for { <https://semopenalex.org/work/W4298857773> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W4298857773 abstract "The phase retrieval problem asks to recover a natural signal $y_0 in mathbb{R}^n$ from $m$ quadratic observations, where $m$ 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 $ell_1$ regularization. While successful in the realm of linear inverse problems, such $ell_1$ methods have encountered possibly fundamental limitations, as no computationally efficient algorithm for phase retrieval of a $k$-sparse signal has been proven to succeed with fewer than $O(k^2log n)$ generic measurements, exceeding the theoretical optimum of $O(k log n)$. 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 $G : mathbb{R}^k rightarrow mathbb{R}^n$ 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 $m = O(kd^2log n)$, where $d$ 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." @default.
- W4298857773 created "2022-10-02" @default.
- W4298857773 creator A5043156635 @default.
- W4298857773 creator A5080237056 @default.
- W4298857773 creator A5089517213 @default.
- W4298857773 date "2018-07-11" @default.
- W4298857773 modified "2023-10-17" @default.
- W4298857773 title "Phase Retrieval Under a Generative Prior" @default.
- W4298857773 doi "https://doi.org/10.48550/arxiv.1807.04261" @default.
- W4298857773 hasPublicationYear "2018" @default.
- W4298857773 type Work @default.
- W4298857773 citedByCount "0" @default.
- W4298857773 crossrefType "posted-content" @default.
- W4298857773 hasAuthorship W4298857773A5043156635 @default.
- W4298857773 hasAuthorship W4298857773A5080237056 @default.
- W4298857773 hasAuthorship W4298857773A5089517213 @default.
- W4298857773 hasBestOaLocation W42988577731 @default.
- W4298857773 hasConcept C102519508 @default.
- W4298857773 hasConcept C106487976 @default.
- W4298857773 hasConcept C11413529 @default.
- W4298857773 hasConcept C114614502 @default.
- W4298857773 hasConcept C121332964 @default.
- W4298857773 hasConcept C129844170 @default.
- W4298857773 hasConcept C134306372 @default.
- W4298857773 hasConcept C135252773 @default.
- W4298857773 hasConcept C153258448 @default.
- W4298857773 hasConcept C154945302 @default.
- W4298857773 hasConcept C159985019 @default.
- W4298857773 hasConcept C163258240 @default.
- W4298857773 hasConcept C163716315 @default.
- W4298857773 hasConcept C192562407 @default.
- W4298857773 hasConcept C204323151 @default.
- W4298857773 hasConcept C207467116 @default.
- W4298857773 hasConcept C2524010 @default.
- W4298857773 hasConcept C2776135515 @default.
- W4298857773 hasConcept C2780992000 @default.
- W4298857773 hasConcept C33923547 @default.
- W4298857773 hasConcept C41008148 @default.
- W4298857773 hasConcept C50644808 @default.
- W4298857773 hasConcept C62520636 @default.
- W4298857773 hasConcept C81793267 @default.
- W4298857773 hasConceptScore W4298857773C102519508 @default.
- W4298857773 hasConceptScore W4298857773C106487976 @default.
- W4298857773 hasConceptScore W4298857773C11413529 @default.
- W4298857773 hasConceptScore W4298857773C114614502 @default.
- W4298857773 hasConceptScore W4298857773C121332964 @default.
- W4298857773 hasConceptScore W4298857773C129844170 @default.
- W4298857773 hasConceptScore W4298857773C134306372 @default.
- W4298857773 hasConceptScore W4298857773C135252773 @default.
- W4298857773 hasConceptScore W4298857773C153258448 @default.
- W4298857773 hasConceptScore W4298857773C154945302 @default.
- W4298857773 hasConceptScore W4298857773C159985019 @default.
- W4298857773 hasConceptScore W4298857773C163258240 @default.
- W4298857773 hasConceptScore W4298857773C163716315 @default.
- W4298857773 hasConceptScore W4298857773C192562407 @default.
- W4298857773 hasConceptScore W4298857773C204323151 @default.
- W4298857773 hasConceptScore W4298857773C207467116 @default.
- W4298857773 hasConceptScore W4298857773C2524010 @default.
- W4298857773 hasConceptScore W4298857773C2776135515 @default.
- W4298857773 hasConceptScore W4298857773C2780992000 @default.
- W4298857773 hasConceptScore W4298857773C33923547 @default.
- W4298857773 hasConceptScore W4298857773C41008148 @default.
- W4298857773 hasConceptScore W4298857773C50644808 @default.
- W4298857773 hasConceptScore W4298857773C62520636 @default.
- W4298857773 hasConceptScore W4298857773C81793267 @default.
- W4298857773 hasLocation W42988577731 @default.
- W4298857773 hasOpenAccess W4298857773 @default.
- W4298857773 hasPrimaryLocation W42988577731 @default.
- W4298857773 hasRelatedWork W1933951994 @default.
- W4298857773 hasRelatedWork W1965977581 @default.
- W4298857773 hasRelatedWork W2007593159 @default.
- W4298857773 hasRelatedWork W2052969112 @default.
- W4298857773 hasRelatedWork W2149216076 @default.
- W4298857773 hasRelatedWork W2894197999 @default.
- W4298857773 hasRelatedWork W2903255608 @default.
- W4298857773 hasRelatedWork W3120181288 @default.
- W4298857773 hasRelatedWork W4247525623 @default.
- W4298857773 hasRelatedWork W4298143267 @default.
- W4298857773 isParatext "false" @default.
- W4298857773 isRetracted "false" @default.
- W4298857773 workType "article" @default.