AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

A Nonconvex Approach for Phase Retrieval: Reshaped Wirtinger Flow and Incremental Algorithms

Huishuai Zhang, Yingbin Liang, Yuejie Chi

JMLR 2017 · Optimization › Implicit bias and generalization · approach: Low-rank and single-index recovery · 99 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use RWF (reshaped, lower-order loss) instead of TWFbatch gradient method · Real Gaussian, n=5000, m=8npasses: 186 -> 72Improved
use RWF instead of WFbatch gradient method · Real Gaussian, n=5000, m=8ntime (s): 54.83 -> 12.66Improved
use RWF instead of AltMinPhasebatch method · Complex Gaussian, n=5000, m=8ntime (s): 9637 -> 122.4Improved
use mini-batch incremental updates (mini-batch IRWF 64) instead of batch RWFRWF · Real Gaussian, n=5000, m=8ntime (s): 12.66 -> 8.076Improved
use incremental update instead of batch (IRWF vs RWF)RWF · Real Gaussian, n=5000, m=8npasses: 72 -> 9Improved
use IRWF update instead of ITWFmini-batch incremental · Real Gaussian, n=5000, m=8ntime (s): 37.38 -> 8.076Improved
use incremental methods instead of batch methodsphase retrieval · Sample complexity, real/complex/CDPsuccess rateImproved
use RWF instead of TWF (truncation)batch methods · Real Gaussian sample complexitysuccess rateHurt
use RWF instead of TWFbatch methods · CDP sample complexitysuccess rateImproved
use RWF instead of TWFbatch methods · Galaxy CDP, L=12passes: 190 -> 70Improved
use RWF initialization instead of WF, TWF, TAFspectral initialization · m=6n, 50 power iterationsrelative error after initializationImproved
use RWF instead of TWFbatch methods · Poisson noise, alpha=1 and 0.001relative errorImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy