Skip to Main content Skip to Navigation
Conference papers

Positive Semidefinite Matrix Factorization Based on Truncated Wirtinger Flow

Dana Lahat 1 Cédric Févotte 1
1 IRIT-SC - Signal et Communications
IRIT - Institut de recherche en informatique de Toulouse
Abstract : This paper deals with algorithms for positive semidefinite matrix factorization (PSDMF). PSDMF is a recently-proposed extension of nonnegative matrix factorization with applications in combinatorial optimization, among others. In this paper, we focus on improving the local convergence of an alternating block gradient (ABG) method for PSDMF in a noise-free setting by replacing the quadratic objective function with the Poisson log-likelihood. This idea is based on truncated Wirtinger flow (TWF), a phase retrieval (PR) method that trims outliers in the gradient and thus regularizes it. Our motivation is a recent result linking PR with PSDMF. Our numerical experiments validate that the numerical benefits of TWF may carry over to PSDMF despite the more challenging setting, when initialized within its region of convergence. We then extend TWF from PR to affine rank minimization (ARM), and show that although the outliers are no longer an issue in the ARM setting, PSDMF with the new objective function may still achieves a smaller error for the same number of iterations. In a broader view, our results indicate that a proper choice of objective function may enhance convergence of matrix (or tensor) factorization methods.
Document type :
Conference papers
Complete list of metadatas

Cited literature [21 references]  Display  Hide  Download
Contributor : Dana Lahat <>
Submitted on : Friday, July 10, 2020 - 2:37:34 PM
Last modification on : Thursday, January 7, 2021 - 8:18:13 PM
Long-term archiving on: : Friday, November 27, 2020 - 10:44:01 AM


Files produced by the author(s)


  • HAL Id : hal-02886419, version 1


Dana Lahat, Cédric Févotte. Positive Semidefinite Matrix Factorization Based on Truncated Wirtinger Flow. European Conference on Signal Processing (EUSIPCO), Jan 2021, Virtual Amsterdam, Netherlands. ⟨hal-02886419⟩



Record views


Files downloads