Research code

previous arrow
next arrow
Slider

RASA @ ICML2019

Riemannian Adaptive Stochastic gradient algorithm on matrix manifolds

The software is available at Github.

Subsampled-RTR @ NeurIPS2018

 Subsampled Riemannian trust-region (RTR) algorithms

The software is available at Github.

The package contains a MATLAB code of Sub-RTR which is presented in the report “Inexact trust-region algorithms on Riemannian manifolds” by HK and B.Mishra in NeurIPS2018 (NIPS2018). See the detailed explanation of the algorithm here.

R-SRG @ ICML2018

Riemannian stochastic recursive gradient algorithm

The software is available at Github.

R-SVRG @ SIOPT2019 and OPT2016

Riemannian stochastic variance reduced gradient algorithm

The software is available at Github.

Low-rank tensor completion @ ICML2016

The software is available at Dr.Mishra’s web site.

SVRMU @ IEEE ICASSP2018

 Stochastic variance reduced multiplicative updates

The software is available at Github.

This code provides a solver of the stochastic MU rule for a large-scale NMF problem, with a variance-reduced (VR) technique of stochastic gradient. This is published in IEEE ICASSP2018.  

OLSTEC @ IEEE ICASSP2016

OnLine Low-rank Subspace tracking by TEnsor CP Decomposition

The software is available at Github.

This package contains an implementation of OLSTEC, an online tensor subspace tracking algorithm based on the Canonical Polyadic decomposition (CP decomposition) (or PARAFAC or CANDECOMP decomposition) exploiting the recursive least squares (RLS), which is published in Neurocomputing and IEEE ICASSP2016. See the detailed explanation of the algorithm here.

HybridPrecodingOpt @ IEEE GlobalSIP2018

Optimization algorithms for hybrid precoding in millimeter wave (mmWave) MIMO systems

The software is available at Github.

This package provides the codes of the proposed optimization algorithms for hybrid precoding. This code includes existing state-of-the arts algorithms, too. The proposed algorithm is published in IEEE GlobalSIP2018. See the detailed explanation of the algorithm here.