Saved in:
Bibliographic Details
Main Authors: Swijsen, Lars, Van der Veken, Joeri, Vannieuwenhoven, Nick
Format: Preprint
Published: 2021
Subjects:
Online Access:https://arxiv.org/abs/2108.00735
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909233246633984
author Swijsen, Lars
Van der Veken, Joeri
Vannieuwenhoven, Nick
author_facet Swijsen, Lars
Van der Veken, Joeri
Vannieuwenhoven, Nick
contents We propose a Riemannian conjugate gradient (CG) optimization method for finding low rank approximations of incomplete tensors. Our main contribution consists of an explicit expression of the geodesics on the Segre manifold. These are exploited in our algorithm to perform the retractions. We apply our method to movie rating predictions in a recommender system for the MovieLens dataset, and identification of pure fluorophores via fluorescent spectroscopy with missing data. In this last application, we recover the tensor decomposition from less than $10\%$ of the data.
format Preprint
id arxiv_https___arxiv_org_abs_2108_00735
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Tensor completion using geodesics on Segre manifolds
Swijsen, Lars
Van der Veken, Joeri
Vannieuwenhoven, Nick
Differential Geometry
Information Retrieval
Machine Learning
15A69, 53C22, 53C30, 65K05, 90C30, 14P10, 15A83
We propose a Riemannian conjugate gradient (CG) optimization method for finding low rank approximations of incomplete tensors. Our main contribution consists of an explicit expression of the geodesics on the Segre manifold. These are exploited in our algorithm to perform the retractions. We apply our method to movie rating predictions in a recommender system for the MovieLens dataset, and identification of pure fluorophores via fluorescent spectroscopy with missing data. In this last application, we recover the tensor decomposition from less than $10\%$ of the data.
title Tensor completion using geodesics on Segre manifolds
topic Differential Geometry
Information Retrieval
Machine Learning
15A69, 53C22, 53C30, 65K05, 90C30, 14P10, 15A83
url https://arxiv.org/abs/2108.00735