Coupled graph tensor factorization

Ahmed S. Zamzam; Vassilis N. Ioannidis; Nicholas D. Sidiropoulos

doi:10.1109/ACSSC.2016.7869683

Coupled graph tensor factorization

Ahmed S. Zamzam, Vassilis N. Ioannidis, Nicholas D. Sidiropoulos

Electrical and Computer Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

5 Scopus citations

Abstract

Factorization of a single matrix or tensor has been used widely to reveal interpretable factors or predict missing data. However, in many cases side information may be available, such as social network activities and user demographic data together with Netflix data. In these situations, coupled matrix tensor factorization (CMTF) can be employed to account for additional sources of information. When the side information comes in the form of item-correlation matrices of certain modes, existing CMTF algorithms do not apply. Instead, a novel approach to model the correlation matrices is proposed here, using symmetric nonnegative matrix factorization. The multiple sources of information are fused by fitting outer-product models for the tensor and the correlation matrices in a coupled manner. The proposed model has the potential to overcome practical challenges, such as missing slabs from the tensor and/or missing rows/columns from the correlation matrices.

Original language	English (US)
Title of host publication	Conference Record of the 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016
Editors	Michael B. Matthews
Publisher	IEEE Computer Society
Pages	1755-1759
Number of pages	5
ISBN (Electronic)	9781538639542
DOIs	https://doi.org/10.1109/ACSSC.2016.7869683
State	Published - Mar 1 2017
Event	50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016 - Pacific Grove, United States Duration: Nov 6 2016 → Nov 9 2016

Publication series

Name	Conference Record - Asilomar Conference on Signals, Systems and Computers
ISSN (Print)	1058-6393

Other

Other	50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016
Country/Territory	United States
City	Pacific Grove
Period	11/6/16 → 11/9/16

Bibliographical note

Funding Information:
A.S. Zamzam and N.D. Sidiropoulos were partially supported by NSF CIF-1525194, ECCS-1231504

Publisher Copyright:
© 2016 IEEE.

Keywords

Tensor factorization
matrix factorization
missing entries
parafac model

Publisher link

10.1109/ACSSC.2016.7869683

Find It

Find It at U of M Libraries

Cite this

Zamzam, A. S., Ioannidis, V. N., & Sidiropoulos, N. D. (2017). Coupled graph tensor factorization. In M. B. Matthews (Ed.), Conference Record of the 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016 (pp. 1755-1759). Article 7869683 (Conference Record - Asilomar Conference on Signals, Systems and Computers). IEEE Computer Society. https://doi.org/10.1109/ACSSC.2016.7869683

Coupled graph tensor factorization. / Zamzam, Ahmed S.; Ioannidis, Vassilis N.; Sidiropoulos, Nicholas D.
Conference Record of the 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016. ed. / Michael B. Matthews. IEEE Computer Society, 2017. p. 1755-1759 7869683 (Conference Record - Asilomar Conference on Signals, Systems and Computers).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Zamzam, AS, Ioannidis, VN & Sidiropoulos, ND 2017, Coupled graph tensor factorization. in MB Matthews (ed.), Conference Record of the 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016., 7869683, Conference Record - Asilomar Conference on Signals, Systems and Computers, IEEE Computer Society, pp. 1755-1759, 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016, Pacific Grove, United States, 11/6/16. https://doi.org/10.1109/ACSSC.2016.7869683

@inproceedings{4b567d6c03804f66ae62575b74eaaeb0,

title = "Coupled graph tensor factorization",

abstract = "Factorization of a single matrix or tensor has been used widely to reveal interpretable factors or predict missing data. However, in many cases side information may be available, such as social network activities and user demographic data together with Netflix data. In these situations, coupled matrix tensor factorization (CMTF) can be employed to account for additional sources of information. When the side information comes in the form of item-correlation matrices of certain modes, existing CMTF algorithms do not apply. Instead, a novel approach to model the correlation matrices is proposed here, using symmetric nonnegative matrix factorization. The multiple sources of information are fused by fitting outer-product models for the tensor and the correlation matrices in a coupled manner. The proposed model has the potential to overcome practical challenges, such as missing slabs from the tensor and/or missing rows/columns from the correlation matrices.",

keywords = "Tensor factorization, matrix factorization, missing entries, parafac model",

author = "Zamzam, {Ahmed S.} and Ioannidis, {Vassilis N.} and Sidiropoulos, {Nicholas D.}",

note = "Funding Information: A.S. Zamzam and N.D. Sidiropoulos were partially supported by NSF CIF-1525194, ECCS-1231504 Publisher Copyright: {\textcopyright} 2016 IEEE.; 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016 ; Conference date: 06-11-2016 Through 09-11-2016",

year = "2017",

month = mar,

day = "1",

doi = "10.1109/ACSSC.2016.7869683",

language = "English (US)",

series = "Conference Record - Asilomar Conference on Signals, Systems and Computers",

publisher = "IEEE Computer Society",

pages = "1755--1759",

editor = "Matthews, {Michael B.}",

booktitle = "Conference Record of the 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016",

}

TY - GEN

T1 - Coupled graph tensor factorization

AU - Zamzam, Ahmed S.

AU - Ioannidis, Vassilis N.

AU - Sidiropoulos, Nicholas D.

PY - 2017/3/1

Y1 - 2017/3/1

N2 - Factorization of a single matrix or tensor has been used widely to reveal interpretable factors or predict missing data. However, in many cases side information may be available, such as social network activities and user demographic data together with Netflix data. In these situations, coupled matrix tensor factorization (CMTF) can be employed to account for additional sources of information. When the side information comes in the form of item-correlation matrices of certain modes, existing CMTF algorithms do not apply. Instead, a novel approach to model the correlation matrices is proposed here, using symmetric nonnegative matrix factorization. The multiple sources of information are fused by fitting outer-product models for the tensor and the correlation matrices in a coupled manner. The proposed model has the potential to overcome practical challenges, such as missing slabs from the tensor and/or missing rows/columns from the correlation matrices.

AB - Factorization of a single matrix or tensor has been used widely to reveal interpretable factors or predict missing data. However, in many cases side information may be available, such as social network activities and user demographic data together with Netflix data. In these situations, coupled matrix tensor factorization (CMTF) can be employed to account for additional sources of information. When the side information comes in the form of item-correlation matrices of certain modes, existing CMTF algorithms do not apply. Instead, a novel approach to model the correlation matrices is proposed here, using symmetric nonnegative matrix factorization. The multiple sources of information are fused by fitting outer-product models for the tensor and the correlation matrices in a coupled manner. The proposed model has the potential to overcome practical challenges, such as missing slabs from the tensor and/or missing rows/columns from the correlation matrices.

KW - Tensor factorization

KW - matrix factorization

KW - missing entries

KW - parafac model

UR - http://www.scopus.com/inward/record.url?scp=85016291628&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85016291628&partnerID=8YFLogxK

U2 - 10.1109/ACSSC.2016.7869683

DO - 10.1109/ACSSC.2016.7869683

M3 - Conference contribution

AN - SCOPUS:85016291628

T3 - Conference Record - Asilomar Conference on Signals, Systems and Computers

SP - 1755

EP - 1759

BT - Conference Record of the 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016

A2 - Matthews, Michael B.

PB - IEEE Computer Society

T2 - 50th Asilomar Conference on Signals, Systems and Computers, ACSSC 2016

Y2 - 6 November 2016 through 9 November 2016

ER -

Coupled graph tensor factorization

Abstract

Publication series

Other

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this