Close Menu

Multi-Modal Data Mining Using Coupled Matrix/Tensor Factorizations


Speaker: Dr. Evrim Acar Ataman, Head of Department, Chief Research Scientist, Simula Research Laboratory

Come Join Us On: Friday, November 13, 2020, 10:30 – 11:45 AM Eastern Time


Hosted by: Eric Miller, Chair of Department of Electrical and Computer Engineering, Tufts University

*please register for the event and obtain the passcode here:



In order to understand the functioning of complex systems such as the brain or human metabolome (i.e., the complete set of small biochemical compounds in the body), the system should be recorded using different sensing technologies. For instance, neuroimaging modalities such as functional Magnetic Resonance Imaging (fMRI) and Electroencephalography (EEG) provide information about the brain function in complementary spatiotemporal resolutions. Such multi-modal data sets have the potential to provide a deeper insight into the system when analyzed together, and there is an emerging need for data mining methods that can jointly analyze data from multiple sources, and extract physiologically meaningful patterns.

Coupled matrix/tensor factorizations have proved useful in various domains as an effective multi-modal data mining (data fusion) approach. In this talk, we will go over different formulations of coupled matrix and tensor factorizations, algorithmic approaches, as well as applications in several domains. We will also introduce a flexible framework that facilitates the use of a variety of constraints, loss functions, and couplings between data sets, and discuss the remaining issues.


About Dr. Ataman:

Evrim Acar is a Chief Research Scientist at Simula Metropolitan Center for Digital Engineering (Oslo, Norway). Her research focuses on data mining, in particular, matrix/tensor factorizations, data fusion using coupled factorizations, and their applications in diverse disciplines. Prior to joining Simula, Evrim was a faculty member at the Chemometrics and Analytical Technology group at the University of Copenhagen. She obtained her PhD in Computer Science from Rensselaer Polytechnic Institute (Troy, NY) in 2008, and held a postdoctoral researcher position at Sandia National Labs (Livermore, CA) between 2008-2010.


Additional Resources:

C. Schenker, J. E. Cohen, E. Acar, A Flexible Optimization Framework for Regularized Matrix-Tensor Factorizations with Linear Couplings. arxiv:2007.09605, 2020

E. Acar, C. Schenker, Y. Levin-Schwartz, V. D. Calhoun and T. Adali. Unraveling Diagnostic Biomarkers of Schizophrenia through Structure-Revealing Fusion of Multi-Modal Neuroimaging Data. Frontiers in Neuroscience, 13:416, 2019

E. Acar, R. Bro and A. K. Smilde. Data Fusion in Metabolomics using Coupled Matrix and Tensor Factorizations. Proceedings of the IEEE, 103:1602-1620, 2015