Over the years, I have adapted my strategy to share codes related to publications. Back during my PhD thesis, I made one big github repo with all my codes. But this approach is not so practical when users want to only try out methods proposed in a single work, or to collaborate with other researchers.  Nowadays, I tend to create a single repo for each paper, therefore the codes for recent works can be found under the PUBLICATION LIST tab next to each relevant paper.


Moreover, with the help of Caglayan Tuna, I am an active contributor to the tensorly toolbox, which allows to perform many standard tensor decompositions and operations while choosing the relevant backend (e.g. numpy, pytorch, mxnet, tensorflow...).

Link to old Github Repository : click here.


If you are allergic to git, I also share some of my older codes on dropbox, links below. (They are NOT updated regularly though)

Matlab Codes

Nonnegative PARAFAC2

Flexible penalization-based algorithm for computing Nonnegative PARAFAC2, constrained on possibly all modes (especially the coupled one).

link to dropbox

Reference: J. E. Cohen, R. Bro, "Nonnegative PARAFAC2, a flexible coupling approach", LVA/ICA 2018, arXiv:1802.05035

Multiple Dictionary Low Rank Factorization (M2PALS)

M2PALS algorithm for sparse coding with multiple available dictionaries.

link to dropbox

Reference: J. E. Cohen, N. Gillis, "Spectral Unmixing with Multiple Dictionaries", IEEE Geoscience and Remote Sensing Letters, vol.15 issue 2, pp. 187-191, 2018. pdf


Codes for Dictionary-based CPD, including MPALS, SMPALS, Fast Gradient, Flexible MPALS.

link to dropbox

Reference: J. E. Cohen, N. Gillis, "Dictionary-based Tensor Canonical Polyadic Decomposition", IEEE Transactions on Signal Processing, vol.66 issue 7, pp. 1876-1889 2018. pdf


Codes for Non-Linear Fluorescence Decomposition (Credits : Xavier Luciani)

link to dropbox

Reference: J. E. Cohen, P. Comon, and X. Luciani, "Correcting inner lter eects, a non multilinear tensor decomposition method," Chemometrics and Intelligent Laboratory Systems, vol. 150, pp. 29-40, 2016.


Codes for Projected-Compressed ALS

link to dropbox

Reference: J. E. Cohen, R. C. Farias, and P. Comon, "Fast decomposition of large nonnegative tensors," IEEE Signal Processing Letters, vol. 22, no. 7, pp. 862-866, 2015.


Codes for Joint Canonical Polyadic Decomposition of 2 tensors with noisy linear coupling.

link to dropbox

Reference : R. Cabral Farias, J. E. Cohen, and P. Comon, "Exploring multimodal data fusion through joint decompositions with flexible couplings," IEEE Transactions on Signal Processing, vol. 64, pp. 4830-4844, September 2016.