Recent Publications

 author = {Ndiaye, E., Fercoq, O., Gramfort, A. and Salmon, J.},
 booktitle = {Proc. NIPS 2016},
 pdf = {},
 title = {{GAP} Safe Screening Rules for {Sparse-Group Lasso}},
 year = {2016}

 author = {Ndiaye, E., Fercoq, O., Gramfort, A., Leclère, V. and Salmon, J.},
 link = {},
 pdf = {},
 title = {Efficient Smoothed Concomitant Lasso Estimation for High Dimensional Regression},
 year = {2016}

 address = {Trento, Italy},
 author = {Jas, M., Engemann, D., Raimondo, F., Bekhti, Y. and Gramfort, A.},
 booktitle = {6th International Workshop on Pattern Recognition in Neuroimaging (PRNI)},
 hal_id = {hal-01313458},
 hal_version = {v1},
 keyword = {magnetoencephalography ; electroencephalography ; preprocessing ; artifact rejection ; automation ; machine learning},
 link = {},
 month = {Jun},
 pdf = {},
 title = {Automated rejection and repair of bad trials in MEG/EEG},
 year = {2016}

 address = {Trento, Italy},
 author = {Bekhti, Y., Strohmeier, D., Jas, M., Badeau, R. and Gramfort, A.},
 booktitle = {6th International Workshop on Pattern Recognition in Neuroimaging (PRNI)},
 doi = {10.1109/PRNI.2016.7552337},
 hal_id = {hal-01313567},
 hal_version = {v2},
 keyword = {Inverse problem ;  MEEG ;  iterative reweighted optimization algorithm ;  multi-scale dictionary ;  Gabor transform.},
 link = {},
 month = {Jun},
 pdf = {},
 title = {M/EEG source localization with multi-scale time-frequency dictionaries},
 year = {2016}

 author = {Eickenberg, M., Gramfort, A., Varoquaux, G. and Thirion, B.},
 doi = {},
 issn = {1053-8119},
 journal = {NeuroImage},
 link = {},
 note = {},
 number = {},
 pages = {-},
 pdf = {},
 title = {Seeing it all: Convolutional network layers map the function of the human visual system},
 volume = {},
 year = {2016}

 author = {Strohmeier, D., Bekhti, Y., Haueisen, J. and Gramfort, A.},
 doi = {10.1109/TMI.2016.2553445},
 issn = {0278-0062},
 journal = {IEEE Transactions on Medical Imaging},
 link = {},
 month = {Oct},
 number = {10},
 pages = {2218-2228},
 pdf = {},
 title = {The Iterative Reweighted Mixed-Norm Estimate for Spatio-Temporal MEG/EEG Source Reconstruction},
 volume = {35},
 year = {2016}

 author = {Laby, R., Gramfort, A., Roueff, F., Enderli, C. and Larroque, A.},
 hal_id = {hal-01167391},
 hal_version = {v1},
 link = {},
 month = {Jun},
 pdf = {},
 title = {Sparse pairwise Markov model learning for anomaly detection in heterogeneous data},
 year = {2015}

Full list of publications

Short Bio

I'm currently assistant professor at Telecom ParisTech and scientific consultant for the CEA Neurospin brain imaging center. My work is on statistical machine learning, signal and image processing, optimization, scientific computing and software engineering with primary applications in brain functional imaging (MEG, EEG, fMRI). Before joining Telecom ParisTech, I worked at the Martinos Center for Biomedical Imaging at Harvard in Boston, and for five years at INRIA in the Parietal Project Team and the Athena Project Team. I am also an active member of the Center for Data Science at Université Paris-Saclay.



Address: 46, rue Barrault, 75013 Paris


  • scikit-learn - A Python project for machine learning.
  • openmeeg - C++ package for low-frequency bio-electromagnetism including the EEG/MEG forward problem. OpenMEEG implements the Symmetric BEM which has shown to provide very accurate solutions. Some features: parallel processing, Python Bindings, Matlab integration with Fieldtrip and BrainStorm.
  • MNE - A complete package to process EEG and MEG data: forward and inverse problems (MNE, dSPM, MxNE), stats, time-frequency analysis.

More on my Github Page




PhD Students



  • Engineer to work on scikit-learn
  • Engineer to work on MNE
  • Post-doc positions on machine learning and signal processing with applications in neuroimaging (MEG, EEG)

This list is fuzzy so please contact me directly for potential opportunities.




Multi-task regression consists in inferring jointly multiple regression models for different prediction tasks. The intuition is that a single joint estimation outperforms several estimations carried out independently if tasks share some similarities, which happens typically when different subsets of features are useful for all regression tasks. This intuition led to several seminal contributions in the machine learning literature, e.g. multi-task Lasso or Multi-task Feature Learning (MTFL) (Argyriou et al. 2006, Obozinski et al. 2006)

However, this assumption of common features between all tasks can be too restrictive. For instance, in the context of functional brain imaging, where features map to brain regions, the multitask assumption implies that the exact same locations are active for all human brains. This assumption is clearly not realistic (Gramfort et. al. 2015).

Various works have attempted to tackle the problem with more or less ad-hoc methods (Kozunov et. al. 2015). The idea of this research project will be to develop a new approach to multi-task regression using regularization terms that will be defined using optimal transport theory, taking advantage of recent computational progress in the numerical computation of Wasserstein distances and their derivatives (Cuturi 2013, Gramfort et. al. 2015, Cuturi et al. 2016).

We will carry out validation experiments and adequate benchmarking starting first with simulations. We will then work with the open dataset from (Wakeman et al. 2015) and start from a preliminary analysis.

Required skills: The successful candidate will have background knowledge in machine learning and optimization at the master level (M2), as well as interest in application of data science. Knowledge of scientific computing in Python (Numpy, Scipy) is encouraged, since we will rely on the MNE for data analysis (Gramfort et al. 2014).

Location and supervision

Internship will take place at INRIA Saclay (Turing building) and ENSAE ParisTech and will be supervised my Marco Cuturi and myself.

To apply

Send me an email


[1] Andreas Argyriou, Theodoros Evgeniou, and Massimiliano Pontil. Multi-task feature learning. In P. B. Schölkopf, J. C. Platt, and T. Hoffman, editors, Advances in Neural Information Processing Systems 19, pages 41–48. MIT Press, 2007.

[2] Marco Cuturi. Sinkhorn distances: Lightspeed computation of optimal transport. In Advances in Neural Information Processing Systems 26, pages 2292– 2300, 2013.

[3] Marco Cuturi and Gabriel Peyré. A smoothed dual approach for variational wasserstein problems. SIAM Journal on Imaging Sciences, 9(1):320–343, 2016.

[4] Alexandre Gramfort, Martin Luessi, Eric Larson, Denis A. Engemann, Daniel Strohmeier, Christian Brodbeck, Lauri Parkkonen, and Matti S. Hämäläinen. MNE software for processing MEG and EEG data. NeuroImage, 86(0):446 – 460, 2014.

[5] Alexandre Gramfort, Gabriel Peyré, and Marco Cuturi. Fast optimal transport averaging of neuroimaging data. In Proc. IPMI 2015, July 2015.

[6] Vladimir V. Kozunov and Alexei Ossadtchi. GALA: group analysis leads to accuracy, a novel approach for solving the inverse problem in exploratory analysis of group meg recordings. Frontiers in Neuroscience, 9:107, 2015.

[7] Guillaume Obozinski and Ben Taskar. Multi-task feature selection. In International Conference on Machine Learning (ICML 2006). Workshop of structural Knowledge Transfer for Machine Learning, 2006.

[8] Daniel Wakeman and Richard Henson. A multi-subject, multi-modal human neuroimaging dataset. Scientific Data, 2:150001 EP –, Jan 2015.