Hamza has successfully defended his Phd thesis entitled : « Estimation efficace cerveau entier de la réponse hémodynamique pour la déconvolution semi-aveugle de l’activité neurale régularisée par variation totale en IRM fonctionnelle », today, in front of the following committee:
Hicham will defend his PhD thesis on March 23rd at 10 am, online.
Title: Advances in Optimal transport and applications to Neuroscience
Brain imaging devices can provide a glimpse at neural activity in multiple spatial locations and time points. Moreover, neuroimaging studies are usually conducted for multiple individuals undergoing the same experimental protocol. Inferring the underlying sources is a challenging inverse problem that can only be tackled by biasing the solutions with prior domain knowledge. Several prior hypotheses have been pursued in the literature such as promoting sparse over dense solutions or solving the problem for multiple subjects at once. However, none take advantage of the particular spatial geometry of the problem. The purpose of this thesis is to exploit the multi-subject, spatial and temporal aspects of magneto-encephalography data as much as possible to improve the conditioning of the inverse problem. To that end, our contributions revolve around three axes: optimal transport (OT), sparse multi-task regression and time series. Indeed, the ability of OT to capture spatial disparities between measures makes it very well suited to compare and average neural activation patterns based on their shape and location over the cortical surface of the brain. For the sake of scalability, we take advantage of the entropic formulation of optimal transport, which we argue has two important missing pieces. From a theoretical perspective, it has no closed form analytical expressions, and from a practical perspective, entropy leads to a significant increase in variance known as entropic bias. We complete this puzzle by studying multivariate Gaussians for which we uncover an entropic OT closed form and propose debiased algorithms to compute fast and accurate optimal transport barycenters. Second, we define a multi-task prior based on OT and sparse penalties to jointly solve the inverse problem for multiple subjects to promote spatially coherent solutions. Our real data experiments highlight the benefits of using OT as a prior over classical multi-task regression penalties. Finally, we propose a loss function to compare and average spatio-temporal data that computes temporal alignments across spatially similar observations of the data via a fast GPU friendly algorithm.
Antonia Machlouzarides Shalit is defending her PhD thesis entitled “Development of subject-specific representations of neuroanatomy via a domain-specific language”
on Dec 15th. Follow it live on https://www.youtube.com/watch?v=jiS3w20K22Q
Jerome Alexis is defending his PhD thesis entitled “Statistical control of sparse models in high dimension” on Friday, Dec 11th at 2pm. The committee comprises:
Mme Chloé-Agathe AZENCOTT – Reviewer
M. Thomas NICHOLS – Reviewer
M. Pierre NEUVIAL – Examiner
M. Christophe AMBROISE – Examiner
M. Joseph SALMON – Co-encadrant de thèse : email@example.com
M. Bertrand THIRION – Directeur de thèse : firstname.lastname@example.org
Lots of great features for plotting and and GLM fitting. Please see
http://nilearn.github.io/whats_new.html#v0-7-0 for a summary of new features.
Thx to all the developers !
Maturin Massias, former PhD student of Parietal, is receiving the PGMO PhD Prize for his work “Sparse high dimensional regression in the presence of colored heteroscedastic noise : application to M/EEG source imaging” and implementation of his ideas in CELER (https://github.com/mathurinm/CELER).
We’re delighted to announce that the following 6 papers have been accepted at NeurIPS:
- Entropic Optimal Transport between (Unbalanced) Gaussian Measures has a Closed Form Hicham Janati, Boris Muzellec, Gabriel Peyré, Marco Cuturi https://arxiv.org/abs/2006.02572
- Modeling shared responses in Neuroimaging Studies through MultiViewICA. Hugo Richard, Luigi Gresele, Aapo Hyvärinen, Bertrand Thirion, Alexandre Gramfort, Pierre Ablin. https://arxiv.org/abs/2006.06635
- PEP: Parameter Ensembling by Perturbation. Alireza Mehrtash, Purang Abolmaesumi, Polina Golland, Tina Kapur, Demian Wassermann, William Wells. https://arxiv.org/abs/2010.12721
- Leaning to solve TV regularised problems with unrolled algorithms. Hamza Cherkaoui, Jeremias Sulam, Thomas Moreau. https://hal.archives-ouvertes.fr/hal-02954181
NeuMiss networks: differential programming for supervised learning with missing values. Marine Le Morvan, Julie Josse, Thomas Moreau, Erwan Scornet, Gaël Varoquaux. https://arxiv.org/abs/2007.01627
Statistical control for spatio-temporal MEG/EEG source imaging with desparsified multi-task Lasso. Jérôme-Alexis Chevalier, Alexandre Gramfort, Joseph Salmon, Bertrand Thirion. https://arxiv.org/abs/2009.14310
Congratulations to all !
We are pleased to share the good news that three papers from our team have been accepted for AISTATS 2020 (International Conference on Artificial Intelligence and Statistics):
Debiased Sinkhorn barycenters
Hicham Janati (Inria and CREST/ENSAE), Marco Cuturi (Google and CREST/ENSAE), Alexandre Gramfort (Inria)
Support recovery and sup-norm convergence rates for sparse pivotal estimation
Mathurin Massias (Inria), Quentin Bertrand (Inria), Alexandre Gramfort (Inria), Joseph Salmon (Université de Montpellier)
Linear predictor on linearly-generated data with missing values: non consistency and solutions
Marine Le Morvan (CNRS), Nicolas Prost (CMAP), Julie Josse (Polytechnique/Inria), Erwan Scornet (École Polytechnique), Gael Varoquaux (Inria)
Kamalaker Dadi is defending his PhD thesis “Population Imaging for Mental Health” on Sept 14th in Alan Turing (and online !)
Here are the committee members:
- Sylvia Villeneuve, McGill University, Canada
- Pierre Bellec, Université de Montréal, Canada
- Camille Maumet, Université Rennes I, France
- Vincent Frouin, CEA Neurospin, Université Paris-Saclay, France
- Michel Thiebaut, Université de Bordeaux, France
- Gaël Varoquaux, Parietal Inria, Université Paris-Saclay, France
- Bertrand Thirion, Parietal Inria, Université Paris-Saclay, France
- Denis Engemann, Parietal Inria, Université Paris-Saclay, France
- Josselin Houenou, UNIACT Lab, Université Paris-Saclay, France
Please join us to discover his great contributions to the field !
We are looking for an engineer to develop cognitive protocols for a massive brain mapping experiments. More details here: https://team.inria.fr/parietal/files/2020/08/engineer_ibc.pdf