ERC consolidator grant for a PI of SeLMA 

Nicolas Gillis has just been awared an ERC consolidator grant, which is a prestigious 5-year grant of about 2 millions euros. The title of the project is “Beyond Low-Rank Factorizations” (eLinoR) and will tackle non-linear and deep low-rank matrix factorization models. The goal will be to better understand these models from a theoretical standpoint, design new efficient algorithms, and use them on real-world applications. This is a follow-up project of SeLMA.

See https://sites.google.com/site/nicolasgillis/projects/erc-cons for more details.  

Mini-symposium on low-rank factorizations and applications

14:15-15:05, room a002 (Euler building, UCLouvain)
Rafał Zdunek (Politechnika Wrocławska)
Low-rank tensor approximations and selected applications

15:10-16:00, room a002
Nicolas Gillis (Université de Mons)
Partial Identifiability for Nonnegative Matrix Factorization

16:30-17:45, room BARB91 (short walk from Euler building)
PhD thesis defense by Cécile Hautecoeur (UCLouvain)
Nonnegative Matrix Factorization using Parametrizable Functions

2022 meeting of the European Research Network on System Identification (ERNSI)

The 2022 meeting of the European Research Network on System Identification (ERNSI) is being organized by ESAT/STADIUS (KU Leuven). The ERNSI 2022 workshop will take place from Sunday September 18 until Wednesday September 21, 2022 at the Irish College in Leuven, Belgium.

ERNSI is the European research network on system identification and at the workshop, about 100 to 150 researchers come together to talk about system identification, data-driven modelling and related subjects. You can find out more about the workshop’s topics on ERNSI 2021, ERNSI 2019, ERNSI 2018.

URL: www.esat.kuleuven.be/stadius/ERNSI2022/

Workshop on Low-Rank Models and Applications

The Workshop on Low-Rank Models and Applications, sponsored by the ERC, will take place at the Faculty of Engineering, University of Mons, Belgium in September 15-16, 2022. 

The workshop will focus on the use of low-rank models in signal processing, data mining and machine learning. This includes but is not restricted to all aspects of low-rank matrix/tensor approximations, such as nonnegative matrix factorization, (subspace) clustering, independent component analysis, low-rank matrix completion, sparse component analysis, dictionary learning, tensor decompositions, etc. The workshop is also interested in optimization algorithms that are central to compute such decompositions. Many aspects of this class of problems will be discussed including theory (complexity, identifiability, etc.), algorithms, and applications.

The plenary speakers are: Jérôme Bobin (CEA Saclay), Eric Chi (Rice University), Lieven De Lathauwer (KULeuven), Nicolas Dobigeon (IRIT/INP-ENSEEIHT, Université de Toulouse), Le Hien (Huawei Belgium Research Center – BeRC, Leuven), Kejun Huang (University of Florida), Yuji Nakatsukasa (University of Oxford) and Elisa Riccietti (ENS de Lyon).

All the information can be found on the following website: https://sites.google.com/view/lrma21/.

The call for papers is available here.  

Workshop on on Algorithmic and Continuous Optimization

This workshop, dedicated to Algorithmic and Continuous Optimization, will take place between July 4 and July 8 2022 at UCLouvain.

The plenary speakers are: Radu Ioan Boţ (University of Vienna), Coralia Cartis (University of Oxford), Benoît Legat (MIT and UCLouvain), Panos Patrinos (KU Leuven), Nelly Pustelnik (ENS de Lyon) and Adrien Taylor (INRIA and ENS, Paris).

All the information can be found on the following website: https://trade-opt-itn.eu/workshop.html

SeLMA Closing Event

To conclude this SeLMA collaboration, a closing event is organised on June 14th. This includes talks by guest speakers Rodolphe Sepulchre and Philippe Toint, progress presentations per work package, open discussions and poster sessions.

URL: homes.esat.kuleuven.be/~sistawww/biomed/selma/closing/

“AI for Time Series” Seminars

This seminar series brings together researchers that are interested in, or are conducting research related to time series, with a varied program of national and international speakers. These biweekly talks are organized by Bart De Moor, Philippe Dreesen and Katrien De Cock (KU Leuven, ESAT-STADIUS) together with Flanders AI Research and VAIA.

More information, including past and upcoming seminars, can be found at: www.vaia.be/en/series/ai-for-time-series.

Release of Tensorlab⁺

Tensorlab⁺ is a reproducible research repository for tensor computations, built on top of Tensorlab. The first release of Tensorlab⁺ includes the algorithms, auxiliary functions, and data needed to recreate the results of 150+ experiments that are discussed in 34 of our research papers, allowing the reproduction of 180+ figures and tables. Moreover, we have included 32 tutorials and demos showing how the proposed algorithms can be used in your own applications. 

The code is free to use for non-commercial research. We hope that you will find it useful and welcome all suggestions and feedback.

Stijn, Martijn, Nico, Michiel, Robin and Lieven.

URL: http://www.tensorlabplus.net 

First ELO-X Seasonal School and Workshop

The first on-site event organized by ELO-X took place in Leuven, Belgium from March 21 to 30, 2022. The event included a four-day seasonal school on training in technical and soft skills at KU Leuven, followed by a two-day workshop on smart industrial control systems hosted by Siemens.

More information can be found at: elo-x.eu/?p=987.

Debate on AI in Healthcare

On Thursday 16 December 2021 Bart De Moor (KU Leuven, ESAT-STADIUS) had a debate with Belgian Secretary of State Thomas Dermine about “AI in Healthcare”, in the series “In digital we trust”. Watch the video here: indigitalwetrust.be/index-nl.html.