Pictures from the workshop “Optimization and Statistical Learning” 2019 co-organized by JM, Alexandre d’Aspremont (CNRS, ENS), Zaid Harchaoui (U. Washington) Julien Mairal (Inria) Philippe Rigollet (MIT).
We welcomed Alexander Rogozin from MIPT for a good research visit of 2 weeks.
Picture of the defense with Wim van Ackooij (EDF Research), JM (DAO), and Mikhail Solodov (IMPA Brazil).
Grenoble Optimization Days is a serie of small-size iternational workshops co-organized by DAO. The broad topic of the 2018 session is “optimization methods and applications in statistical learning”, and its goal is to foster exchanges between scientific communities working on high-dimensional data, and to report and discuss recent advances in optimization algorithms for machine learning and optimal transport. This session also celebrates the axe Grenoble-Moscow, and is co-organized with Alexander Gasnikov (MIPT, Moscou). The workshop will gather researchers, PhD students, and master students on mathematical optimization and statistical learning from Grenoble and russian partners from MIPT and HSE in Moscow. The targeted audience for the presentations are the researchers on optimization and learning as well as the PhD students of the Grenoble campus and the master students from the MSIAM Master track “Data Science”.
Details on the website: https://ljk.imag.fr/membres/Jerome.Malick/Oday18.html
F. Iutzeler obtained a funding for 8kE from CNRS INSMI and INS2I – Intelligence artificielle et apprentissage automatique (I3A) call with M. Clausel (IECL, U. Lorraine, and former team member), M.-R. Amini (LIG, Grenoble).
The 2018 French Days on Optimization and Decision Science
Grenoble is hosting the 2018 edition of the French Days on Optimization and Decision Science (Journées SMAI-MODE 2018), the biennial meeting of the MODE special interest group of the French Society for Industrial and Applied Mathematics (SMAI). The event will take place from March 28 to March 30, 2018, and aims at bringing together students, researchers and practitioners in the broad areas of optimization, game theory, control, and decision science in general. It will be an opportunity for researchers to present their recent advances in the field and to initiate new and fruitful collaborations, as was the case with the previous editions.
The 2018 SMAI-MODE meeting will be held in the Escandille conference center in Autrans, in the heart of the Vercors Massif (about one hour from Grenoble). Topics of interest include (but are not limited to):
- Optimization, nonlinear analysis
- Game theory
- Optimal control and optimal transport
- Stochastic optimization and learning
- Operations research
- Mathematical modelization in finance, economics and social sciences
SMAI-MODE 2018 is co-organized by the European Network for Game Theory (COST Action CA16228). Just before the MODE meeting, the French research group on optimization (GdR MOA) is organizing a short course on “Optimization and learning: applications to image processing”. The course will take place on March 26-27, and will be given by Julien Mairal.
The whole team is participating in the organization of this event and J. Malick is general chair.
On Thu., Sep 28th, J. Lelong defended its Habilitation to direct researches entitled “Quelques contributions aux méthodes numériques probabilistes et à la modélisation stochastique” at the Batiment IMAG, on the Grenoble Campus in front of a jury composed of:
- Eric MOULINES (Professeur, Ecole Polytechnique)
- John SCHOENMAKERS (Humbold University, Berlin)
- Denis TALAY (directeur de recherche, INRIA Sophia Antipolis Méditerranée)
- Anatoli IOUDITSKI (Professeur, Université Grenoble Alpes)
- Adeline LECLERCQ SAMSON (Professeur, Université Grenoble Alpes)
- Gilles PAGES (Professeur, Université Paris 6)
M. Clausel and F. Iutzeler are in the organizing committee of CAp 2017 — http://cap2017.imag.fr , the annual meeting of the francophone Machine Learning community.
The conference is a place of exchange and sharing of experiences in the field of machine learning with presentations of original research results. In 2017, the 19th edition of CAp will be held in Grenoble. As the image of the city, characterized by an interdisciplinary culture, the field of machine learning is at the interface of several scientific communities such as computer science, applied mathematics, neuroscience and human sciences. This multidisciplinary has played a leading role in the recent development of the area, both in terms of methodology and applications. This year, CAp will also feature a learning challenge over Tweets (see below). The organizing committee is chaired by Massih-Reza Amini (LIG, Univ. Grenoble Alpes).
Léon Bottou, Facebook
Yves Grandvalet, CNRS and Université de technologie de Compiègne
Yurii Nesterov, Université Catholique de Louvain
Call for papers:
CAp is an interdisciplinary gathering of researchers at the intersection of machine learning, applied mathematics, and related areas. Authors of accepted papers will be invited for oral presentation of their work and for a posters session. This session is an opportunity to have constructing and rigorous feedbacks, as well as to establish contacts with members of the French machine learning community. PhD Students are particularly welcome and encouraged to submit papers. Contributions will be freely distributed on the conference website, subject to approval by the authors. The conference and program chairs of CAp 2017 invite those working in areas related to any aspect of machine learning to submit original papers for review. Solicited topics include, but are not limited to:
– Learning theory, models and paradigm: Learning, Bandit algorithms, Matrix and tensor factorization, Kernel methods, Stochastic processes, Graphical models, Neural networks and deep learning, Game theory
– Optimization and related problems: Large-scale and distributed optimization, Machine learning and structured data (spatio-temporal, tree, graph), Classification with missing values
– Applications: Social network analysis, Bioinformatic, Data mining, Neuroscience, Natural language processing, Information retrieval, Computer vision
Submitted papers can be either in English or in French and we encourage two types of submissions:
– Full research papers on the theme of machine learning theory and its applications should not exceed ten pages in CAp double-column format (including references and figures).
– Short papers can be up to four pages using the same format as Full papers. They present original ideas and provide an opportunity to describe significant work in progress.
This year CAp features a challenge on “Named Entity Recognition in Tweets” which concerns the classification of textual segments of data in a predefined set of categories (13 types). A training dataset of 3,000 tweets will be available to the participants. For more information please visit http://cap.imag.fr/Challenge.html. A prize of 600 Euros will be awarded to the best system.
14 Avril 2017 : Papers Due
26 Mai 2017 : Notification
28-30 Juin 2017 : CAp 2017 Conference
On Thu., Jan 26th, J. Malick defended its Habilitation to direct researches entitled “Variational-analysis look at combinatorial optimization and other selected topics in optimization” at the Maison Jean Kuntzmann, on the Grenoble Campus in front of an international jury composed of:
- Alexandre d’Aspremont, ENS Ulm
- Jérome Bolte, Toulouse School of Economics
- Gérard Cornuéjols, Carnegie Mellon University
- Jean Lasserre, LAAS-CNRS
- Nabil Layaida, Inria
- Yurii Nesterov, Université de Louvain (Belgique)
- Michael Overton, Courant Institute of Mathematical Sciences, NYU
The jury unanimously enjoyed the clarity of the manuscript and presentation as well as the pertinence of his research in modern optimization.