Page 4 - Weekly highlights 6690

IV

Weekly highlights

Eight Moroccan AI talents in top laboratories

M ANY Moroccan research
talents are working in interna-
tional laboratories. Benefiting
from conditions that enable them to
excel, they are enjoying successful ca-
reers. Among them are eight promising
young researchers working in the field
of artificial intelligence (AI). They will
be meeting next February in Benguerir,
as part of the Science Week organized
by the Mohammed VI Polytechnic Uni-
versity (UM6P), where they will present
their work.
«These researchers were not selected
at random. They have achieved an ex-
traordinary feat: that of having a scien-
tific paper accepted at the most presti-
gious AI conference in its modern form:
NeurIPS», notes Rachid Guerraoui,
who will bring these young researchers
together in the event’s «Computing
day» on February 13. Guerraoui, Direc-
tor of the Distributed Computing Labo-
ratory at the Ecole Polytechnique Fédé-
rale de Lausanne (EPFL), is himself
among the world’s top 2% of science ter science and applied mathematics the aim of my research is to analyze and learning». Here, the young doctoral
researchers, according to Stanford Uni- (algorithms, optimization, and control develop learning algorithms capable of engineer develops algorithms that are
versity’s 2020 rankings. «It is as if they systems). He graduated in engineering interacting with complex environments, robust to manipulation, can be proven,
had been selected for Wimbledon in ten- from Ensimag in Grenoble, with a using as little data as possible», explai- and preserve the confidentiality of their
nis or for the World Cup in football», he double degree from the Royal Polytech- ned the young researcher. His ambition users’ data. His passion is research,
says. And with good reason: Neurips nic School of Stockholm (KTH). And it is to embark on a career as a teacher- which he hopes to continue. In acade-
is where the most advanced results in was at KTH that he completed his PhD researcher in the field of machine lear- mia or industry? He hasn’t made up his
machine learning from both academia in reinforcement learning. Today, Yassir ning. mind yet.
and industry are unveiled. «This is Jerda is a post-doc at MIT, where he is Youssef Allouah is a PhD student in Some of these young talents have
where new AI techniques are presented, conducting research into reinforcement computer science at EPFL. He des- passed through the Lydex preparatory
making it possible to predict diseases, learning. «This type of learning is ap- cribes himself as « a theoretician by courses in Benguerir, like Oussama
forecast natural disasters, invent new plied in many fields, such as robotics, training, at the intersection of compu- Boussif, currently attached to the MILA
medicines, drive a driverless car, and medical treatment, and recommender ter science and mathematics». His field Institute.o
so on. It is also where new versions of systems, among others. In a few words, of research is «trustworthy machine Ahlam NAZIH
ChatGPT are presented», explained
Rachid Guerraoui.
The eight young talents work in some AI will change our lives!
of the world’s leading laboratories
(see illustration), including those at the HOW will AI evolve over the next few years? For Yassir Jedra (MIT), it is
Massachusetts Institute of Technology clear that AI will change our lives, both socially and in the workplace. «We’re
(MIT), the Montreal Institute for Lear- already seeing AI-based tools like ChatGPT that have become indispen-
ning Algorithms (MILA), the Ecole sable in the day-to-day work of programmers», says the young resear-
Polytechnique Fédérale de Lausanne cher to illustrate his idea. For him, AI will also play a major role in
(EPFL), the Ecole Polytechnique Fédé- future scientific innovations, particularly in fields such as medicine,
rale de Zurich (ETHZ) and the Ecole biology, physics, and even mathematics. «We can already cite
Nationale de la Statistique et de l’Admi- a few examples, such as Alphafold from Deepmind (a Google
nistration Économique (ENSAE, Paris). subsidiary), which predicts the structure of proteins from their
amino acids alone, or, just recently, Alphageometry, also from
«Reinforcement learning», DeepMind, making it possible to solve fairly complex geometry
problems, at the level of the International Mathematical Olym-
«trustworthy machine piad», added Yassir Jedra.
learning»... For Youssef Allouah (EPFL), AI, which is attracting an increasing
number of talented researchers and entrepreneurs, some of whom

They are positioned in cutting-edge have branched off from their original fields, has a bright future ahead
specialties, and work on cutting-edge of it. Nonetheless, «there will inevitably be problems of reliability and
research themes. Yassir Jerda (MIT), regulation as AI is deployed on a large scale. Our research is ideally placed
for example, specializes in compu- to support this deployment», he concluded.o
Friday 26 January 2024