• Aucun résultat trouvé

The University of Mons is a key player in this field,


Academic year: 2022

Partager "The University of Mons is a key player in this field, "

En savoir plus ( Page)

Texte intégral



The recent increase in computing power (GPU), coupled with access to very large databases (Big Data) and high-performance learning algorithms (Deep Learning), have resulted in the rapid development of Artificial Intelligence (AI).

The University of Mons is a key player in this field,

with its teaching, continuing education, research,

business support and creation of spin-offs

contributing to advances in the domain.



CHIST-ERA - 1,141,431 € In collaboration with INRIA, Univ. of Sherbrooke, Univ Lille 1, KTH & Univ. de


Use of deep neural learning techniques for automatic language understanding in a multimodal context.

MOBFACENET Win2Wal - 556,096 €

In collaboration with MoodMe

Deep neural networks embedded in a smartphone for face analysis tasks (prediction of age, sex, emotions, etc.).

ROADSTEP 220,319 €

In collaboration with ULiège, A.R.S.I.A., & Centre des Technologies Agronomiques (CTA) The project aims at monitoring cattle with the use of the official electronic identification and AI in order to predict disease.


Convolutional Neural Networks, Deep Reinforcement Learning, Generative Adversarial Networks and even Auto- encoders, are already seen as basic tools which our students and graduates require knowledge about and feel the need to master. This programme specifically offers them not only a broad understanding of recent developments in AI but also the know-how that is requested by the industrial sector.

The Hands on AI programme is spread over 16 evenings and a hackathon week-end to learn about and practise using AI tools (for engineers and computer scientists).

76 participants in 2018-2019

AI Challenges (4 ECTS)

Three applied challenges coming from different domains (such as fire detection, image analysis for applications in Alzheimer’s patients, SLAM for robot vision, etc.).

AI Seminars (1 ECTS)

12 hours of seminars to appreciate the application of AI in a wide range of domains (such as health, ethics, transport, etc.).

AI Workshops (5 ECTS)

A full week-end for all students registered on the Certificate, targeting team work for the realisation of an ambitious AI project.


Some key AI projects


ELIA (Belgium’s electricity transmission system operator) The project aims to model the reactive behaviour of transmission/

distribution interfaces at various time horizons, in the context of increasing penetration of decentralised generation.


(Technologies for Early Diagnosis of Developmental Impairments through

Data Mining) PhD Thesis - FNRS Development of AI tools trained on brain data, allowing clinicians to make early diagnosis of mental disorders in children.

PARKAR WALInnov - 1,951,972 € In collaboration with AlterFace and


Innovative solutions to achieve interactions of individuals in augmented reality with small groups of people.

AI Challenge, 2018-2019

Tangerine (First Spin-off)

Development of tangible (non-virtual) games to learn programming.

AI for Breast Tumor Response Prediction PhD Thesis - UMONS Deep Learning approach predicting breast tumour response to neoadjuvant treatment using two DCE-MRI exams for patients’ data provided by the Jules Bordet Institute.

Acapela (1996) Personalised digital

voice creation

Hovertone (2016) Interactive technology-

based installations

Ittention (2016) Predictive visual impact

Dephten (2019) Image analysis and automatic tagging


Reasons to join ?

Staggered schedule (evenings and 1 week-end) Open to PhD students as well as to adults returning to education

Hands-on training with concrete projects and practical challenges

Ideal place to network with AI experts and make industrial contacts

Teaching and continuing education

UMONS integrates a large number of AI-related courses into its various Masters degrees in engineering and computer science. Starting in the 2019-2020 academic year, artificial intelligence will be explicitely integrated as a specialist focus of three UMONS Master’s degrees in engineering and computer science. This gives the students the opportunity to access a large range of AI-focused courses and to deepen their expertise with an internship and Master dissertation in the field. Artificial intelligence is also part of a wide variety of curricula in other faculties, including those of the Faculty of Architecture and Urban Planning, and the Faculty of Psychology & Education.

Since 2018, UMONS has also developed a university certificate worth 10 ECTS in AI : Hands on AI.

CLEO (First Spin-off)

Applying AI to the reconstruction of the 3D internal structure of a bone from simple x-rays.

EGC (First Spin-off)

Automatic and polyphonic transcription tool for guitar providing notes, chords and playing techniques automatic detection.

Audio, Speech and

Music Multimedia

Information Retrieval


Medical Imagery

AI for Energy Transition

Smart Rurality Education

with AI

Edge AI

Smart Spaces The University of Mons (UMONS) has structured the organisation of

its research around 10 institutes in order to prioritise the quality and the impact of its research and to ease technology transfer to the benefit of society.

Three of these institutes are dedicated to information and communication technologies (ICT) : Numediart, Infortech and Complexys. They gather more than 200 researchers, many of whom are now tackling AI-related issues.

Naturally, AI also impacts research in other UMONS institutes be it in energy (smart grids) or social and educational sciences.

Motion Capture


Reasoning Operational

Research Time Series




UMONS Innovation Centers in AI

Contact us

Thierry Dutoit

President of the Numediart Institute thierry.dutoit@umons.ac.be

Ecosystem building and regional impact

At the heart of a rich ecosystem, UMONS collaborates on many AI initiatives in close connection with a wide range of partners. It also benefits from a strong connection to Multitel and CETIC and to local innovation and financing structures such as Creative Valley, Digital Attraxion and IDEA. It is also at the origin of two AI-related organisations : CLICK and the Deep Learning Academy.

In 2017, UMONS, along with UCLouvain and Multitel, initiated the so-called “Deep Learning Academy” with the aim of organising quarterly gatherings for dozens of researchers from academia, research centres and companies (mainly SMEs) and to share experiences and best practices around AI. An « AI Mons Meetup » group and related mailing list has also been created to facilitate local networking around AI and provide information on AI events and scientific breakthroughs.

Deep Learning Academy

UMONS has recently joined the Walloon AI network, which gathers together Walloon companies, research centers and universities. It collaborates with the network on a regular basis to provide it with research and training cartography and to better analyse regional industrial needs on that topic. UMONS is also member of the AI4Belgium coalition and has collaborated on the writing of the AI federal report.

UMONS in the AI ecosystem

Sandrine Brognaux

Technology Transfer Officer in ICT sandrine.brognaux@umons.ac.be Pierre Manneback

President of the Infortech Institute pierre.manneback@umons.ac.be

Thomas Brihaye

President of the Complexys Institute thomas.brihaye@umons.ac.be

- www.umons.ac.be/ia -

Created in 2019, the Innovation Centers associate UMONS and their long-term strategic partners. UMONS and its Innovation Centers form a unique structure in Belgium, allowing us to carry out fundamental, strategic and applied research with our partners, who are recognised for their excellence. The research carried out can benefit any socio-economic organisation, regardless of the TRL of the results.

MULTITEL : Research centre in Telecommunications, Signal and Image Processing, founded by the the Engineering Faculty of Mons.

CLICK’ LIVING LAB: Where AI meets creativity At the heart of the creative Walloon ecosystem and benefiting from the expertise of the NUMEDIART Institute (UMONS), CETIC and Multitel, CLICK’ is the chosen partner for project owners wishing to develop innovative products and services. CLICK’s methodology includes exploration, experimentation and evaluation, in the lab and in the field, and is based on co-creation and feedback from the target users (UX). It is a place of experimentation for short prototyping projects, mixing AI with sounds, images, texts, gestures, etc.



Documents relatifs


In Europe, it is reinforced by the creation of the European Higher Education Area and the Bologna Process ( www.ehea.info ), and the focus of this study, the University of

Key Words: conservation biology; genetic resources; local seeds; mode of knowledge production; native species; problem-finding; Pyrenees; seed transfer zones;

(1991) observed that adults and 1+ juveniles were concentrated in the vicinity of the of the collector canal, with the exception of Central Mudminnow which was distributed

- En s’appuyant sur les chartes existantes, et pour chaque type de risques, des procédures d’engagement dans l’expertise doivent être préparées en avance des

of elliptic curves having complex multiplication with special Γ-values (cf. Function field, Kronecker limit formula, Eisenstein series, Drinfeld modules... This work was supported

EMS a settings) with different capabilities, on the same microgrid environment with rare events: a basic DDQN agent (without any improvement), an agent with only the combined

For practical reason, this system is written using the vector potential A C of the incident laser field, the vector potential A R of the Raman component, the electric field

When the SNMP agent receives a PDU containing an SNMP GET request for a variable that a sub-agent registered with the agent, it passes an SNMP DPI GET packet to the

The routing protocol SHOULD support the wireless worker with fast network connection times of a few of seconds, and low command and response latencies to the plant behind the

This report is about recent progress on semi-classical localization of eigenfunctions for quantum systems whose classical limit is hyperbolic (Anosov systems); the main example is

But I disagree with him that this norm is more basic than the norm of knowledge (2001: 289- 90) 15 The objection is well formulated by David 2001: “Although knowledge is certainly

As previously stated, hemocyte sub-populations may be differentially involved in functions of bivalve biology and physiology, including nutrient transport and digestion,

Le deuxième régime, avec les nombres de Rossby -1, -1.5 -2.5 et -3, est surtout caractérisé par une invariance axiale de la vitesse angulaire, dans l'intérieur de la zone limitée par

or differ from a scalar density only by a pure divergence expression (the latter must be a divergence expression without assuming any relation between all field

They showed that the problem is NP -complete in the strong sense even if each product appears in at most three shops and each shop sells exactly three products, as well as in the

If an abstract graph G admits an edge-inserting com- binatorial decomposition, then the reconstruction of the graph from the atomic decomposition produces a set of equations and

This nding suggests that, even though some criminals do not commit much crime, they can be key players because they have a crucial position in the network in terms of

However, the current relative uncertainty on the experimental determinations of N A or h is three orders of magnitude larger than the ‘possible’ shift of the Rydberg constant, which

Assuming the correctness of QED calculations in electronic hydrogen atom, the problem can be solved by shifting by 5 standard deviations the value of the Ryd- berg constant, known

Historical materialism, thoroughly modernist in formation and outlook, conceives history from the point of view of the oppressed and defeated, those whose life possibilities have

“The Sublime Anthropocene.” Salvaging the category of the “formless” from Kant’s Critique of Judgment, Williston argues that sublime encounters with formless nature