Search results

Abstract Predictive maintenance (PdM) is increasingly being used to improve system reliability, reduce downtime, and optimize operating costs in complex industrial environments through anomaly detection. As industrial systems become more complex, with components interacting with one another and monitored by ever-expanding sensor networks, anomaly detection faces new methodological challenges. In particular, there is a gap in the literature regarding anomaly detection using traditional machine learning (ML) methods based on a Single Label Classification (SLC) approach, which is unsuitable when multiple failures can occur simultaneously.Traditional anomaly detection approaches based on ML methods using an SLC formulation assume that each observation belongs to a single anomaly category. This assumption does not hold when multiple failures occur simultaneously on the same equipment, resulting in a loss of information. We begin by illustrating this limitation through a concrete example showing that using an SLC approach instead of a multi-label formulation degrades the classification of simultaneous anomalies. To address this issue, anomaly detection is reformulated as a Multi-Label Classification (MLC) problem, enabling the detection of multiple failures within a single time step. Following this reformulation, the issue of selecting multi-label classifiers suited to the context of Predictive Maintenance (PdM) must be examined. The general conclusion from the literature is that the performance of classifiers depends heavily on the context and that while there are many detection methods, none is universally dominant. Although in-depth comparative studies exist in the field of MLC, datasets from complex systems are absent from these comparisons, even though they exhibit specific characteristics such as multivariate time-series structures, significant label imbalance, and complex interactions between components.To address this gap, a structured and reproducible evaluation protocol is proposed to evaluate eight state-of-the-art ML methods in an MLC framework across three public industrial datasets. The results confirm that the performance of classifiers depends heavily on the characteristics of the dataset and that no single method consistently outperforms others across all scenarios. However, conducting this type of comparison for each new industrial context is costly in terms of time, computational resources, and data requirements, which limits its feasibility for real-world industrial deployment. The thesis also addresses this issue through dimensionality reduction and variable selection methods. PM systems generate large multivariate datasets in which only a subset of the variables is actually relevant for anomaly detection. A self-adaptive evolutionary strategy is proposed to perform wrapper-style variable selection and obtain a subset containing only the most informative variables. Experimental results show that reducing the variable space improves computational efficiency and, in most cases, anomaly prediction performance. A comparison of the proposed method with state-of-the-art metaheuristic approaches on three PdM datasets and five anomaly detection methods confirms its competitiveness in terms of predictive performance and the optimal size of the selected feature subset. Finally, feature space reduction not only optimizes the resources required for detection but also reduces the number of signals that technicians must analyze during root cause investigations. This simplification aims to strengthen confidence in the system. Finally, the thesis addresses the practical aspects of deploying detection methods in complex systems. Among these, there is a lack of trust and transparency among maintenance technicians and decision-makers. An analysis of the interpretability and robustness of anomaly detection methods is provided to directly contribute to their deployment in real-world conditions. First, the proposed variable selection strategy is evaluated on a public PdM dataset using SHAP-based explanations to verify the consistency between the selected variables and those identified as the most important by tools dedicated to interpretability. The results show that the selected variables are consistent with those identified as the most important by these tools. Furthermore, visual analysis of the most influential variables reveals that anomalies are associated with a limited and well-defined subset of sensor signals, which helps improve transparency and strengthen confidence in automatic detection systems. At the same time, it is necessary to conduct a study of the robustness of multi-label classifiers in controlled sensor degradation scenarios, including different types of degradation and various levels of severity. The results show that the performance of classifiers can deteriorate significantly, particularly in the presence of gradual drifts. To mitigate this effect, introducing sensor degradation during training appears to be a relevant strategy. Finally, this thesis proposes a structured framework for multi-label anomaly detection in complex systems, covering classification methods, feature space optimization, interpretability, and robustness. For each of these aspects, the thesis adopts an industrial perspective and takes into account the differences between controlled laboratory experimental conditions and the additional constraints of real-world complex environments. The jury Prof. Katrien Beuls - University of Namur, BelgiumProf. Elio Tuci - University of Namur, BelgiumProf. Patrick Heymans - University of Namur, BelgiumProf. Jenni Raitoharju - University of Jyväskylä, FinlandMr. Fabio Pinna - Telespazio, BelgiumMr. Vito Trianni - ISTC-CNR, Italy Sign me up

Défense de thèse - Antoine Hubermont 04/09/2026 à 15h à l'auditoire PA02. Name First name E-mail address Will attend the reception following the defense Yes ( optional ) No ( optional ) Need a parking sticker Yes ( optional ) No ( optional ) Would like a certificate for defense assistance Yes ( optional ) No ( optional ) In order to process your request, you must complete all fields marked "optional". When you submit this form, the completed data will be transmitted to UNamur and used to process your request. Learn more about your data protection and your rights This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Faced with the ecological crisis and soaring energy prices, the energy transition has become an undeniable emergency. Every day, at UNamur, researchers from a wide range of fields - geology, chemistry, physics, computer science - are thinking about innovative ways of dealing with this perilous future.

On January 23 and 24, 2025, UNamur experts were present at the SETT (School Education Transformation Technology) trade show for its sixth edition. A must-attend event for digital education in the Wallonia-Brussels Federation, dedicated to principals, teachers and technical-pedagogical advisors.

During the Japanese Emperor’s state visit to Belgium, two professors from the University of Namur had the opportunity to represent their institution at two highlights of the official program. Isabelle Parmentier and Alexandre Mauroy reflect on an experience they won’t soon forget.

On 20 June 2023, the F.R.S.-FNRS published the list of winners of the various doctoral and post-doctoral mandates. Among them, 16 researchers from UNamur obtained funding.

Jamie Wright, a researcher at the Namur Digital Institute at UNamur, participated in a study conducted by Pompeu Fabra University (UPF) in Barcelona and published in the prestigious journal *Nature Human Behaviour*.  The study shows that languages around the world tend towards lexical economy, reusing words to denote different concepts when doing so does not lead to communicative confusion.

Detecting a failure before it occurs: that is the goal of the research being conducted by Antoine Hubermont, a doctoral student at UNamur. This project, named Monsater, is funded by SPW Research as part of the Win4Doc program in collaboration with the space company Telespazio Belgium. It addresses a key strategic challenge: ensuring the reliability of complex systems, particularly in the space sector. 

Intended primarily for master’s and doctoral students, this program will aim to guide participants through a step-by-step process of mastering the fundamentals of diplomatic science. To this end, it will combine lectures delivered by members of the team, opportunities to meet and exchange ideas with internationally recognized specialists in the field, and practical exercises using unpublished documentary collections held at the State Archives in Namur. More information

Abstract Deep learning has become a central part of everyday life, and has given us powerful digital tools such as machine translators, voice assistants, and large language models. The Deaf communities, however, benefit verylittle from this progress. Sign languages are full natural languages, but they have no widely used written form, so the large text corpora and pretrained models that support these tools for spoken languages have no direct equivalent. As a result, sign language processing is held back by two related problems: a persistent shortage of annotated data, and a lack of tools to analyze the languages themselves.This thesis focuses on one component of the sign language processing pipeline that has received comparatively little attention: sign language segmentation, the temporal partitioning of a continuous signing video into individual sign units. Segmentation sits between raw video and almost every symbolic task built on top of it, yet it remains poorly understood. The goal of this work is twofold: to advance the segmentation task itself, and to turn sign language processing research into tools that are usable by the Deaf community, with French Belgian Sign Language (LSFB) as the main setting.The central contribution of this thesis is a set of segmentation models. We first study approaches based on recurrent neural networks, and show that their main difficulty lies in the modeling of the transitions between signs, the brief ambiguous movements that separate one sign from the next and that are easily confused with the signs themselves. This insight motivates the Hydra framework, a new approach that detects each sign as a whole unit rather than deciding, frame by frame, whether a sign is being produced. Concretely, instead of classifying every frame, the model predicts the distance from each frame to the nearest sign boundaries, which makes it less sensitive to the ambiguous regions and improves consistently over prior methods. Around this core, we present further contributions: a sign language-to-text dictionary that recognizes signs from an ordinary webcam, supporting work on isolated sign recognition, and a collaborative platform that gathers new sign data as a side effect of everyday use. Together, these contributions narrow the gap between sign language processing research and the people whose language it concerns. Jury Prof. Benoit Frénay - University of Namur, BelgiumProf. Anthony Cleve - University of Namur, BelgiumProf. Katrien Beuls - University of Namur, BelgiumProf. Barbara Hammer - Bielefeld University, GermanyProf. Michèle Gouiffès - Paris-Saclay University, FranceDr. Mathieu De Coster - Ghent University, Belgium I am registering to attend Pierre Poitier's thesis defense

September 14, 2026 What's on the schedule for everyone9:00 a.m. | Welcome at Pedro Arrupe, Rue de Bruxelles 67, 5000 Namur9:30 a.m. | Ceremony at Pedro Arrupe 11:00 a.m. | Back-to-school celebration at Saint-Loup Church—to be confirmed, rue du Collège, 5000 Namur—followed by a welcome for students by the student clubs. Today's Rates September 15For Block 1 (Room I02)* - Welcome Session10:40 a.m.: Introduction by the Dean/Associate Dean (Anthony Cleve - Marie-Ange Remiche)11:10 a.m.: Introduction by the academic advisor (Géraldine Grandjean)11:30 a.m.: Introduction by the Academic Coordinator (Fanny Boraita)11:50 a.m.: Introduction to the Student Services Office (Cédric Aerts)12:10 p.m.: Introduction to CSLabs (Hugo Raskin) For students in the first 60 credits of the bachelor’s program (first-year students only*) – Room I022:00 p.m.: Introductory session for the English courseAttendance at these sessions is mandatory.For UES** students and new Master’s students (Room I30) – Welcome session2:00 p.m.: Introduction by the Dean/Associate Dean (Anthony Cleve – Marie-Ange Remiche)2:30 p.m.: Introduction by the academic advisor (Géraldine Grandjean)2:50 p.m.: Introduction by the Academic Coordinator (Fanny Boraita)3:10 p.m.: Introduction to the Student Services Office (Cédric Aerts)3:30 p.m.: Introduction to CSLabs (Hugo Raskin) September 16 For Block 1 (Room to be determined—check the schedule)*8:30 a.m.: Mathematical Foundations for Computer Science course (M. De Vleeschouwer)Attendance at this session is mandatoryFor all students: Classes begin (see schedule)Information session and PAE setupSeptember 18: Bachelor’s Program, Block 1 (non-first-year students): 1:00–2:00 p.m. – Academic HallSeptember 17: Bachelor’s Program, Blocks 2 and 3: 1:00–2:00 p.m. – Academic HallSeptember 16: MA1 and UES (non-first-year students): 1:00–2:00 p.m. in I30* First-time students: Students enrolled for the first time in a computer science program at UNamur, whether they are coming from high school, a college, another university, or are enrolled in a staggered-schedule program. ** UES: Additional course units for the master’s program (bridge year) Staggered-Schedule Classes Bachelor's and Master's 60Saturday, September 12—Classes BeginFor First-Year Students (Block 1 and UES):9:00 a.m.: Presentations in Room I02 by the Associate Dean, Ms. Marie-Ange REMICHE; the academic advisor, Ms. Géraldine GRANDJEAN; and the IT coordinator, Mr. Cédric AERTS. Lecture Hall I02 (ground floor of the School of Computer Science). Attendance at this session is mandatory. The presentation from the orientation session will be posted on the BVE afterward.10:00 a.m.: Classes begin for all students Specialized Master's Degree in Computer Science and Innovation: Business Analysis and IT Governance For students who need to take prerequisite coursesFriday, September 18, at 9:00 a.m., Seminar Room I22 on the 2nd floor of the Faculty building.For all new studentsClasses begin on Saturday, October 16, in the academic hall on the 4th floor of the School of Computer Science, starting at 8:30 a.m. And before school starts? Give yourself the best possible chance—enroll in preparatory courses!Preparatory courses allow you to get a taste of university-level education by reviewing the subjects essential to your future program of study.Designed exclusively for students graduating in 2026, these preparatory courses—scheduled between mid-August and early September—are tailored to each university program.Learn more about the schedules for the different sessions and register for the preparatory courses...  To make the most of your first semester at the university, be sure to attend the orientation days!The orientation days will take place on September 10 and 11, 2026.On the agenda: a tour of the campus and the city, a student housing fair (KàPs), themed workshops (student rights, digital tools, budgeting, student engagement, sports, and more), a barbecue, and a super-enthusiastic team ready to help you discover life on campus!Thursday, September 10: 7:30 a.m. – Student check-in; 8:30 a.m.–12:40 p.m. - Workshops for students; 12:40 p.m. – Lunch – Barbecue available by registration (€3.60 per sausage sandwich / falafel sandwich) and a bar run by the AGE; 2:00–6:00 p.m. - Project-based student housing fair; 8:30 p.m.–1:00 a.m. – Bunker reserved for first-year studentsFriday, the 11th: 8:30 a.m.–12:30 p.m. - Workshops for students; 12:30 p.m. - Lunch; 2:00 p.m. - Depending on the faculty: organized by the faculty or open to all.Reserved for new students - Information will be sent after registration at UNamur.   

How can anomaly detection on a production line be improved? It was in response to a very concrete problem encountered at the company where he works that Arnaud Bougaham, with the support of his employer, decided to devote his thesis to this topic at the UNamur School of Computer Science. The goal? To develop an artificial intelligence model that assists operators in detecting anomalies on an industrial production line by reducing false alarms and making the diagnosis easier to understand. This approach also holds promise for deployment in the medical field.