Background
The International School on Signal Processing and its Applications has been set up in 2004 to promote signal processing research activities in developing countries, starting with the Maghreb region. This school is dedicated to signal and image processing research. It is a forum that gathers leading researchers and postgraduate students from universities and industrial groups involved in the development of new applications and/or carrying out common research projects.
The first edition of the school took place at Ecole Nationale Polytechnique (ENP), Algiers, in 2004. It has been followed by 5 successive editions organized in Mostaghanem (2005), Jijel (2006), Boumerdes (2007), Kuala Lumpur (Malaysia 2008), and Oran (2009), respectively. After that, the organisation of this thematic school has been resumed in 2017 where it has been organized at the University of Guelma followed by the last edition organized at the University of Annaba in 2019. Several areas related to signal processing are of interest and have been the focus of previous issues.
The 9-th Edition
The focus of this 9-th edition will be on Signal Processing and Artificial Intelligence. Indeed, AI became in the last recent years one of the hottest topics in data science including signal processing.
This new edition celebrates the 20th Anniversary of ISSPA at Ecole Nationale Polytechnique, Algiers, Algeria, and its first place!
Honorary General Chair
Prof. Abdelouahab MEKHALDI, Ecole Nationale Polyetchnique, Algeria
General Chair
Prof. Adel BELOUCHRANI, Ecole Nationale Polyetchnique, Algeria
Advisory Board
Prof. Adel BELOUCHRANI, Ecole Nationale Polyetchnique, Algeria
Prof. Karim ABED-MERAIM, University of Orleans, France
Prof. Nguyen LINH TRUNG, Vietnam National University, Vietnam
Organizing Committee: TBA
Submission Deadline: 15 April 2024
Acceptance Notification: 1 May 2024
Accepted candidates will present their research work at a POSTER session
Preliminary Program: TBA
Summaries and speaker’s short biographies
Speaker 1: Prof. Sid-Ahmed BERRANI, National School of Artificial Intelligence, Sidi Abdallah, Algeria
Artificial intelligence and its applications
Artificial intelligence (AI) has significantly evolved over the last decade, in particular thanks to the availability of data and thanks to the large available computing power. Its applications are very varied and affect many areas: industry, health, entertainment, defense, telecommunications, security, biometrics, etc. Developed AI-based solutions are fundamentally changing our daily lives. It is not just about optimizing operating costs or saving money, AI is creating new uses and transforming the way we work. Thanks to the use of AI in robotics and process automation, the productivity of industrial companies has been significantly improved. In the field of image/video processing, AI now makes it possible to recognize objects in images better than a human being does for many tasks. AI is also capable of making automatic skin cancer diagnosis, transcribing an audio stream to text and translating it into another language simultaneously. The objective of this lecture is to introduce the basic concepts of AI, its scientific foundations through a historical retrospective that shows the evolution of this field since the 1950s to date. This involves demystifying the concept while highlighting the assets and factors that have enabled this major and rapid evolution. Modern machine learning and deep learning techniques will also be presented. These will be illustrated through examples of real applications that show the capabilities of AI-based technologies and their use in practice.
Bibliography: Sid-Ahmed BERRANI obtained an engineering degree in computer science from the University of Sidi Bel-Abbès in 2000; a Ph.D. in computer science from the University of Rennes 1 in 2004; a “HDR” from the University of Rennes 1 in 2014. He has been a researcher at Orange Labs from 2004 to 2011. Then, from 2011 to 2015, he headed the Orange Labs R&D unit which is specialized in Artificial Intelligence and multimedia content analysis and indexing. From 2015 to 2018, he held several positions in the ICT sector in Algeria (Deputy CEO of Algérie Telecom, advisor to the CEOs of Algérie Télécom and Mobilis…). From 2019 to 2023, he has been an associate professor at École Nationale Polytechnique (Algiers), in the electronics department, and since October 2023, he is an associate professor at the National School of Artificial Intelligence – Sidi Abdallah. His research activities focus on image and video indexing, machine learning, multidimensional data analysis and Artificial Intelligence. He is author and co-author of around fifty scientific publications. He also filed 13 patents.
Speaker 2: Prof. Hacéne BELBACHIR, University of Science and Technology Houari Boumediene, Algeria
Graph theory and its applications
Our aim is to discuss the basics of graph theory and its application. Some concepts are the following: oriented and non-oriented graphs; paths, chains, cycles, circuits, trees, connectivity and planarity; Eulerian and Hamiltonian graphs; partial graphs and sub-graphs; colouring, clique and stable set; bipartite graphs; independent edge set (matching), covering and dominating set. We also discuss some modern applications of graph theory, such as mobile communications systems, wireless networks, traffic networks, navigable networks, molecular epidemiology and big data.
Bibliography: Hacene Belbachir has received the highest diploma in mathematics (D.E.S.), Option: Probability and statistics from the University of Science and Technology Houari Boumediene (USTHB). In 1991, he received the magister diploma in algebra and number theory from the USTHB, in 1994, the magister diploma in applied statistics and economy from the National Institute of Planification and Statistics (INPS, now ENSSEA), in 1996, and the doctorat d’Etat (PhD) degree, in mathematics on combinatorial number theory from the USTHB, in 2007. He was with the Department of Mathematics, USTHB since 1994 as an associate professor and from 2007 as full professor. Actually, he is the head of the Team Combinatory, Arithmetic and Theoretical Informatics (CATI Team), RECITS Laboratory, USTHB, associate member of LITIS Laboratory, Team Combinatory and Algorithms from Rouen University in France, and member of the Euro-Maghreb Laboratory of Mathematics and their Interactions (LEM2I).
Speaker 3:Prof. Eric MOULINES, Ecole Polytechnique, Paris, France
Divide-and-Conquer Posterior Sampling for Denoising Diffusion Priors
Interest in the use of denoising diffusion models (DDM) as priors for solving inverse Bayesian problems has recently increased significantly. However, sampling from the resulting posterior distribution poses a challenge. To address this problem, previous work has proposed approximations for the bias of the drift term of the diffusion. In this work, we take a different approach and utilize the specific structure of the DDM prior to define a set of intermediate and simpler posterior sampling problems, resulting in a lower approximation error compared to previous methods. We empirically demonstrate the reconstruction capability of our method for general linear inverse problems using synthetic examples and various image restoration tasks.
Bibliography:Eric Moulines earned a degree in engineering (1984) from the Ecole Polytechnique, and a PhD in electrical engineering (1990) from the Ecole Nationale Supérieure des Télécommunication. In 1990, he joined the Signal and Image Processing Department at Télécom ParisTech, where he was appointed full professor in 1996. In 2015, he joined the Centre for Applied Mathematics at Ecole Polytechnique, where he is currently Professor of Statistics. His areas of expertise include computational statistics (Monte Carlo simulations, stochastic optimization), probabilistic machine learning, statistical signal processing and time series analysis (sequential Monte Carlo, non-linear filtering). His current research topics include high-dimensional Monte Carlo sampling, stochastic optimization, and generative models (variational autoencoders, generative adversarial networks). He applies these different methods to uncertainty quantification, Bayesian inverse problems, and the control of complex systems. He has published more than 120 articles in leading journals in the areas of signal processing, computational statistics and applied probability and more than 300 proceedings papers in major signal processing and machine learning conferences. In 1997 and 2006, he received the Best paper Award from the IEEE Signal Processing Society (for publications in IEEE Trans. On Signal Processing). He has served on the editorial boards of IEEE Trans. On Signal Processing, Signal Processing, Stochastic Processes and Applications, Journal of Statistical Planning and Inference, Electronic Journal of Statistics. From 2013-2016, he was the Editor-in-Chief of Bernoulli. E. Moulines is a EURASIP and IMS Fellow. He was awarded the Silver Medal of the Centre National de Recherche Scientifique in 2010 and the Orange Prize of the French Academy of Sciences in 2011. In 2016, he was a Fellow of the IMS. In 2020, he received the technical achievement award from EURASIP. In 2017, he was elected to the Academy of Sciences.
Speaker 4:Prof. Karim ABED-MERAIM, University of Orleans, France
Tensor decomposition for neural network
The era of “Big Data”, which deals with massive datasets, has brought new analysis techniques for discovering new valuable information hidden in the data. Among these techniques is multilinear low-rank approximation (LRA) of matrices and tensors, which has recently attracted a lot of attention from engineers and researchers in the signal processing and machine learning communities. A tensor is a multidimensional array and provides a natural representation of high-dimensional data. Low-rank approximation of tensors (t-LRA) can be considered as a multiway extension of LRA of matrices (which are two-way) to higher dimensions. Generally, t-LRA is referred to as tensor decomposition, which allows factorizing a tensor into a sequence of basic components. As a result, t-LRA provides a useful tool for dealing with several large-scale multidimensional problems in modern data analysis, which would be, otherwise, intractable by classical methods. In this lecture, a brief review of different tensor concepts and different tensor decomposition algorithms with illustrative application examples is first provided. Then, we focus on the application of tensor decomposition for model reduction in neural network and deep learning
Bibliography:Karim Abed-Meraim was born in Algiers, Algeria, on August 10, 1967. He received the State Engineering Degree from the cole Polytechnique, Palaiseau, France, in 1990, the State Engineering Degree from the Ecole Nationale Supérieure des Télécommunications (ENST), Paris, France, in 1992, the M.Sc. degree from Paris XI University, Orsay, France, in 1992, and the Ph.D. degree in the field of signal processing and communications from ENST, in 1995. From 1995 to 1998, he was a Research Staff with the Electrical Engineering Department, The University of Melbourne, where he worked on several research project related to Blind System Identification for Wireless Communications, Blind Source Separation, and Array Processing for Communications. From 1998 to 2012, he has been an Assistant Professor, then an Associate Professor with the Signal and Image Processing Department, Télécom ParisTech. In September 2012, he joined the University of Orleans, France (PRISME Laboratory), as a Full
Professor. His research interests include signal processing for communications, adaptive filtering and tracking, array processing, and statistical performance analysis. He is the author of more than 500 scientific publications, including book chapters, international journal and conference papers, and patents. Dr. Abed-Meraim is a Senior Area Editor of the IEEE Transactions on Signal Processing. Prof. Karim Abed-Meraim has been named a 2019 IEEE Fellow.
Speaker 5:Prof. Mèrouane DEBBAH, Khalifa University, UAE
Large Language Models: Challenges and Opportunities
In recent years, the development and deployment of Large Language Models (LLMs) have revolutionized various fields across academia and industry. These models, with millions or even billions of parameters, have unlocked new possibilities in natural language understanding and generation. This talk delves into the theory behind LLMs and explores their diverse applications, shedding light on their impact on our digital world. The first part of the talk will focus on the theoretical underpinnings of LLMs. We will discuss the architectural foundations, including attention mechanisms and self-attention, as well as the training methodologies, such as pre-training and fine-tuning. Understanding the theoretical framework of LLMs is essential for appreciating their capabilities and limitations. The second part of the talk will showcase the wide-ranging applications of LLMs across various domains with a special focus in the telecom domain. The talk should provide a comprehensive overview of Large Language Models, bridging the gap between theory and practice. By the end of the talk, attendees will gain a deeper understanding of LLMs’ theoretical foundations and their transformative potential in an array of applications.
Bibliography: Mérouane Debbah is a researcher, educator and technology entrepreneur. Over his career, he has founded several public and industrial research centers, start-ups; and is now Professor at Khalifa University of Science and Technology in Abu Dhabi and founding Director of the Khalifa University 6G center. He is a frequent keynote speaker at international events in the field of telecommunication and AI. His research has been lying at the interface of fundamental mathematics, algorithms, statistics, information and communication sciences with a special focus on random matrix theory and learning algorithms. In the Communication field, he has been at the heart of the development of small cells (4G), Massive MIMO (5G) and Large Intelligent Surfaces (6G) technologies. In the AI field, he is known for his work on Large Language Models, distributed AI systems for networks and semantic communications. He received multiple prestigious distinctions, prizes and best paper awards (more than 35 best paper awards) for his contributions to both fields and according to research.com is ranked as the best scientist in France in the field of Electronics and Electrical Engineering. He is an IEEE Fellow, a WWRF Fellow, a Eurasip Fellow, an AAIA Fellow, an Institute Louis Bachelier Fellow and a Membre émérite SEE. His recent work led to the development of NOOR (upon it release, largest language model in Arabic) released in 2022 and Falcon LLM (upon its release, top ranked open source large language model) released in 2023. These two models have positioned the UAE as a global leader in the generative AI field.
Speaker 6: Prof. Adel BELOUCHRANI, Ecole Nationale Polytechnique, Algeria
Source Separation: Model driven versus Data driven approaches
Blind signal separation is now a mature field of research with a broad range of applications. It is motivated by practical problems that involve several source signals and several sensors. Each sensor receives a mixture of the source signals. The problem under consideration consists of recovering the original waveforms of the source signals without any knowledge of the mixture structure. The latter may be instantaneous linear mixture, convolutive mixture or nonlinear mixture. This talk concentrates on Model driven approaches with some applications in various fields of engineering. A discussion on Data driven approaches versus Model driven ones will be provided.
Bibliography:Adel Belouchrani was born in Algiers, Algeria, on May 5, 1967. He received the State Engineering degree from Ecole Nationale Polytechnique (ENP), Algiers in 1991, an MSc degree in signal processing from the Institut National Polytechnique de Grenoble, France in 1992, and his PhD in signal and image processing from Télécom Paris, France in 1995. He was a postdoctoral fellow at the Electrical Engineering and Computer Sciences Department, University of California at Berkeley, CA, USA, from 1995 to 1996 and was with the Department of Electrical and Computer Engineering, Villanova University, Villanova, PA, USA, as a research associate, from 1996 to 1997. From 1998 to 2005, he was with the Electrical Engineering Department, ENP, as an associate professor. Since 2006, he has been a full professor with ENP. His research interests are in statistical signal processing, (blind) array signal processing, time-frequency analysis, and time-frequency array signal processing with applications in biomedical and communications. Dr. Belouchrani is a founding member of the Algerian Academy of Science and Technology. Dr. Belouchrani was awarded an Arab-American Frontiers Fellowship of the U.S. National Academy of Sciences, Engineering and Medicine to conduct research on the blind identification of power sources in processors design at Brown University, Rhode Island, USA in 2016. He has served as Associate Editor of the IEEE Transactions on Signal Processing for two terms from 2013 to 2017 and as Editorial board member of the Digital signal processing Journal (Ed. Elsevier) from 2011 to 2019. He has been Senior Area Editor of the IEEE Transactions on Signal Processing for two terms from 2017 to 2021 and an elected member of the EURASIP Special Area Team (SAT)-Theoretical and Methodological Trends in Signal Processing for the 2018-2020 term, as well as an elected member of the IEEE Sensor Array and Multichannel Technical Committee for 2019-2021 term. Prof. Adel Belouchrani has been named a 2020 IEEE Fellow.
Speaker 7: Prof. Nguyen LINH TRUNG, VNU, Vietnam
Signal and Information Processing Methods for Brain Studies
Recent advances in signal processing and machine learning have led to opportunities and challenges in brain studies. In this talk, we will present several research studies targeting three brain conditions: epilepsy, cognitive conflict, and Alzheimer’s disease. The first study, for epilepsy, develops a multi-stage system for single-channel EEG-based epileptic spike detection and a tensor decomposition method for multi-channel spike detection. The second study selects special computer-vision methods for EEG-based peak detection in cognitive conflict processing. The third study applies machine-learning methods for PET-based feature extraction and data-driven brain atlas construction in classification of Alzheimer’s disease. The fourth study applies machine-learning methods for MRI-based feature extraction in prediction of pathways between multivariate brain areas and multivariate disease/behavior outcomes.
Bibliography: Nguyen Linh Trung obtained his B.Eng. and Ph.D. degrees, both in Electrical Engineering, from Queensland University of Technology, Brisbane, Australia, in 1998 and 2005. He joined VNU University of Engineering and Technology, Vietnam National University, Hanoi (VNU), in 2006, where he is currently an associate professor of electronic engineering and founding director of the Advanced Institute of Engineering and Technology. His broad technical interests are theories, methods and applications of signal processing; in particular, detection and estimation, adaptive processing, subspace and tensor methods, blind separation, time-frequency analysis, graph processing, machine learning, and their applications in communications, networks and biomedicine. He has served as technical editor-in-chief of the Journal of Research and Development on Information and Communication Technology published by the Ministry of Information and Communication of Vietnam, founding chair of the IEEE Signal Processing Vietnam Chapter, founding chair of the Asia-Pacific Signal Processing and Information Vietnam Chapter general co-chair of the 2023 IEEE Statistical Signal Processing workshop.