The Third Edition of the Study Days on Advances of AI and its Societal and Industrial Applications

Biography

Ahmed Guessoum is a professor of computer science and artificial intelligence at the National Higher School of Artificial Intelligence (ENSIA). After a 5-year degree in Computer Science from the University of Science and Technology Houari Boumediene (USTHB), he obtained a Master’s in CS from the University of Southern California in Los Angeles, and then a PhD in Artificial Intelligence from the University of Bristol in the United Kingdom. He represented Algeria as an AI expert at the 17th Meeting of Arab Ministers of Higher Education and Scientific Research and at two UNESCO intergovernmental expert meetings on the Recommendation on the Ethics of Artificial Intelligence.

Plenary Talk 1 & 2

Understanding Generative AI (Part 1 & 2)

Abstract: Generative Artificial Intelligence has rapidly transformed the field of AI, enabling machines to generate coherent text, images, code, and multimodal content with unprecedented fluency and adaptability. This plenary talk provides a structured introduction to the conceptual and technical foundations underlying modern Generative AI systems, with particular emphasis on Large Language Models (LLMs). The talk begins by revisiting vector semantics and word embeddings, examining the Transformer architecture, and exploring pretraining, autoregressive generation, and RAG architectures.

Plenary Talk 3

Towards Transforming Education in Algeria Using Large Language Models

Abstract: The talk explores the growing role of Artificial Intelligence in transforming the education sector and reshaping modern learning environments. It begins with an overview of the major benefits of AI in education, including personalized learning, intelligent tutoring systems, and adaptive assessment. Particular emphasis is placed on NLP applications, discussing the experience of teaching NLP at ENSIA, and addressing the specific challenges facing AI integration in education in Algeria.

Biography

Dr. Nassima DIF is an Associate Professor at the Higher School of Computer Science ESI-SBA, where she teaches several Artificial Intelligence courses, including Machine Learning and Deep Learning. She is actively involved in training engineering students and supervising research projects in AI and data science. Her research focuses on Artificial Intelligence, Deep Learning, computer vision, multimodal learning, explainable AI, and intelligent data analysis.

Plenary Talk

Recent Advances in Deep Learning: From Medical Imaging to Deepfake Detection

Abstract: Artificial Intelligence has seen significant progress in recent years, due to advances in deep learning and the availability of large-scale datasets and powerful computational resources. This presentation explores recent trends in deep learning across different application domains, ranging from medical imaging to deepfake detection. Special attention is given to modern architectures such as Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), attention mechanisms, and generative models, which have greatly improved the performance of modern AI systems.

Biography

Professeure à l’Université Oran 1 au sein du département d’informatique, elle est également cheffe de l’équipe AIR du Laboratoire d’Informatique d’Oran. Depuis 2024, elle est responsable de la Maison de l’IA. Ses travaux de recherche portent sur le traitement du langage naturel, l’apprentissage automatique, l’analyse de sentiments et leurs applications dans le domaine médical. Elle participe à deux projets de recherche mixtes menés en collaboration avec l’EHU d'Oran et le CHU d'Oran.

Plenary Talk

Analyse des usages de l’IA générative dans l’enseignement supérieur : cas de l’Université Oran 1

Abstract: Cette étude vise à analyser les différents usages de l’IA générative au sein de l’université Oran 1 auprès des enseignants, chercheurs, personnels administratifs et étudiants et en considérant plus particulièrement ceux qui sont en préparation de projets de fin d’études ainsi qu’aux porteurs de projets innovants et de startups. La méthodologie adoptée repose sur une enquête quantitative par questionnaire diffusé auprès des différentes composantes de l’université Oran 1 afin d'analyser les différences d'usage et de formuler des recommandations.

Biography

Dr. Skander HAMDI is an Associate Professor, AI Researcher, and Senior Mobile Engineer specializing in Machine Learning and Deep Learning architectures. Holding a Ph.D. in Intelligent Systems, his published research spans speech recognition, acoustic signal processing, and medical computer vision. Dr. Hamdi’s work focuses on utilizing attention mechanisms, transfer learning, and ensemble methods to solve complex, high-impact pattern recognition tasks.

Workshop 1

Deep Learning for Image Classification and Model Explainability

Abstract: This workshop introduces the fundamental concepts of Deep Learning through a practical image classification task using Convolutional Neural Networks (CNNs). Participants will learn the complete workflow for building, training, and evaluating a deep learning model on image datasets (CIFAR-10 or Fashion-MNIST) using TensorFlow/Keras and Google Colab. A special focus will be placed on model interpretability (Grad-CAM, feature maps).

Workshop 2

Deep Learning for Sentiment Analysis and NLP Explainability

Abstract: This workshop presents the foundations of Deep Learning for Natural Language Processing (NLP) through a practical sentiment analysis application using recurrent neural networks (RNNs/LSTMs). Participants will build and evaluate a text classification model using the IMDB movie reviews dataset. The session covers text preprocessing, word embeddings, sequence representation, and SHAP/LIME tools.