This course teaches future AI engineers to build, train and deploy data-driven systems that are lawful by design, using the EU GDPR as the reference framework and Algeria's Law 18-07 (as amended by Law 25-11 of 24 July 2025) as the national anchor, compared with the GDPR.

    This course introduces the fundamental concepts and methods of Reinforcement Learning (RL), a branch of machine learning in which an agent learns to make decisions by interacting with an environment and receiving rewards or penalties.

    Students will study the main components of reinforcement learning, including states, actions, rewards, policies, value functions, and Markov Decision Processes (MDPs). The course covers the mathematical foundations of RL, policy evaluation and improvement, dynamic programming, Monte Carlo methods, and temporal-difference learning. It also introduces important algorithms such as SARSA and Q-Learning, as well as advanced approaches including Deep Q-Networks (DQN), policy gradient methods, actor–critic methods, and multi-agent reinforcement learning.

    Through lectures, tutorials, and practical sessions, students will develop the theoretical understanding and practical skills needed to model sequential decision-making problems, implement reinforcement learning algorithms, and evaluate their performance in artificial intelligence applications.

    This course aims to introduce students to the paradigm of Constraint Programming, widely used to model and solve complex combinatorial problems. The objectives are:

    1. Understand the concept of a constraint satisfaction problem (CSP).
    2. Learn to model a real problem in terms of variables, domains, and constraints.
    3. Study the main resolution techniques (backtracking, filtering, heuristics).
    4. Apply constraint programming to concrete cases: scheduling, timetabling, graph coloring, optimization.

    This advanced course offers a comprehensive journey from traditional pattern recognition pipelines to state-of-the-art end-to-end deep learning representations. Students will explore complex feature extraction, non-linear dimensionality reduction, and self-supervised contrastive learning, before advancing to sophisticated clustering techniques and cutting-edge visual architectures, including modern CNNs and Vision Transformers (ViT). The curriculum also delves into generative models (GANs, VAEs) for synthetic data augmentation, alongside critical frameworks for advanced model evaluation, adversarial robustness, and Explainable AI (XAI). Supported by hours of intensive practical labs, learners will gain hands-on experience implementing these theories—from fine-tuning EfficientNets and analyzing ViT attention maps to deploying interpretability tools like Grad-CAM—equipping them to build, evaluate, and interpret highly robust visual recognition systems.

    This course introduces metaheuristics as powerful tools for solving complex optimization problems (NP-hard, non-linear, multimodal) where exact methods fail. You will study evolutionary algorithms (genetic algorithms, evolution strategies, genetic programming, NSGA-II and SPEA2) and other nature-inspired methods (simulated annealing, PSO, ant and bee colonies). You will also see how metaheuristics are combined with Machine Learning and Deep Learning for hyperparameter tuning, neural architecture search and feature selection. 

    This course offers a comprehensive exploration of the theoretical foundations and practical applications of Generative Artificial Intelligence. Designed to cultivate both rigorous mathematical understanding and hands-on engineering capabilities, the curriculum covers the core paradigms driving modern AI generation—including Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), Diffusion Models, and Large Language Models (LLMs).

    Students will gain experience designing, implementing, and evaluating generative architectures across diverse data modalities—such as computer vision, natural language processing, audio synthesis, and multimodal integration. Beyond technical implementation, the course critically examines pressing societal implications, including ethical considerations, security vulnerabilities, intellectual property challenges, and protocols for responsible deployment.

    Academic Master’s Degree
    Specialty: Artificial Intelligence
    Field: Computer Science

    Course Title: Natural Language Processing (NLP)
    Semester: 3
    Type: UEM

    Total Teaching Hours (VHS): 63 hours
    Weekly Teaching Hours (VHH): 3 hours
    Lectures: 1h30
    Tutorials (TD): 1h30
    Practical Work (TP): 1h30

    Personal Study Hours: 42 hours
    Coefficient: 2
    Credits: 5