This course delves into the mathematical optimization techniques essential for developing and refining machine learning algorithms and AI applications. Focusing on theoretical foundations, this course explores deep neural network initialization, gradient descent techniques, automatic differentiation and backpropagation, and adaptive learning rate algorithms such as Adam and RMSProp. Additionally, it covers principal component analysis (PCA), density estimation algorithms, and support vector machines (SVM). Students will learn to solve unconstrained and constrained optimization problems, apply these methods to neural networks, and enhance model performance. The course provides a comprehensive understanding of optimization’s role in AI, equipping students with the theoretical knowledge to tackle complex challenges in various AI domains.

Temps présentiel : 30 heures

Charge de travail étudiant : 70 heures

Méthode(s) d'évaluation : Examen final