Speaker: Dr. Andrea Manno
Dates: July 12, 13, 19 and 20
Time: 16:00-18:00
MS Teams code: 9kkf5ny
Outline of the course (8 hours)
(4 hours) Introduction to nonlinear optimization
- unconstrained optimization: optimality conditions, algorithms, con-
vergence and regularization - hints on constrained optimization
(2 hours) Nonlinear optimization for Neural Networks training
- a loss function minimization problem and the backpropagation
- adapting general algorithms and regularization for neural networks
training - advanced strategies
(2 hours) The training of Support Vector Machines as a constrained optimization problem
- linear SVMs for the linearly separable case
- a mathematical formulation for non-linearly separable datasets
- nonlinear SVMs
[COURSE] Nonlinear Optimization for Machine Learning Architectures