Optimization methods are the primary language for describing most sections of modern applied mathematics - from determining the minimum of a scalar function to finding the best parameters of a neural network in a huge dimensional space.
At the course we will:
build an optimal investment portfolio
build a recommendation system of choosing TED talks as a solution to a linear programming problem
consider different regularization techniques in machine learning problems
look at what humanity knows so far about neural network training.
We will use various libraries and frameworks for machine learning and optimization in Python - Cvxpy, Scipy, Jax, PyTorch, Autograd, Keras, Optuna, Pyomo.
Course instructor: ΠΠ»Π΅ΠΊΡΠ΅ΠΉ ΠΡΡΠ»ΠΎΠ²
The course is a presentation of modern results and approaches in solving applied optimization problems. Despite the focus on applications, the course contains the necessary theoretical background to understand why and how specific methods work.