3.0

Average Course Rating

The course provides foundations and algorithms addressing optimization problems in modern data science. It covers smooth unconstrained descent methods (including deterministic and stochastic gradient), smooth constrained optimization and some non-smooth situations in the convex case. Each of these optimization problems will be related to specific data science training algorithms (such as logistic regression, neural networks, support vector machines or lasso). Prerequisites include multivariable calculus and linear algebra. Homework will include some programming components and students will be expected to have basic proficiency in computer languages such as Python (preferred) or Matlab. Recommended Course Background: Multivariable Calculus and Linear algebra.