Dr. Italo Zoppis and Dr. Riccardo Dondi
University of Milano-Bicocca, Milano, Italy
This course gives a graduate-level introduction to machine learning algorithms. While the theoretical aspects will be covered, the mathematics will be confined to the main intuitive concepts and definitions. The primary goal is to provide students with the tools and principles needed to solve both traditional and novel data science problems found in practice. In particular, the course will cover the following topics:
- Fundamental issues in Statistical Learning, and Probably Approximately Correct (PAC) learnability.
- Linear Models for Classification and Regression.
- Recurrent vs. Feed-forward Neural Networks. Back-propagation.
- Discovering Structures. Graphical Models, Algorithmic aspects, and Learning in Graphs.
- Kernel Methods. Support Vector Machines (SVM), Duality, Applications and Construction of Kernel Functions. Kernelization, and Structured Input.
- Sampling Methods.
- Learning with TensorFlow. Deep Architectures.