This course is moved to Autumn/2022

This two-day course gives a fundamental approach to machine learning models, towards designing an interpretable model to a given problem. From key statistical, probabilistic and computational principles, the course provides heuristics on how and when to choose a particular machine learning approach to a problem. The course focuses on supervised and unsupervised approaches, and model selection.

The course is organized in hybrid fashion: on site and zoom. Hands-on exercises will be done using the Python language in CSC Notebooks environment (https://notebooks.csc.fi/).

Learning outcomes: To obtain ideas on what to look out for when a given problem can be solved using supervised or unsupervised learning tools, and focus on designing interpretable models.

Prerequisites: Basics of the Python language is assumed. Additionally basic notions of statistics and probability will be beneficial, however basic notions will be explained as methods and approaches are introduced.

Schedule:

both days from 09:00 to 16:00

Day 1

Course introduction
Supervised Learning
Support Vector Machines
Artificial Neural Networks

Day 2

Nearest Neighbor
Ensemble methods
Model selection
Unsupervised learning