Course content

Data representation: representation of various data sources such as images, sound and text, current techniques for processing data.

Unsupervised learning: various clustering algorithms, reduction of dimensions, and other current methods.

Supervised learning: including logistic regression and different types of neural networks.

Learning outcome

Knowledge

The candidate can give an account of:

• different ways of representing data
• different methods for grouping and classifying data
• which machine learning methods are appropriate to use for given problems
• limitations of machine learning

Skills

The candidate can:

• create full-fledged machine learning solutions using a framework
• use representation algorithms that make it easier for machine learning methods to give better results for a given data set
• select and adapt a machine learning method that is relevant to a given problem
• assess whether machine learning methods can give good results for a given problem based on a given data set

General competence

The candidate must be able to find and adapt solutions to new problems based on previous applications of machine learning.

Learning methods and activities

Lectures, exercises and project.

Compulsory assignments

• Mandatory exercises

Specific conditions

Recommended previous knowledge

Linear algebra and statistics.

Course materials

Programming examples, presentations and operating instructions with auxiliary literature through external resources.

Credit reductions