In this course, the students will learn about the relationship between data, models, and algorithms to understand how to process and draw conclusions from data through data mining and machine learning. The course introduces some theory on machine learning but focuses mainly on current applied methods. Successful machine learning applications need to be designed through a critical engagement and understanding of data, the algorithms that can be applied based on the kind of features the data exhibits, and choosing the right paradigm of machine learning. This course provides a fundamental basis for using machine learning in an ethical and responsible manner. What are the predominant paradigms in machine learning, and in what situations are they best used? What perspectives should we consider when we design machine learning applications? Why is a critical perspective important for developing machine learning?
During the course, participants will set up their own AI model and train it on either their own (if available) or a provided dataset. For more information about the course, please contact Tobias Nyberg (tnybe@kth.se).
The following topics are included in the course:
• Statistical and probabilistic methods for data analysis.
• Different methods for data mining.
• Algorithms for supervised and unsupervised machine learning.
• Neural networks and deep learning.
• Data extraction: purpose and typical use cases.
• Routines for importing, combining, converting and selecting data for learning and validation.
• Validation methods and performance measures.
• Ethics and regulations concerning use and processing of personal data.