Course content

Mathematical foundations of estimation and learning. Optimization methods in context of machine learning. Machine learning algorithms. Data science workflow, including feature engineering, model tuning. Bayesian methods and probabilistic programming. Integration of predictive models into cyber-physical production systems.

Learning outcome

Knowledge

• The candidate is in the forefront of knowledge within the fields of the selected machine learning methods. The candidate can evaluate the expediency and application of machine learning methods in manufacturing-oriented research and development projects.

Skills

• The candidate can formulate machine learning problems and plan the associated data science workflows.
• The candidate can implement computational routines for data preprocessing, feature engineering, and training and tuning of predictive models.

General competence

• The candidate has the ability to communicate and lead discussions on recent research about machine learning methods and their application to manufacturing-related problems.
• The candidate has the ability to evaluate and critique methods for data preprocessing, feature engineering, and machine learning.

Learning methods and activities

The course is based on regular on-campus seminars centered around the individual topics and backed by the corresponding reading materials. Every seminar will contain an in-depth group discussion.

Coursework requirements:

• On a given topic, prepare and give one seminar consisting of an introductory lecture followed by a group discussion.
• Attend at least 75% of the seminars.

Further on evaluation

Candidates must provide one paper where the applicability of the topics of the course to her/his own thesis work is thoroughly discussed. The paper should be in the form of a scientific paper, include examples of the application of machine learning methods, and preferably constitute a basis for a future publishable scientific paper.Grading scale: Pass/Fail

Course materials

Selected research papers and book chapters.