Machine Learning & Energy


Teacher: Prof. Dr. Florian Steinke , Tim Janke

Content: Digitalization makes the analysis and interpretation of data ever more important, also for engineers. Smart Grids are a term to describe a host of novel data-based services in the field of generation, distribution, consumption, and marketing of (renewable) energy. The lecture presents the recent developments and the underlying machine learning methods.

For a start we will describe in a structured way the different problem areas that machine learning method deal with (classification, regression, clustering, dimensionality reductions, time series models, …). For each problem setting we will present applications in the energy sector (prediction of renewables, fault detection and prediction, data visualization, robust investments decisions, probabilistic load flow, …).

After such an introductory overview we will review probability distributions and probabilistic graphical models in detail. Based on these ideas we will study for each problem class one or more algorithms in detail (e.g.  SVMs, Deep Learning, Collaborative filtering,…) and examine their applications in the energy domain.

Practical exercise with Matlab will deepen the understanding and support student’s actively useable skills.

After the course, students should understand important machine learning problem settings and know some key methods for each task. Moreover, the students will be able to apply those methods independently to new applications (not only from the energy domain).

Type of course: 2 SWS lecture+  1 SWS exercise + 1 SWS practical session

Each week an assignment with mostly programming tasks will be given. Students should first try to solve the tasks at home. They can then continue development during the practical session under the supervision and in interactive collaboration with the teacher. In the end an exemplary solution will be presented.

Language: English

Dates: Tue 9:50-11:30, Thu 13:30-16:05

Exam: Thu 27.02.2020, 14:00-16:00  in room S306/051


  • K.P. Murphy: Machine Learning. A Probabilisitic Perspective
  • C.M. Bishop: Pattern Recognition and Machine Learning
  • J. Friedman, T. Hastie, R Tibshirani: The elements of statistical learning (
  • D. Koller, N. Friedmann: Probabilistic Graphical Models. Principles and Techniques

Requirements for successful participation:

Good knowledge of linear algebra and the foundations of numerical optimization (e.g. from the course 18-st-2010 Energieanagement & Optimierung) are required. Using Matlab for programming the practical examples should pose no difficulty.

TUCAN ID: 18-st-2020

Print | Legal note | Privacy Policy | Sitemap | Search | Contact
to topto top