EE4C12: Machine Learning for Electrical Engineering
Graduate course, TU Delft, Faculty of Electrical Engineering, Mathematics and Computer Science, 2024
EE4C12 is a graduate-level course designed for MSc students in Electrical Engineering at TU Delft. This course provides a comprehensive introduction to machine learning concepts, with a focus on applications relevant to electrical and energy systems. Students will gain theoretical knowledge and hands-on experience, equipping them to apply machine learning techniques effectively in real-world problems.
Course Details
- Degree Program: MSc Electrical Engineering
- ECTS: 5
- Language: English
- Education Period: Quarter 1 (early September to late October)
- Exam Periods: Quarter 1 and Quarter 2
- Contact Hours / Week: 4/0/0/0
- Prerequisites:
- Basic programming skills in any programming language.
- Elementary knowledge of probability theory and statistics.
- Instructors:
- Dr. J.L. Cremer (J.L.Cremer@tudelft.nl)
- Dr. J.H.G. Dauwels (J.H.G.Dauwels@tudelft.nl)
- Dr. P.P. Vergara Barrios (P.P.VergaraBarrios@tudelft.nl)
- Dr.ing. R.K. Bishnoi (R.K.Bishnoi@tudelft.nl)
- Dr. S.H. Tindemans (S.H.Tindemans@tudelft.nl)
Course Content
The course covers the following topics:
- Introduction to data analytics
- Regression
- Linear, Ridge, and Lasso Regression
- Gradient Descent
- Classification
- Logistic Regression
- Support Vector Machines
- Feature engineering and selection
- Feature scaling and normalization
- Principal Component Analysis (PCA)
- Imputation techniques for missing data
- Neural Networks
- Perceptron and Fully Connected Networks
- Activation Functions
- Backpropagation Algorithm
- Developing a machine learning workflow
- Geometric Deep Learning
- Graph Neural Networks
- Convolutional Neural Networks
- Tree-based methods
- Reinforcement learning
- Markov Decision Processes
- Tabular methods, SARSA, and Q-Learning
- Hardware in ML
- Overview of platforms (CPU, GPU, FPGA, ASIC)
Study Goals
By the end of the course, students will be able to:
- Analyze data and draw conclusions in the context of electrical engineering.
- Compare and evaluate machine learning concepts and algorithms.
- Use the Scikit-learn package in Python for practical applications.
- Apply machine learning training strategies to solve electrical engineering problems.
- Design a machine learning workflow for electrical engineering use cases.
Education Method
The course employs interactive, intuition-based lectures and hands-on coding labs. Students engage in practical implementations of machine learning models, integrating theoretical knowledge with real-world applications.
Literature and Study Materials
- Primary Textbook: Machine Learning Refined, Foundations, Algorithms, and Applications by Jeremy Watt, Reza Borhani, and Aggelos K. Katsaggelos (Cambridge University Press, 2nd edition, 2020, ISBN: 9781108480727).
- Supplementary Texts:
- An Introduction to Reinforcement Learning by Richard S. Sutton and Andrew G. Barto.
- Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges by Michael M. Bronstein et al.