Upon successful completion of the course, the student can:
- Understand the mathematics behind machine learning algorithms
- Understand the mathematics behind deep learning algorithms
- Deploy deep learning models in production
- Understand and design machine learning procedures algorithms and pipelines
- Have an understanding of the intrinsic link between probability and statistical machine learning
|
|
|
Introduction to optimization techniques, Gradient descent, Model Selection, Bias-Variance Tradeoff, Gaussian Processes, Bagging, Boosting, Decision Trees. An introduction to data structure optimization, Stochastic Gradient descent and all its variants, batch learning. Automatic differentiation computational graphs and learning optimization , Deep learning, neural networks and backpropagation.
|
|
|
Essential knowledge to be able to pass the course::
- Probability theory, calculus and linear algebra
- Basic Python programming
- Basic understanding of computational complexity
- Basic knowledge of a Unix/Linux operating system
Advised:
- Measure theory
- Stochastic calculus
- Advanced programming and mathematical logic knowledge
|
|
The test will exist of four larger individual assignments
There will be one course grade which is the average of the assignments
Possibility to resit of each type: Assignments can be handed in again for grading if and only if the original grade was lower than 5.5
Grading for each part is registered in Brightspace, only the final grade is published in Osiris.
|
|
The reading material is non-obligatory where students can go and look for extra material or deeper understanding.
|
|