Machine Learning 2

Course guarantor:

Filip Šroubek

Lecturer:

Filip Šroubek

Tutor:

Soňa Drocárová, Tomáš Kerepecký, Adam Novozámský, Filip Šroubek

Supervisor:

Department of Mathematics

Synopsis:

1.Fundamental topics from the probability theory and machine learning (classical distributions, Bayes theorem, Kullback-Leibler divergence, curse of dimensionality, overfitting, maximum likelihood and maximum a posteriori estimators, Principle Component Analysis)

2.Deep feed-forward networks (hidden units, nonlinear activation functions, output units, loss functional, ML principle) )

3.Optimization for training deep models (stochastic gradient descent, back-propagation algorithm, algorithms with adaptive learning rates, implicit and explicit regularization)

4.Advanced network architectures (convolutional, recurrent, transformers)

5.Unsupervised learning (generative adversarial networks, normalizing flows, variational autoencoders, diffusion models)

6.Applications of deep learning (classification, segmentation, image reconstruction, language models, image generators)

Requirements:

Credit award: Participation in the final exercise, during which student teams present the results of their semester project.

Syllabus of lectures:

For more information, visit https://su2.utia.cas.cz/

Syllabus of tutorials:

For more information, visit https://su2.utia.cas.cz/

Study Objective:

The course focuses on understanding the principles of deep learning. In addition to learning theory and deep network optimization, advanced architectures of convolutional and recurrent networks, transformers, and the principles of generative models will be introduced.

Study materials:

Key references:

[1] Prince S.: Understanding Deep Learning, MIT Press, 2023.

Recommended references:

[2] Goodfellow I., Bengio Y., Courville A.: Deep Learning, MIT Press, 2016.

[3] Bishop, Christopher M.: Pattern Recognition and Machine Learning. Springer, 2006.

[4] Géron A: Hands-On Machine Learning with Scikit-Learn and TensorFlow, 2017.

[5] Chollet, F.: Deep Learning with Python, 2018.

[6] online resources: pytorch.org/tutorials/, playground.tensorflow.org, tensorflow.org/learn/

Note:

Further information:

https://su2.utia.cas.cz/

Time-table for winter semester 2025/2026:

Time-table is not available yet

Time-table for summer semester 2025/2026:

Time-table is not available yet

The course is a part of the following study plans:

• Aplikované matematicko-stochastické metody (compulsory course in the program)
• Aplikace informatiky v přírodních vědách (elective course)