---

Topic 1 - Introduction to Machine Learning. Regularized Regression.

Lecture block 1 - Basic ML principles and regularized regression
Reading: Slides | Chapters 1.1-1.3, 3.1-3.2 in PRML | PDSH: what is ML? | PDSH: intro to scikit-learn | PDSH: hyperparameters and model validation | PDSH: feature construction
Extras: Demo linear regression in scikit-learn | Bias-Variance illustration | Variable selection illustration

Lecture block 2 - Trees, forests and beyond
Reading: Slides | PDSH: Decision trees and Random forest | ESL: 9.2, 10.1-10.3 | XGboost article

Computer lab 1 - Regularized regression and Tree models
Problem 1
Problem 2

Lecturer: Mattias Villani
Lab assistants: Amanda Olmin and Caroline Svahn

Topic 2 - Classification

Lecture block 3 - Learning and Classification
Reading: Slides | Chapters 1.2.3, 1.2.6 in PRML | PDSH: Naive Bayes | PDSH: Decision trees and Random forest
Extras: Maximum likelihood optimization | Real-time digit recognition

Lecture block 4 - Unsupervised learning
Reading: Slides | Chapters 9.1-9.2 in PRML | PDSH: k-means | PDSH: Gaussian mixture models

Computer lab 2 - Classification and Learning
Problem 1
Problem 2

Lecturer: Mattias Villani
Lab assistants: Amanda Olmin and Caroline Svahn

Topic 3 - Neural Networks and Deep Learning

Lecture block 5 - Neural networks and convolutional neural networks
Reading: Slides | Chapters 6, 7, 8, 9 in Deep Learning.
Code:
Other material: Tensorflow playground, Keras documentation

Lecture block 6 - Generative adversarial networks and recurrent neural networks
Reading: Slides | Chapters 10, 11, 20.10.4 in Deep Learning.
Code:
Other material:

Computer lab 3 - 2D CNNs
Image classification
Image segmentation Data

Cloud instructions

Lecturer: Anders Eklund
Lab assistant: David Abramian

Topic 4 - Reinforcement Learning

Lecture block 7 - Q-Learning Algorithm
Reading: Slides | Chapters 1-7 in RLI.

Lecture block 8 - REINFORCE Algorithm
Reading: Slides | Chapters 9, 10, 12, 13 and 16 in RLI.

Computer lab 4 - Grid worlds
Lab Lab2

Lecturer: Jose M. Peña
Lab assistant: Joel Oskarsson

Course literature

• Bishop Pattern Recognition and Machine Learning, Springer, 2006. [PRML]
• VanderPlas Python Data Science Handbook, O'Reilly, 2016. [PDSH]
• Hastie, Tibshirani, Friedman Elements of Statistical Learning, Springer, 2008. [ESL]
• Sutton, Barto Reinforcement Learning: An Introduction, MIT Press, 2018. [RLI]
• Goodfellow, Bengio, Courville Deep Learning, MIT Press, 2016. [Deep Learning]
• Other material distributed on this web page under each topic.

Teachers

Lecturers

Mattias Villani
Professor of Statistics
Focus: Bayesian Statistics and Machine Learning

Jose M. Peña
Associate Professor in Computer Science
Focus: Graphical Models and Causality

Anders Eklund
Associate Professor in Medical Informatics
Focus: Deep Learning and Neuroimaging

Lab assistants

Caroline Svahn
WASP Industrial PhD student in Statistics
Focus: Machine Learning for 5G networks

Amanda Olmin
WASP PhD student in Statistics
Focus: Deep Learning and Bayesian Machine Learning

David Abramian
Phd student in Medical Informatics
Focus: Machine Learning for Neuroimaging

Joel Oskarsson
Master student in Applied Physics and Electrical Engineering
Focus: Machine Learning and Reinforcement Learning

Course information

The typical participant has a degree in engineering, finance or other quantitative fields. We recommend that participants have taken at least one course in each of the following subjects:

• linear algebra
• calculus
• statistics
• programming

Examination and Grades: The report from each computer lab will be graded Pass or Fail. Successful completion of all five labs gives the grade Pass on the whole course.

Course organization The course is organized in four topics, each containing six lecture hours and a computer lab.