# Machine Learning (SS 2019)

## News

**16.08.2019:**We have to postpone the notification of the results of the final exam by at least 2 weeks, i.e., they will be published after August 28, 2019. Please be patient.**05.08.2019 (Important!):**There are two different rooms for the exam tomorrow, according to the first letter of your last name: Audimax (A030): A - M, B 201: N - Z. Both rooms are in the Main Building (Geschwister-Scholl-Platz 01)**26.07.2019**As announced, due to a lot of writing, the solutions to exercise 12 were uploaded.**25.07.2019: Please note the guidelines for the final exam.****25.07.2019**: Because of the sudden room change regarding the last lecture, there will be no record of it. The reason for that was, that the administration put the wrong information about the end of the lecture in the system 'LSF' (lecture directory, Vorlesungsverzeichnis). The consequence was that the room has shown up to be free for a final exam which obviously took place. Hence, we had to change the room in the morning and the video team had no time to set up their equipment. We are highly sorry for these circumstances.**25.07.2019**: There was a misunderstanding in the administration of the university. Therefore the last lecture will take place in E 120 (Große Aula)**25.07.2019**: Professor Tresp will be delayed today. Therefore, the last lecture will start about 15min later.**10.07.2019**: We uploaded the solution for exercise 10-1. In the tutorials on Tuesday, i calculated the values with 99% for specificity and sensitivity instead of 99.9%. The steps were right, but the values were wrong (sry for that!) - CF.**26.06.2019**: From July 1st onwards you can register for the final exam via UniWorX.**26.06.2019**: Because many students missed the tutorials on Whit Tuesday, you can now find the solution for exercise sheet 06 online.**17.06.2019**: This week's tutorials will regularly take place. We are highly sorry for the delay in uploading exercise sheet 07**07.06.2019**: A proof of the invertibility of the matrix occuring in the analytical solution of the linear regression problem with regularization can be found under tutorial 3. Please note that a modified version of the identity matrix has to be used.**17.05.2019**: The final exam will be on August 6th, 2019 from 10am - 12pm. Further details will be announced on time.**25.04.2019**:**Important**- the tutorials will start on May 7th, resp., May 8th, due to the*Labor Day*on May 1st**25.04.2019**: The lecture will be recorded. You can find the videos here.**28.02.2019**: The registration for this course via UniWorx will be open from April 1st, 2019 onwards.

## Organisation

**Course:**3+2 hours weekly (equals 6 ECTS)**Lecture:**Prof. Dr. Volker Tresp**Assistants:**Christian Frey, Sabrina Friedl**Required:**Professional skill of at least one programming language**Audience:**The course is directed towards master students in informatics, bioinformatics and media informatics**Course Language:**English

## Time and Locations

All times are c.t. (cum tempore)

Component | When | Where | Starts at |

Lecture | Thu, 9:00 - 12:00 h | Room M 218 (Geschwister-Scholl-Platz 1) | 25.04.2019 |

Tutorial 1 | Tue, 14:00 - 16:00 h | Room B 015 (Geschwister-Scholl-Platz 1) | 07.05.2019 |

Tutorial 2 | Tue, 16:00 - 18:00 h | Room B 015 (Geschwister-Scholl-Platz 1) | 07.05.2019 |

Tutorial 3 | Wed, 14:00 - 16:00 h | Room B 015 (Geschwister-Scholl-Platz 1) | 08.05.2019 |

Tutorial 4 | Wed, 16:00 - 18:00 h | Room B 015 (Geschwister-Scholl-Platz 1) | 08.05.2019 |

## Content

Machine Learning is a data-driven approach for the development of technical solutions. Initially motivated by the adaptive capabilities of biological systems, machine learning has increasing impact in many fields, such as vision, speech recognition, machine translation, and bioinformatics, and is a technological basis for the emerging field of Big Data.

The lecture will cover:

- Supervised learning: the goal here is to learn functional dependencies for classification and regression. We cover linear systems, basis function approaches, kernel approaches and neural networks. We will cover the recent developments in deep learning which lead to exciting applications in speech recognition and vision.

- Unsupervised Learning: the goal here is to compactly describe important structures in the data. Typical representatives are clustering and principal component analysis

- Graphical models (Bayesian networks, Markov networks), which permit a unified description of high-dimensional probabilistic dependencies

- Reinforcement Learning as the basis for the learning-based optimization of autonomous agents

- Some theoretical aspects: frequentist statistics, Bayesian statistics, statistical learning theory

The technical topics will be illustrated with a number of real-world applications.

## Course Schedule

## eLectures

The lecture will be recorded. The videos will be uploaded after each lecture and can be found here

## Final Examination

**Date:**August 6th, 2019**Time:**10am - 12pm**Location:**Hauptgebäude (Main Building) Audimax (A030): A - M // B 201: N - Z**Registration**UniWorX (will be open from July 1st onwards) - For students with no UniWorX account it is mandatory to write an email to the assistants such that we can register you manually**Guidelines:**guidelines.pdf

Please note: There will be **no **second exam!

## Recommended Literature:

- Deep Learning.
*Ian Goodfellow, Yoshua Bengio and Aaron Courville* - The Elements of Statistical Learning: Data mining, Inference and Prediction.
*Hastie, Tibshirani, Friedman* - Machine Learning: a Probabilistic Perspective.
*Kevin Murphy* - Bayesian Reasoning and Machine Learning.
*David Barber* - Pattern Recognition and Machine Learning.
*Christopher M. Bishop* - Artificial Intelligence: A Modern Approach.
*Russel and Norvig*

## Links to Tutorials

- Introduction to Python: Dive Into Python 3
- Introduction to PyTorch: Basics of PyTorch
- Playground TensorFlow