# Master courses and my personal evaluation

**Disclaimer**
> The information presented in this review represents personal opinions and evaluations concerning the Master's courses I have completed. It is intended for informational purposes only and does not constitute academic or professional advice. While I strive for accuracy, I cannot guarantee the completeness or timeliness of the information presented. Course syllabi, instructional methods, and assessment procedures are subject to change. Please consult the respective educational institution for the most current and official information. Additionally, results may vary; what I found beneficial may not be applicable to all students, and my experiences should not be considered a guarantee of anyone else’s outcomes.

Last Update : 29.08.2023

General comments:
* Always take the VL and UE together
* The courses have a certain sequence that should be followed.
* If you have to do extra courses you should do them first before you start with master courses.
* Stick to the given schedule.
* Find a study group or group with which you can discuss.
* Exercises can take a lot of time do not underestimate it.

## 1st Semester (WS)

### Computer Vision (3 VL)
The lectures are recorded and the videos are available until the end of the semester. At the end of the semester, you have a Moodle exam that you have to write on-site, for example in Linz, Vienna, or Bregenz. The exam consists mostly of multiple choices and small calculations. A very good preparation is to look at old exams. There is no attendance requirement in lectures.

Level of difficulty: 5/10

Prior knowledge: Nothing

### Computer Vision (1.5 UE)
In the exercise, a project is undertaken. This project is usually done in groups of 4-5 people. You should be good at programming for this. Attendance was required in the exercises. The progress of the project was presented every few weeks.

Level of difficulty: 6-7/10

Prior knowledge: Python

### Deep Learning and Neural Nets I (3 VL)
It's an intro course. You don't need much prior knowledge to start but it's good if you have some basic knowledge. The lectures are recorded and the videos are available until the end of the semester. There is no attendance requirement in lectures. Moodle exam in the end.

Level of difficulty: 5/10

Prior knowledge/ good to have: supervised techniques

### Deep Learning and Neural Nets I (1.5 UE)
There are four assignments and one is a bonus. At the end, you have a small test with 10 points that you must pass to get the grade from the exercise.

Level of difficulty: 6/10

Prior knowledge: Python

### LSTM and Recurrent Neural Nets (3 VL)
The lectures are recorded and the videos are available until the end of the semester. There is no attendance requirement in lectures. There will be two compulsory Moodle exams.

Level of difficulty: 6/10

Prior knowledge/ good to have: NLP

### LSTM and Recurrent Neural Nets (1.5 UE)
There are four assignments. You were chosen at random and had to present it. There were many groups and time slots.

Level of difficulty: 6/10

Prior knowledge: Python

### AI and Law I (3 VL)
The lectures are recorded and the videos are available until the end of the semester. There is no attendance requirement in lectures. Moodle exam in the end.

Level of difficulty: 1-2/10

Prior knowledge: Nothing

### Artificial Intelligence in Society (1.5 KV)
The lectures are recorded and the videos are available until the end of the semester.

Mandatory:
• Final writing assignment (70% of the grade)
• 6-7 pages written report on a on a topic of the course
• Two assignments (30% of the grade)
• Presentation or 1-pager
• Deadlines: 3 weeks after lecture 1 and 2
• In groups up to 5 people
• Group registration will be done via Moodle

Level of difficulty: 1-2/10

Prior knowledge: Nothing

## 2nd Semester (SS)

### Deep Learning and Neural Nets II (3 VL)
It build up on DLNN1. The lectures are recorded and the videos are available until the end of the semester. There is no attendance requirement in lectures. Moodle exam in the end.

Level of difficulty: 6/10

Prior knowledge: Deep Learning and Neural Nets I

### Deep Learning and Neural Nets II (1.5 UE)
There are four assignments and one is a bonus. At the end, you have a small test with 10 points that you must pass to get the grade from the exercise.

Objectives
■ Lecture
□ Overview of the most important state-of-the-art methods
□ Understand the current challenges in the field of RL
□ Recognize strengths and weaknesses of different
methods
□ Analyze and discuss Deep Reinforcement Learning
systems

Level of difficulty: 6/10

Prior knowledge: Deep Learning and Neural Nets I

### Deep Reinforcement Learning (3 VL)
There is no attendance requirement in lectures. Moodle exam in the end.

■ Lecture
□ Overview of the most important state-of-the-art methods
□ Understand the current challenges in the field of RL
□ Recognize strengths and weaknesses of different
methods
□ Analyze and discuss Deep Reinforcement Learning
systems

Level of difficulty: 6/10

Prior knowledge: Reinforcement Learning

### Deep Reinforcement Learning (1.5 UE)

Objectives:
■ Exercise
□ Implement basic Deep Reinforcement Learning methods
□ Experience common pitfalls when training Deep RL
methods

Topics
■ Imitation Learning
□ Behavioral Cloning & DAgger
■ Value-based methods
□ Deep Q-Learning
■ Policy gradients methods
□ Proximal Policy Optimization

Exercise Mode
■ Exercise sessions are 45 minutes
□ Attendance is not mandatory
□ Streamed live via Zoom (link on Moodle)
□ Recording available afterwards on Moodle

Level of difficulty: 6/10

Prior knowledge: Python

### Theoretical Concepts of Machine Learning (3 VL)
There is no attendance requirement in lectures. Moodle exam in the end.

Level of difficulty: 7/10

Prior knowledge/ good to have: supervised and unsupervised techniques

### Theoretical Concepts of Machine Learning (1.5 UE)
There are four assignments and 1-2 assignments are bonus.

Level of difficulty: 7-8/10

Prior knowledge: Python

### AI and Law II (1.5 VL)
There is no attendance requirement in lectures. Moodle exam in the end.

Level of difficulty: 1-2/10

Prior knowledge: Nothing

### Robopsychology (3 KV)
The course has mini-tasks for reviewing scientific papers and an online study. There's also a group task focused on practical AI design, and the course wraps up with a closed-book written exam.

The grade results from the following parts:

Written exam (mandatory): max. 24 points (12 points must be achieved as a minimum)
Mini tasks (optional): 4 mini tasks à max. 2 points each = max. 8 points
Group task (optional): max. 8 points

Level of difficulty: 1-2/10

Prior knowledge: Nothing

## 3rd Semester (WS)

### Explainable AI

The course consists of a lecture and a practical lab.

The lecture is held weekly in a hybrid format (HS + online), with a written exam at the end of the semester.

The lab is held irregularly in hybrid format (JKU + online) with practical assignments throughout the semester. Attendance in the lab sessions is compulsory. Generally, you should join the slot to which you are assigned but you can join other groups in exceptional cases.

Level of difficulty: 5/10

Prior knowledge: Python, DLNN1 and DLNN2

### Communicating AI

Course requirements (How to get graded):

Total points that can be achieved in the course: 50
Gorup task paper: max. 35 points
AI image survey task: max. 9 points
Breakout task 1: max. 3 points
Breakout task 2: max. 3 points

What happens there?
• Link to Zoom online lecture room
• Q & A forum
• Slides and materials
• Recordings of sessions (only partly available)
• Assignments for group task
• Submission of individual tasks, breakout tasks and group task

Level of difficulty: 1-2/10

Prior knowledge: Nothing

### Probabilistic Models (3 VL)
The lectures are held in a lecture hall where I can attend in person, but they're also live-streamed via Zoom. Plus, they record the lectures and make the videos available to watch anytime during the semester. In the end Moodle exam.

Level of difficulty: 6/10

Prior knowledge: Nothing

### Probabilistic Models (1.5 UE)

In this course, I get hands-on practice through Problem Sets that follow major lecture topics. I have two weeks to complete each set, and afterward, there are videos that explain sample solutions. The class format is flexible—I can attend in person, via Zoom, or watch prerecorded videos. Occasionally, some of us are chosen to present our solutions live.

Level of difficulty: 6-7/10

Prior knowledge: Python