Second semester at the best university in the EU
The biggest advantage of the United Kingdom being out of the EU is that I can now claim I study at the best university in the EU — Technical University of Munich — at least according to QS rankings. This semester, I have fought a serious illness and therefore was focusing only on 3 courses corresponding to 23 credits. Two of the three courses were very challenging to get into. When applying, I even had to send my CV. Luckily, I got in.
Overcoming challenge
I was diagnosed with cancer that has a very high chance for survival and since it was caught at a very early stage I knew the treatment would take a few months at most. Luckily for me, this semester was still in an online format and I could keep my hopes of finishing a few courses. The most natural course to focus on was a seminar where people read a research paper and then do a presentation about it. I also wanted to finish a more challenging practical course where people do weekly programming tasks.
The biggest challenge was planning around chemotherapy. I knew beforehand that it would be very difficult to do any kind of work during this time. Unfortunately, this proved to be correct. Chemotherapy affected my body and my mind. My short-term memory, concentration, and stamina were greatly affected. Plus I had to spend a lot of time in the hospital. To overcome this, I focused on just two courses and did extra work before the treatment started. I also asked course instructors for an extension for some tasks and thankfully they agreed. After I was done with these two courses, I’ve used my free time during the summer to prepare for the Foundations of Data Analysis retake exam.
Courses
[IN4273] Practical Course: Learning for self-driving cars and intelligent systems (2.0)
The course was divided into two parts. In the first four weeks, we were given four individual tasks. The tasks were very diverse — from estimating optical flow, generating artificial images from different camera points of view, to even developing an autonomous driving system in CARLA. I enjoyed this part a lot. After that, we picked a project and worked on it in a group of two. Our project was about image localization and re-localization. In short, given an image, we were trying to find the exact position the image was taken from based on a database of images and their positions. Our supervisor — Qadeer Khan — had weekly meetings with us and gave us assignments that even challenged the state-of-the-art. It wasn’t easy, but we learned a lot.
After the lecture period ended, we gave a presentation about our project. The date of the presentation coincided with my holiday in Croatia, so I ended up giving a presentation from the beach. At this point, the course still wasn’t over. Ahead of us was a viva — oral examination that tested our theoretical understanding of the concepts we implemented. I would say these questions were more difficult than the questions I’ve received during my bachelor thesis presentation. On the 8th of September, almost two months after the lecture period ended, we had finished the course. Overall, the course wasn’t easy and students were expected to spend a lot of time on it, but if you want to challenge yourself a little bit, I recommend the course.
[IN4977] Seminar: The Evolution of Motion Estimation and Real-time 3D Reconstruction (1.3)
Imagine having a car with a camera attached to it. You drive it around and in real-time you have a system that takes images from the camera and creates a map based on these images. This system could also keep the current position of the car on the map. This is called SLAM and the seminar focused on this problem. We studied state-of-the-art approaches to solving the SLAM problem. Each student read one paper, created a four-page summary of the paper, and did a 20–30 minute presentation. Each student also had one on one consultations regarding the studied paper and the final presentation. Both instructors, Lukas Koestler and Christiane Sommer were the reason why was the course so enjoyable. After the seminar was over, Christiane and Lukas organized an in-person meet-up at a local Beer Garden. I very much recommend the course.
[MA4800] Foundations of Data Analysis (3.0)
The most feared course of the Data Engineering and Analytics program. The only mandatory course taught by the mathematics department. My last and most difficult mandatory course. That’s Foundations in Data Analysis. It’s a math course that covers everything from advanced linear algebra, statistics like the Johnson–Lindenstrauss lemma to convex and non-convex optimization. For people with a computer science degree, the course is especially difficult as it may be their first course where they need to write mathematical proofs. Sure, I’ve encountered proofs during my bachelor’s studies, but in this course writing proofs was our daily bread-and-butter. Most of the proofs focused on using inequalities (such as Markov, Hoeffding, etc.). Even geometrical proofs such as proving that most of the volume of an n-dimensional ball is concentrated near the equator were done using inequalities. And the exam was no different. I still vividly remember writing proof that some set is a subset of some subdifferential on whole a A4 piece of paper. The worst thing was, there were 10 such non-trivial problems and only 90 minutes to solve them.
Writing proofs is one of those skills where it’s difficult to immediately see its usefulness. I would say it enhances a person’s problem-solving ability and it broadens the horizon of people with a computer science background. Because in a field such as AI/ML — which is complex and unpredictable — generalists, not specialists, are primed to excel. For me, it taught me not to be too afraid of difficult problems and to exert effort and not to give up as easily when solving a such problem. It is definitely a skill that I am glad learned a bit.
Overall results
This semester, I have finished all three courses I’ve focused on. And given the circumstances, I consider it a success. For the second time in a row, I can say “This was the most difficult semester I’ve experienced”. Next semester, I am looking forward to finally attending in-person classes and showing you more of the university facilities.
+----------------+-----+-----+-----+-----+-----+
| Semester | 1st | 2nd | 3rd | 4th | 5th |
+----------------+-----+-----+-----+-----+-----+
| GPA (1.0 best) | 2.2 | 2.2 | | | |
+----------------+-----+-----+-----+-----+-----+
| ECTS (credits) | 33 | 23 | | | |
+----------------+-----+-----+-----+-----+-----+
