Last week I had the opportunity to attend the 1st Transylvanian Deep Learning Summer School (TMLSS). It took place in Cluj, Romania and its main focus was Deep Learning and Reinforcement Learning. Here is the link to website that includes the programme and the full list of lecturers. (link)
The aim of the organizers was
"to popularize topics around machine learning and artificial intelligence more broadly in Europe and specifically in Eastern Europe, and to encourage research in these fields."
I found this objective to be a great idea and I firmly believe that initiatives such as this help the research community grow and expand with people from diverse backgrounds.
Speaking of diversity the majority of the lecturers came originally from Eastern European countries. The student mix was 56% from Eastern Europe, 36% Western Europe and US and 8% from the rest of world. In total 30 different countries were represented including far way places such as Philippines, Brazil and Iran. From a gender perspective 43% of the students were women, which is also an impressive ratio.
Since this was my first summer school I have no point of reference to compare it with, but I found it to be a great experience. Looking at the programme one can see that it was intense, with plenty of lectures and labs, as well as social gatherings and opportunities to meet and greet with other students and teachers.
I was immensely impressed by how humble and approachable all the lecturers were. In a full week of interacting with all these accomplished scientists I never heard one of them bragging about their accomplishments or the field in general. It shows that people at the forefront of DL & RL research do not necessarily embrace the hype around it. This was also reflected in the classroom, where part of the discussion was the current open research problems and the limitations of what we know and can do.
When it comes to the content itself, it was a mix of introductory and more advanced lectures and labs in Mathematics of Deep Learning, Convolutional Neural Networks, Recurrent Neural Networks, Unsupervised Learning and Reinforcement Learning. As expected, some classes were easier to follow than others, but I believe that there is always an opportunity to learn something at any circumstance. I also viewed the past week as a way to gauge my own level of knowledge in certain subjects and even more importantly I viewed it as an opportunity to get a peek in new areas that serve as a potential source of new research directions. One of the topics that I found extremely interesting and potentially applicable to Network Security was the Graph Neural Networks, which funnily enough was the last lecture of the school. So, in a way they saved the best for last as far as I'm concerned.
After a full immersion for a week, I feel inspired not just by the knowledge I acquired, but also by the interaction with all these interesting people -teachers and fellow students alike- and the exposure on different ways of thinking about DL and RL research.
Last but not least I have to say that the organization before, during and after the school was impeccable. A lot of effort was put from a good number of people and that was evident in the fact that there were virtually no problems or at least no problems that we, as participants, observed.
Thank you for taking such a good care of us and making this a wonderful experience!