From Russia With Love

Filip packed his bags and left for Moscow to participate in the DeepBayes summer school, learning a thing or two about using Bayesian methods within deep learning. And, of course, indulging in the traditional Ukrainian and Georgian cuisine.

There’s something that quality summer schools in machine learning do best: they make you feel stupid. That’s exactly how our Filip felt after a few introductory classes at the second edition of the DeepBayes summer school held in Moscow. After its successful first edition held exclusively for Russian participants, DeepBayes opened its doors for everyone interested in the application of Bayesian methods within deep learning.

Filip taking a break from the classes.

Organized at the Soglasie Hall in Moscow, the DeepBayes summer school provided six completely packed days of classes and practical sessions, covering a wide range of topics. Almost 100 school participants tried their hands at things like variational autoencoders, distributional reinforcement learning, and Bayesian neural networks armed with variational dropout. If we were to trust Filip, both classes and practical sessions were challenging, yet extremely eye-opening and useful.

The official group photo.

All DeepBayes activities were also intertwined with a fair amount of touristic wondering, visiting famous locations and indulging in traditional cuisine. Filip climbed the Ostankino tower, the highest building in Europe, but also gave boarding the space station Mir a try as well. As a real gourmand, Filip lived off traditional Ukrainian and Georgian cuisine, making a life enemy out of okroshka, a traditional Russian cold soup based around kvass (fermented beverage) and veggies.

On the flip side, he made life friends out of Bayesian methods after six amazing days and departed, as written in the famous movie’s title, “From Russia with love”. 🙂

View from the Ostankino tower.

Filip hanging out with his former Amazon colleague Alexey.

Space station Mir at the Cosmonaut museum.