HSE University and Yandex organized the 3d Joint Autumn School on Generative Models

The Autumn School took place from November 23 to 26 in mixed format. The classes were hosted at Yandex School of Data Analysis. Those who were not able to attend on-campus joined online.

 

The school was aimed at young undergraduate/graduate students planning to develop a research career. This intense four-day workshop consisted of 3 interdisciplinary mini-courses, a list of invited talks, and a lively discussions. The three mini-courses were:

• Generative models by Denis Derkach (HSE University),
• Diffusion Generative Models by Dmitry Vetrov (HSE University),
• Efficient Sampling Techniques by Alexey Naumov (HSE University).

The participants were satisfied with the school: the average score according to the feedback form is 4.8 out of 5. Below we cite some of the participants.

Maxim Tingaev: "Most of all I liked that the school was held online in the zoom, you could write in the zoom chat or in the telegram chat. Also, all the materials of lectures and practices were provided, which is important, when I begin to analyze these topics in more detail, it will be where to start the study. As an improvement, I don't even know what to add, in principle there were a lot of format and activities. Maybe some online tests or validated problem solving, maybe practical problems with solution validation or demonstration".

Julia Smolkina: "I loved networking and random coffee bot idea, also I appreciate openness of speakers and how they were ready to answer the questions and to continue communication afterwards".

Lilia Kurchenko: "In my opinion, on-campus classes were more effective than zoom. As for presentations, one of the most interesting was from Samsung Research. It would be interesting to listen to several more lectures by Dmitry Vetrov. After the school and meeting so many motivated people I feel inspired to enter YASD. Sometimes the topics were challenging, for example not all students were familiar with stochastic differential equations. I also liked that active participation was encouraged with souvenirs. And thanks a lot for all the snacks!"

The school is organised by three HSE laboratories:

• LAMBDA (Laboratory of Methods for Big Data Analysis)
• DeepBayes (Centre of Deep Learning and Bayesian Methods)
• HDI Lab (International laboratory of stochastic algorithms and high-dimensional inference)

Yandex is an industry partner of the school.

This school is supported by the RSF grant N19-71-30020, "Applications of probabilistic artificial neural generative models to the development of digital twin technology for Non-linear stochastic systems".