Second Summer School on Bayesian Methods in Deep Learning announced

On November 11, at Tesla Space Sberbank Data Science Day was held, a one-day event that gathers Moscow experts in data analysis to discuss trends in the field and to share experiences. Sberbank Data Science Day is being held for the second time, and both times it is attended by several hundred specialists in the field of machine learning. Like last year, Dmitri Vetrov spoke at the event with a lecture. In addition, it has become a tradition to announce on this day about holding the Summer School on Bayesian methods in Deep Learning, which is organized by the International Laboratory of Deep Learning and Bayesian Methods.

In 2017, the summer school was held in August, gathered about a hundred selected participants and had good reviews, as well as wishes for a repeat of the school next year. In 2018 the school will be held at the international level and in English. The laboratory plans to invite several foreign lecturers. Subjects of the school will remain the same, but with an even greater bias in the latest developments in the field of neurobayesian methods. The official announcement of the summer school will appear in the near future, and the selection is scheduled to begin in February.

A lecture by Dmitry Vetrov on the Sberbank Data Science Day was devoted to the same topic - the neurobayesian methods. In it, Dmitry spoke about the opportunities that Bayesian methods open in deep learning: regularization, ensembling, automatic model selection and building new models on the basis of already trained ones. To learn more about these approaches, Dmitry invited everyone to apply for a summer school.