Structural Inference in High-Dimensional Models 2, Pushkin, August 26-30, 2019

On August 26–30, 2019, the International Laboratory of Stochastic Algorithms and High Dimensional Inference (HDI Lab) held the Second Annual Conference ‘Structural Inference in High-Dimensional Models’. The focus of the event this year was on the following topics:
– High-dimensional probability and statistics;
– Statistical and computational optimal transport;
– Markov Chain Monte-Carlo (MCMC);
– Deep-network mathematics.

The conference took place in Kochubey Mansion (Pushkin, St. Petersburg). The event included a number of mini-courses, topical sessions, and a poster session. Traditionally, we invited a truly stellar lineup of speakers. Speeches and presentations were delivered by Professor Stephane Mallat, President of the German Mathematical Society (DMV) Friedrich Götze, Member of the Russian Academy of Sciences Ildar Ibragimov, Professor Mikhail Lifshits, Alexandre Tsybakov, Anatoli Yuditsky, Marco Cuturi, Axel Munk, Christophe Andrieu, Olga Klopp, and many other world-class scientists, as well as HDI Lab researchers. On top of that, a number of young participants made their poster presentations.

The educational part of the conference consisted of the following four mini-courses:

• Stephan Mallat (College de France) ‘Mathematics of Deep Convolutional Neural Networks’
• Olga Klopp (ESSEC Business School) ‘Matrix Completion: Old and New’
• Mikhail Lifshits (SPBU) ‘How Complex is a Random Picture’
• Christophe Andrieu (University of Bristol) ‘Non-Reversible Markov Chain and Process Monte Carlo Methods’

The scientific part of the conference included the following sessions:

• MCMC Session, organized by Eric Moulines (Ecole Polytechnique, HSE University)
• Statistical Session, organized by Anatoli Yuditsky (Université Grenoble Alpes)
• High-Dimentional Probability Session, organized by Alexey Naumov (HSE University)
• Statistical Optimal Transport Session, organized by Vladimir Spokoiny (WIAS, Berlin; HSE University)

The complete list of participants, as well as abstracts and slides of presentations, mini-courses, and posters can be found on the Conference Page. 

The next conference will take place in Bordeaux in August 2020. Details will be available soon.

Meanwhile, we are pleased to invite students, PhD students and early career researchers to register for the HDI Lab winter school MML2020: Math of Machine Learning 2020 in Sochi, co-organized by Sirius University.