Laboratory for Methods and Technologies For Working With Ultra-Large Archives of Space Monitoring Data

Head of the Lab

Ph.D. Andrey A. PROSHIN

development of new methods and approaches to organize the storage and parallel processing of ultra-large data, which are used for analysis of long-term observations of various processes and phenomena;
development of the physical and mathematical models to assess and forecast the state of high-latitude ionosphere;
research and development of machine learning methods, deep learning methods for processing and analysis of astrophysical data and reinforcement learning methods.

The laboratory was established in 2019 under the "Science and Universities" National Project with the main goal to develop new approaches to big data in space research

New methods developed to organize the work with ultra-large remote sensing data archives. In particluar, dynamic block access technology was implemented, which inreased the efficience of satellite data processing and facilitiated users' access.
Parameters of longitudinal currents during magnetic storms of 24th solar cycle were derivated, using measurements of SWARM polar satellites. Small-scales structures of extra-high density were revealed and their link to local auroral arcs and electrojets.
Deep-learning algorithms were developed for segmentation and search for objects in the data of microwave sky surveys and morphological classification of galaxies on the images in the visual range. Mahine learning models were developed, which are used to divide into classes and measure redshifts of X-ray objects together with the table data of multi-wave photometric sky surveys.

Its results are summarized in the annual reports (.pdf, in Russian):

As of today, 8 people work in the laboratory. Staff members of the Space Plasma Physics and High-Energy Astrophysics Departements also take part in the works run by the lab.

The laboratory's page on the website of the department "Earth Remote Sensing Technologies"