Laboratory for Algorithms and Technologies for Networks Analysis

• We have developed new methods of research of network structures in models built from statistical data. For the first time we have developed a general concept of uncertainty of network structures on the basis of the theory of statistical processes with multiple solutions. Within the developed approaches we have investigated ways of reduction of uncertainty (errors) of networks analysis. We have found that usage of alternative proximity measures allows to significantly reduce uncertainty. Results can be applied to analysis of financial markets.

• On the basis of the notion of «critical graph classes» we have developed new approaches to analysis of computational complexity of problems of combinatorial optimization in hereditary graph classes. We have found critical classes of graphs for a wide range of problems of discrete optimization.

• Our researchers have developed efficient algorithms for solving problems with high computational complexity based on network structures. New approaches have been developed for solving problems in logistics and transportation that feature a large number of variables and limitations and have major practical importance.

• We have researched ways of increasing accuracy of processing of images of faces based on preliminary extraction of characteristic features using deep neural networks. A regularizing method has been proposed for criteria of search of the nearest neighbor that is based on maximization of credibility of distances between input images and all the samples. We have demonstrated that such an approach allows to increase accuracy of identification of faces on videos compared to the best of the existing similar algorithms. The most relevant applications are related to grouping faces for video surveillance systems.

Implemented results of research:

• Our researchers have completed applied projects according to agreements with organizations and companies: «Tander» JSC («Magnit» supermarket chain), «Intel» and others.

• In collaboration with «Yandex» we prepared the conference in networks analysis «NET–2018».

Education and career development:

• We have developed and are running a masters program «Intelligent data analysis» (specialization «Applied mathematics and informatics»).

• Students and postgraduates of the Nizhniy Novgorod branch of the Higher School of Economics participated in scientific projects of the Laboratory.

Organizational and structural changes: The Laboratory was transformed into an International Laboratory of the Higher School of Economics in 2016 is financed from extra-budgetary funds.

Other results:

• Every year the Laboratory conducts two international conferences and two scientific schools for students and young scientists: School for Operations Research and Applications (2011 – 2018) and School In Data Analytics (2016 – 2018).

• Young members of the academic staff of the Laboratory have received: a medal of the Russian Academy of Sciences in mathematics, a grant from the President of the Russian Federation, a grant from the Russian Science Foundation, a scholarship from the President of the Russian Federation, a Segalovich Scholarship from the «Yandex» company, a personal scholarship named after academician G.A. Razuvaev, as well as won a «IBM PhD Fellowship» competition.

• A team of students of the Laboratory were in top-10 of the best teas in the final of the international competition «Data Science Game» in Paris.

Collaborations:

• University of Florida (USA), University of Pittsburgh (USA): collaboration agreements

• S.A. Melentiev Energy Systems Institute of the Siberian Department of the Russian Academy of Sciences, S.L. Sobolev Institute of Mathematics of the Siberian Department of Russian Academy of Sciences, Moscow Institute of Physics and Technology (Russia), Omsk Research Center of the Siberian Department of Russian Academy of Sciences: organizing international scientific conferences

• A.A. Kharkevich Institute for Information Transmission Problems of the Russian Academy of Sciences, M. V. Keldysh Institute of Applied Mathematics of the Russian Academy of Sciences: organizing scientific schools