Our research is conducted in the following research groups:

  • Computer Vision (Victor Lempitsky’s group)
    Computer vision group research is about designing computer systems that can extract, organize, and quantify information contained in images of various types and origin. For this purpose, the group develops new machine learning techniques (deep learning in particular) and optimization techniques that are robust and flexible enough to handle and to adapt to the diversity of image data in the modern world.
  • Scientific Computing (Ivan Oseledets’s group)
    Our research focuses on developing breakthrough numerical techniques for solving a broad range of high-dimensional problems. The key ingredient is the effective decomposition of multidimensional arrays (tensors). Our recent interests also involve graph mining, recommender systems, and topological shape optimization.
  • Advanced Data Analytics in Science and Engineering (Evgeny Burnaev‘s group)
    ADASE group research is about developing efficient machine learning methods and their use for solution of applied engineering problems and industrial analytics. In particular, we are working on regression based on Gaussian processes and kernel methods for multi-fidelity surrogate modeling and optimization, Deep Learning for 3D Data Analysis and manifold learning, on-line sequence learning for prediction and non-parametric anomaly detection.
  • Multiscale modelling (Alexander Shapeev’s group)
    Our research is devoted to all kinds of problems with several spatial and temporal scales, and in particular to problems arising in molecular modelling. One of our current thrusts is developing efficient and accurate models of molecular interaction.
  • Laboratory “Tensor networks and deep learning for application in data mining”

    Laboratory Tensor networks and deep learning for applications in data mining under the guidance of Prof. Andrzej Cichocki on the basis of the Skolkovo Institute of Science and Technology develops new fundamental approaches for training, testing and storing parameters of deep neural networks based on tensor decomposition techniques. These approaches allow to reduce by orders of magnitude computational complexity and required memory for the operation of the network, while maintaining a high quality of prediction.

    The Laboratory mission is to pursue cutting-edge research in the design and analysis of deep neural networks, tensor decompositions, tensor networks and multiway analysis with many potential practical applications.

    The Laboratory brings together several professors and young researchers in the fields of machine learning, computer vision, artificial intelligence, robotics, large-scale data analysis, mathematics as well as computational neuroscience and bioinformatics.

  • AeroNet Lab
    The research of the lab is devoted to the creation of an open benchmark for remote sensing data and their products – a set of images with the labels of different classes, obtained from various aircrafts, in particular satellite imagery, UAV and MA survey. Our lab activities are  to build adaptation procedures of modern intelligent algorithms for data analysis, and also to develop fundamentally new approaches to the processing of multidimensional/multi-fidelity remote sensing data (semantic segmentation, classification, detection of changes, consolidation of heterogeneous data, etc.)

The main research interest of the group is in finding new ways in chemical informatics that are based on a combination of physical chemistry methods with the machine learning techniques for prediction of properties of organic compounds. Our primary goal is to develop methods that, on one hand, are accurate and, on the other hand, are universal and have wide applicability domains.