Science

Scientific Activity

The research activities of the Department of KMAD are focused on the development of modern methods, models, and information technologies in the fields of intelligent data analysis, machine learning, and mathematical modeling of complex systems.

The scientific work of the department is closely related to the educational program “Intelligent Data Analysis” and encompasses both fundamental and applied research in data analysis, decision-making, intelligent systems, and control under uncertainty.

Research results are actively implemented in information technologies, decision support systems, diagnostic systems, and in solving control problems of complex technical, economic, and logistics systems.

Research Areas

At the KMAD Department, scientific research is conducted in the following main areas:

Methods of Intelligent Data Analysis and Machine Learning:

  •  intelligent decision support systems based on Preference Learning and Learning to
    Rank;
  • semi-supervised learning for aggregated data;
  • construction of composite indicators using kernel-based machine learning methods;
  • predictive analytics of nonlinear time series

Mathematical and Computer Modeling of Systems:

  • interval models of aggregated Markov systems;
  • modeling of non-contact surface diagnostics processes;
  • optical diagnostic systems using artificial intelligence.

Information Technologies for Managing Complex Systems:

  • consensus control of multi-agent systems;
  • management of logistics systems and supply chains under uncertainty and risk.

Key Scientific Results

Among the most significant results achieved at the department:

  •  development of methods for robust consensus control of multi-agent systems under uncertainty (disturbances, delays, variable network topology) based on invariant ellipsoids and linear matrix inequalities;
  • creation of methods for constructing nonlinear models of integral indicators, preference functions, and ranking based on kernel machine learning methods and aggregated expert-statistical data;
  • development of approaches to modeling discrete stochastic systems using interval uncertainty models;
  • creation of information technologies for intelligent optical diagnostics based on machine learning methods

State Scientific and Technical Projects:

The department has extensive experience in carrying out research and applied projects, including those funded by the state budget and contractual agreements.

State Scientific and Technical Projects:

  • “Development of information technology for forming portfolios of national-level projects…” (2015–2018);
  • “Methods for solving inverse problems of diagnostics and control of nonlinear systems…” (2011–2014);
  • “Statistical and neural network methods for computer monitoring of complex systems…” (2010–2013).

Contractual and Commercial Projects:

  •  development of semi-automated machine learning algorithms (client — Samsung Electronics Ukraine);
  • clustering of large-scale text data (Samsung Electronics Ukraine);
  • ranking models for search systems (Noosphere Ventures, UK);
  • web-based decision support systems for financial markets (USA);
  • modeling of energy systems (Republic of Tajikistan).