Program Supervisor
e-mail: kharlamov@analyst.ru
Can you train your computer deep? Do you understand how hardware artificial brain works?
The program focused on the theory and application of modern neural networks and machine learning methods for solving human level problems. Strong and wide theoretical background provides you various tracks for researching: signal, text and image processing, robots control, cybersecurity, neural hardware design.
Neural Network Theory and Practice
Cyber Security
Image, Text and Signal Processing
Neurocontrol
Neural Network Theory - various architectures of Neural Networks and learning: deep and shallow, supervised and unsupervised, Reinforcement Learning, Spiking Networks, Soft Computing: Fuzzy nets, Wavelets, Evolution methods, etc.
Neuromathematics – applications of neural networks to mathematical problems: linear and nonlinear systems of equations solution, grids and graphs problems, multidimensional modelling, inter- and extrapolation, classification and clustering, etc.
Neurocontrol – applications of neural networks to control tasks: kinematics and dynamics, nonlinear controllers, robots control, etc.
Neurocomputers – hardware for neural networks acceleration, past, current and future: Graphical and Tensor Processing Units, Neural Chips, Optical Computers, Memristors, etc.
Applications of Neural Networks in Cyber Security; Image, Text and Signal Processing; Robot Control, etc.
|
|
Centre of Information Technology and Systems Chair of Intelligent Information Systems and Technology |
|
CITaS is a government organization for digitalization of all process in government responsibilities: Defence, Security, Medicine, Economics, Social etc. Chair has many years experience in both theory and application of Artificial Neural Networks and parallel computing: 8\16 GPU cluster for MIPT with software for: wind load prediction; hydrolocator data processing etc.; metal mill measuring and control, radiolocator data processing; object recognition over vibration of optical cable and many others. |
|
Full stack of neural networks architectures. Approaches of various tasks solution with neural networks. Particular task solved with neural networks. As example from previous years:
