Laboratory of adative methods of data processing

Phone: +7 495 939 46 19

Site: http://nnga.narod.ru

 

Laboratory of adaptive methods of data processing

The Laboratory of adaptive methods of data processing (LAMDP) is organized in 2009. It's Head is Professor, DSc. Igor Persiantsev.

Activity of the Labratory concerns the development and using of the modern adaptive methods for data processing and analysis for mathematical processing of physical measurements data. The basic methods are artificial neural networks (ANN), genetic algorythms, methods for cluster arguments recording, fuzzy logics, basic components analysis, wavelet analysis, etc. The problems under consideration are the problems of forecasting, pattern classification and recognition, data clusterization, inverse problems from the fields of space physics, optical and gamma-spectroscopy, geological exploration, etc.

Students and post-graduates are widely involved into the scientific research of the Laboratory. The scientists have developed a specialized optional course "Introduction to neural networks and genetic algorythms" for the students, post-graduates and scientists of all departments of the University. Those specialists who do not work at MSU, are able to attend the same lectures at the University Extension courses and get state standard certificate.

This celcture course includes the following topics:

• the basic ANN models (multilayer perceptrons, recurrent networks, content addressable memory, Kohonen networks, networks with generalized regression, probabilistic neural networks, Hopfield networks);
• fundamentals of data origination and pre-processing
• method of cluster arguments recording, genetic algorythms (GA), genetic programming;
• additional ANN methods and algorythms;
• fuzzy logic;
• using of ANN and GA.

Computer practical studies on the following topics are also involved into the course:

• visual demonstration of the basic algorythms of neural network. "Neuropractice";
• practical studies with neural networks - fundamentals, solution of practical problems (basing on the programme packs NeuroShell 2 and NeuroShell Series);
• practical studies with genetic agorythms - fundamentals, solution of practical problems (basing on the programme packs GeneHunter and ChaosHunter);
• solitary work of the students on the solution of their problems under the supervision of the lecturer.

Detailed description of the course program is presented on the site http://nnga.narod.ru.