Computer models of complex processes and methods of machine learning and knowledge extraction

Computer models of complex processes and methods of machine learning and knowledge extraction

Research activity concentrates on two scientific fields:

computer modeling and simulation of complex systems,

machine learning and knowledge extraction.

Despite the differences in the classical understanding of these issues, they are closely related. Modeling is associated with the process of creating abstract models of reality by an observer, their realization in virtual “computational” spaces, validating any made assumptions, and explaining and predicting events. The modeling process consists of a series of subprocesses such as observations (measurements) of the real world, information selection and extraction, classification, hypothesis generation, and the formulation of laws and rules as well as the creation of theoretical and computational models and the assimilation of real data. Along with the analysis of obtained results and their visualizations, implementing a model in a multiprocessor-machine environment or metacomputer computing space is associated with the current information-processing technology. The above procedures (closed in a positive feedback loop) create a three-component scheme of knowledge extraction and generation that consist of experiment, theory, and computer simulation.

Machine-learning methods are an important element of the modeling process. They integrate the environment with the observer and real data. Machine-learning methods enable the perception of reality; i.e., the transformations of perceived objects, their evolution, and interactions into the formalism of mathematical models. They also allow for the selection and extraction of useful information and the generation of specialized predictive models.

Keywords: computer modeling, machine learning

Contact: prof. dr hab. inż. Witold Dzwinel BPP AGH


Research Team: CSG

Team Leader: prof. dr hab. inż. Jacek Kitowski BPP AGH