Department of Methods of Artificial Intelligence and Applied Mathematics (DMAIAM) is an interdisciplinary group of specialists working within the Faculty of Computer Science. DMAIAM’s research activities are mainly focused on the following topics:
Data analysis and application of classifiers and regression models based on data (technical, economic, medical) - using standard methods and methods of artificial intelligence.
Solving optimization tasks - using classical methods and methods of artificial intelligence.
Development of mathematical models of dynamic plants and technical and economic processes.
Development and testing of a variety, including classical and modern (fuzzy, neuro-fuzzy), control algorithms of technical objects.
Development of learning and control algorithms for a mobile robot or a group of robots.
Additionally, the following research’s activities are in the spectrum of DMAIAM’s interests:
Application of the reinforcement learning of a behaviour policy for mobile robots
Decision making under uncertainty and risk, processing uncertain information
Local regression models and their application for uncertain and incomplete data
Machine learning and data mining ( data analysis, classification and regression algorithms, theoretical background on machine learning techniques, application to medical and industrial problems)
Probabilistic Inference (probabilistic modeling (Bayesian networks, HMMs) Probabilistic computation (statistical computation, distribution of statistics, propagation of uncertainty, system reliability, probabilistic arithmetic, etc.)
Artificial Intelligence ( Graph searching algorithms and related discrete optimization problems)
Rapid prototyping and testing of control algorithms using programmable controllers
DMAIAM has expertise in:
A software design and implementation for a specific task, an method assessment on a model and a development of technical documentations and reports, related to data analysis and application of classifiers and regression models, optimization methods, development of mathematical models, development and testing of control algorithms
DMAIAM is involved in some international and national projects, has experience at the development a software design and implementation mainly for a data analysis and control. DMAIAM cooperates with scientific partners from different sectors – including medicine, control, maritime technology etc. and with industrial partners.
- Technical, economic, medical data analysis based on classical methods and methods of artificial intelligence
- Solving optimization tasks
- Development of mathematical models of dynamic plants and technical and economic processes
- Development and testing of a variety, including classical and modern (fuzzy, neuro-fuzzy), control algorithms of technical plants
- Development of tuning and supervisory algorithms dedicated to control algorithms
- Development of learning and control algorithms for a mobile robot or a group of robots.
- Application of classifiers and regression models based on technical, economic, medical data
Modern infrastructure: laboratories, equipment: programmable controllers (Bernecker&Rainer (X20s, Power Panels, Automation Panels, etc.), GE, Siemens, National Instruments); software : Wonderware Industrial Software (ActiveFactory, Application Server
Device Integration, Historian Client, Historian Server, InBatch etc.).