Research Projects

Methods of ontology alignment involving semantics and valuations of ontology concepts

Principal Researcher: Professor Ngoc Thanh Nguyen
Researcher: M.Sc Marcin Pietranik
Funding Organization: National Science Centre
Duration: 2011-2013

The goal of this research project is developing methods of finding mappings between ontologies. The main contribution to this widely discussed topic will be expanding basic building blocks of these structures(which are class' attributes) with explicitly given semantics and incorporating formal criteria of identifying relationships between them into the process of designating valid alignments.


Method of determining a personalized learning scenario in E-Learning systems

Principal Researcher: Professor Ngoc Thanh Nguyen
Funding Organization: Polish Ministry of Science and Higher Education
Duration: 2009-2011

The main goal of this research project is to work out a method for determination of an effective scenario for a student using personalization methods. Next this method will be implemented in an intelligent e-learning system.


Method of knowledge integration in selected problems of collective intelligence

Principal Researcher: Professor Ngoc Thanh Nguyen
Funding Organization: Polish Ministry of Science and Higher Education
Duration: 2009-2012

The main goal of this research project is work out a set of tools for knowledge integration using methods for collective intelligence. The focus of this project is on such aspects that processing inconsistency of knowledge, communication languages in multi-agent environments, recommendation systems, and incomplete data processing.


Modelling Computational Collective Intelligence by Using Consensus Theory

Principal Researcher: Professor Ngoc Thanh Nguyen
Funding Organization: Polish Academy of Science (PAN – Poland) and National Research Foundation (NRF-Korea)
Duration: 2010-2011

The main goal of this bilateral research project is to work out a set of tools for collective intelligence with using Consensus Theory. The focus of this project is on such aspects that processing inconsistency of knowledge in web-based systems and ontology integration.


Computer Methods in Knowledge Processing Problems in Autonomous Systems

Principal Researcher: Professor Radosław Katarzyniak
Funding Organization: Polish Ministry of Science and Higher Education
Duration: 2010-2013

The main target of this project is to create and verify effective computational tools for resolution of knowledge management problems by autonomous systems. Particular research problems cover many issues related to knowledge aquisition, processing and semantic communication by BDI-like systems/agents (belief-desire-intention). New computational methods for knowledge creation, knowledge structures' development as well as linguistic summarization extraction from autonomous data bases are elaborated in this project. Examples of detailed problems are: elaboration of an original semantic communication model for BDI agents, elaboration of a formal and embodied ontology management module for autonomous agents, elaboration of autonomous strategies for ontology mapping in multiagent populations, elaboration of new interaction protocols for BDI agents with collective knowledge processing.


Multiple model prediction methods for dynamic regression problems

Principal Researcher: Dr. Bogdan Trawiński
Funding Organization: Polish National Science Centre
Duration: 2011-2014

Main goal of the project is to elaborate new models and prediction methods based on multiple model and hybrid approaches. It is planned to work out methods ensuring appropriate balance among four basic criteria: accuracy, stability, interpretability, and efficiency. The criteria are very important in some application areas particularly in long-term valuation, e.g. the valuation of real estate or debt packages purchased and sold on the free market.

Due to the incremental occurrence of data used for generalization learning systems are required to revise their current knowledge as soon as new observations occur. Thus, data instability is an essential issue in the prediction of dynamic regression problems. The aim of the project is to devise methods of evolving and incremental learning for regression problems, which allow for consideration of time variable data characteristics.

There are five principal research areas considered within the project, i.e. evolving fuzzy systems applied to ensemble systems, self-adapting genetic algorithms employed to optimization fuzzy systems, incremental algorithms for ensemble models, multiple model prediction of complex structures (sequences, graphs, and multigraphs), incremental learning using feature subspaces and instance subsamples.


Methods of Data Propagation to Solving Problems of Collective Intelligence

Principal Researcher: Dr. Dariusz Król
Funding Organization: Polish Ministry of Science and Higher Education
Duration: 2010-2011

The main goal of the project is to elaborate the effective methods of data propagation in contemporary web systems and their application to solve selected problems of collective intelligence. Such methods are applied when the subjects constitute autonomous and distributed sources of data (knowledge) and to solve problems the integration of that data (knowledge) is needed. Due to big complexity of up-to-date network systems, such as P2P, social networks or multi-agent systems, nature inspired algorithms, multi-criteria optimization and advanced network programming should be employed.