PL EN

Information and communication technology

Author of the topic: prof. dr hab. inż. Khalid Saeed:   +48 85 746 91 96 ,   k.saeed [at] pb.edu.pl

Scope

Studying the available methods of human physiological biometric features such as face, iris, fingerprints. Then selecting one of the features to show the advantages and disadvantages of its known approaches, methods and algorithms. The aim will then be to find literature gaps and develop their own approach to retrieve and work out a suitable solution to the existing problems.

Author of the topic: prof. dr hab. Leon Bobrowski:   +48 85 746 90 50,   l.bobrowski [at] pb.edu.pl

Scope

The aim of the work will be to develop new techniques for structuring and representing text files based on the criterion of linear separability. One of the techniques for assessing the degree of linear separability of data sets representing particular categories is based on the minimization of the perceptron criterion function with the use of the basis exchange algorithm.

Author of the topic: prof. dr hab. Leon Bobrowski:   +48 85 746 90 50,   l.bobrowski [at] pb.edu.pl

Scope

The aim of the work will be to develop and implement multidimensional scaling methods for the purposes of structuring and geometric representation of knowledge in the form of a large number of observed relations. One possible application area there may be structuring of genomic databases.

Author of the topic: dr hab. Stanisław Jarząbek, prof. PB:   +48 85 746 91 03,   s.jarzabek [at] pb.edu.pl

Scope

Software modernization aims at improving the quality of aging programs and/or at reengineering them for reuse. For reuse, a prerequisite is identification of reusable program components. Proposed research theme will focus on analysis of commonalities and differences among program components that is a central theme in program modernization, but especially in analysis for reuse. Such analysis is usually done on multiple versions of software components resulting from software evolution. The PhD candidate will explore better ways to present commonalities and differences in a family of program versions, than conventional version control systems allow us to do.

Author of the topic: dr hab. Stanisław Jarząbek, prof. PB:   +48 85 746 91 03,   s.jarzabek [at] pb.edu.pl

Scope

Regression testing tools run test cases to validate a program after changes, and report – often long – lists of errors. How do we decide which errors should be corrected first? In addition to well-known error prioritization for correction based on their business importance, we could also take into an account dependencies among errors, to correct errors in an order that reduces the overall debugging effort. The idea is to identify and correct first root causes of failures, i.e., errors that may be causing many other program failures. As test cases are often mapped to relevant software requirements, we could use those mappings to infer error dependencies and prioritize errors. The candidate will explore requirement-based heuristics for determining a suitable error correction order.

Author of the topic: dr hab. Zenon Sosnowski, prof. PB:   +48 85 746 91 14,   z.sosnowski [at] pb.edu.pl

Scope

he main objective of the work will be to develop new methods to improve the efficiency of inference in rule-based expert systems by changing the structure of the knowledge base (modification of the RETE network architecture). In addition, the optimization of inference processes will be possible thanks to taking into account the incompleteness of knowledge. The effectiveness of inference will be measured by the time gain resulting from the lack of the need to review all the rules in the knowledge base and the number of new facts generated during the inference process.