My name is Tomasz Arodz. I am a research assistant at the Department of Computer Science of the AGH University of Science and Technology, a leading technical university in Poland. The university is situated in the magical city of Kraków.
 
I graduated with a PhD in Computer Science from AGH in 2007, with a thesis on constructing ensemble classifiers and on their biomedical applications. I work under supervision of prof. Witold Dzwinel in a group led by prof. Jacek Kitowski.
 
In my research I explore classification and feature selection methods in high-dimensional, high-volume data. I also show usefulness of the developed machine learning algorithms in applications rising from medicine and biology. In summary, my research is centred on pattern recognition and bioinformatics.
 

Why do we need pattern recognition?
Because we are flooded with loads of data: images, genomes, other. And "faced with information overload, we have no alternative but pattern recognition", noted Marshall McLuhan.
 
But what is it exactly?
It is a methodology of data processing. Consider Ray Kurzweil's metaphor: "We apply a pattern recognition type of methodology (...). Think of a child catching a fly ball. At some point, she senses to take a few steps back, and then raises her hand (...). But she is not computing all the differential equations (...), even in an unconscious way."
 
So, it is trying to make computers handle data as humans do it. After all, "Homo sapiens is about pattern recognition" as William Gibson wrote in one of his novels.
 
It is making the machine find regularities, create abstractions, discard noise - understand the data. And, Isaiah Berlin observed, "To understand is to perceive patterns".

In bioinformatics, computer science methods meet problems from life sciences. Algorithmic analysis and mining of biomedical data is at the core of this research field.
 
I focus on three domains of bioinformatics. First, computer support for drug discovery and QSAR. Second, medical image analysis. Third, analysis of biological networks.