Graduate Seminar, 26th February

Speaker: Szymon Nakoneczny

Title: Natural language processing methods in biological activity prediction and quasar detection in large scale photometric surveys.

Abstract: The first part will focus on my master ‘s thesis. Virtual screening is a process in which databases of chemical compounds are searched in order to f ind structures characterized with a high biological activity, possible drug candidates. The goal is to classify active compounds by combining the natural language processing methods with a text representation of compounds. The second part is going to cover an approach of detecting quasars with means of machine learning. Because it is not possible to fully sample extragalactic objects with spectroscopy, quasars analysis requires a proper automated approach based on photometric surveys. We are going to focus on describing a problem and showing preliminary results of our approach.