Julian Zubek

Research

I firmly believe that most human cognition occurs in interaction, concerns other people, and is influenced by the interactive environment. I adopt a dynamical perspective, emphasizing the importance of studying real-time, embodied processes rather than just their final outcomes. In my research, I frequently use pose tracking techniques to quantify movement patterns during naturalistic interpersonal interactions. Additionally, I develop agent-based models to explore interactive phenomena.

• Formal models of categorization, links between human category learning and data complexity.
• Machine learning with the emphasis on ensemble classifiers, classifiers diversity, data complexity, and methods for building consensus from multiple predictions.
• Applications of machine learning in bioinformatics. Facilitating reconstruction of protein interaction network by predicting not yet observed interactions based on the available experimental data.