معرفی
Martin Mozina serves as an Assistant Professor and Researcher at the Faculty of Computer and Information Science, University of Ljubljana, where he has been affiliated with the Artificial Intelligence Laboratory since 2004. His primary institutional role focuses on advancing machine learning methodologies integrated with domain knowledge.
His research centers on developing machine learning algorithms that incorporate prior knowledge alongside training data, with significant contributions to model visualization through nomograms and automated chess position analysis. Key innovations include argument-based machine learning frameworks and techniques for resolving knowledge acquisition bottlenecks in intelligent systems.
Analysis of his 2004-2009 publications reveals consistent emphasis on interpretable AI, combining symbolic reasoning with statistical learning. His work bridges theoretical machine learning with practical applications in medical diagnostics, game AI, and decision support systems, particularly through nomogram-based classifier visualization and chess tutoring systems.
He actively participates in major research initiatives including:
- DRIFT (L2-4436): Real-time optimization of low-voltage networks using deep incentivized learning (2022-2025)
- Umetna inteligenca in inteligentni sistemi: Agency-funded AI research program (2015-2020)
- Strojno učenje v gradnji inteligentnih sistemov: Machine learning for intelligent tutoring systems (2011-2014, 2013-2014)
- Molecular markers for lung cancer research (2011-2014, 2016-2018)
- X-MEDIA: Large-scale knowledge sharing across media (EU project, 2006-2009)
Mozina's work within the Artificial Intelligence Laboratory demonstrates sustained focus on making machine learning more transparent and applicable to complex real-world problems through hybrid symbolic-statistical approaches.
