Marco Federici

PhD Researcher

Short Bio

I graduated in computer science in 2016 at the University of Trento in Italy. Alongside my studies, I collaborated with a research group at Bruno Kessler Foundation first as an intern than as a consultant, working in the areas of business process prediction and sentiment analysis.

I completed the Master course in Artificial Intelligence offered by the University of Amsterdam in 2018, working on research related projects in the areas of Bayiesian compression, uncertainty estimation for graph neural networks and information theoretic approaches for generative modeling.

Selected Publications

Improved Bayesian Compression
https://arxiv.org/pdf/1711.06494.pdf

I have other publications in the field of busines processes and knowledge-based sentiment analysis that are not on ArXiv:
https://scholar.google.com/citations?user=TfInmkIAAAAJ&hl=en

Research

My research focuses on the application of information theoretic tools for representation learning and generative modeling, aiming to design and implement more robust and explainable neural networks.