
Who we are
We are a research group at the University of Turin with two main interests: computational biology and theory of neural networks (and sometimes these two topics overlap!)
Computational Biology
We apply methods from physics and mathematics to many different topics in biology, including gene differentiation and reprogramming, whole-genome duplication, and sequencing. By integrating bioinformatics, statistical modeling, and machine learning, we aim to understand the complex networks that govern these phenomena. Additionally, one of our main interests is the study of the scale laws that govern component systems, trying to find the universal principles underlying biological organization. Another research topic is the intrinsic dimensions of biological datasets, using dimensionality reduction and manifold learning methods to find meaningful trends from complex biological data. We focus on simple, interpretable models, rejecting black-box algorithms in favour of probabilistic models, network theory, and component systems.
Neural Networks
Neural networks have been the most influential algorithms in modern machine learning and AI. We believe it is crucial to focus on their theoretical aspects, since their development relies less on deep understanding than on trial and error. We use tools from physics, such as statistical mechanics and dynamical systems theory, and apply them to these systems to (hopefully) better understand their underlying principles.