Research Topics in LFR

Network science provides a powerful framework to understand a wide range of complex systems by representing their components as nodes and their interactions as links. We focus on uncovering universal structural patterns such as hubs, communities, and hierarchies, that shape system-level behavior across domains from biology and social systems to infrastructure and technology. By combining graph theory, sta+tistical physics, and data-driven modeling, network science moves beyond isolated components to explain emergence, robustness, and vulnerability. In our group, we focus on the following core topics:

Infrastructure and Mobility Networks

Network Medicine

Community Detection

Epidemic Spreading

Deep neural networks and biological brains share common learning dynamics that can be described using non-equilibrium statistical physics. We show that equations governing neuronal avalanches in brains also apply to activity cascades in artificial networks, which learn best in a quasi-critical regime between inactive and active phases. Rather than exact criticality, maximal susceptibility emerges as a better predictor of learning performance, offering a physics-based blueprint for designing more effective AI systems. In particular, we focus on:

Deep learning and brains

The growing availability of large-scale digital data on research funding, publications, citations, collaborations, and scientist mobility has opened new possibilities for understanding how science itself evolves. The science of science takes a quantitative approach to studying the relationships among researchers, institutions, and ideas across time and geography, aiming to uncover the mechanisms that drive creativity, discovery, and scientific impact. By combining methods from natural, computational, and social sciences, the field leverages big data, empirical analysis, and generative modeling to map the development of scientific fields, institutions, and careers. We work on identifying the processes shaping knowledge production, from citation dynamics and team formation to the measurement of performance, the quantification of novelty, and the emergence and evolution of research areas, ultimately informing tools and policies designed to foster and accelerate scientific progress. Our main topics are:

AI for Scientific Creativity

Novelty of AI content generation

Scientist independence

Society is a complex system. Simple interactions among individuals can give rise to unexpected large-scale patterns. Through repeated conversations in pairs or small groups, entire communities may eventually reach consensus, share cultural traits or languages, or even display coordinated collective behavior. Much like phase transitions in physics emerge from basic interactions among neighboring particles, social order can arise from local human interactions. Computational social science builds on this parallel by applying tools from statistical physics, network science, and data analysis to understand phenomena such as opinion dynamics, cultural diffusion, language evolution, and the formation of social hierarchies. In particular we focus on:

Opinion Dynamics