Speaker
Description
Statistical validation of complex networks through maximum-entropy null models is explored as a robust framework to identify significant structures in large-scale data. These methods are applied to computational social science, including the study of polarization, echo chambers, language complexity, and information diffusion in online discourse. Simulations of social systems with large language models and machine learning illustrate the potential of computational approaches, while also emphasizing the irreducible complexity of social dynamics. Furthermore, extensions of network methods, such as Laplacian renormalization techniques, are investigated to analyze multiscale network behavior, with applications ranging from social dynamics to neural time series. Overall, the research highlights the versatility of network-based methodologies across diverse domains of complexity.
