Machine Learning for Evolving Systems
Overview
Machine Learning for Evolving Systems (MLES) leverages the power of machine learning to study the evolution of complex systems. Rather than focusing on predefined models or explicit equations, MLES utilizes data-driven approaches to uncover patterns, trends, and the dynamics of change across a wide variety of domains. From academic research to sports, MLES helps to understand how systems evolve over time by analyzing large-scale data, identifying emergent behaviors, and predicting future trends. The methods used in MLES—such as clustering, network analysis, and dynamic modeling—are particularly suited to uncovering shifts in performance, collaboration, and knowledge dissemination. Whether it’s tracking the evolution of research topics, identifying performance trends in sports, or understanding system dynamics in any other context, MLES provides valuable insights that can drive strategic decision-making and foster innovation. The primary focus is on applying machine learning techniques to data, uncovering evolving patterns without relying on explicit mathematical models or assumptions about how these systems evolve.
Detail
SciSci applies artificial intelligence and network science to the study of science as a complex, evolving system. By analyzing large-scale data on publications, collaborations, and research topics, we map how knowledge is produced, disseminated, and transformed over time. Our methods include topic modeling, heterogeneous network analysis, and dynamic clustering to uncover thematic convergence and fragmentation in research fields. We also examine the relationship between collaboration structures and scientific impact, helping institutions and researchers make evidence-based strategic decisions. SciSci aims to advance the meta-science of collaboration and knowledge production, contributing to a deeper understanding of how science evolves and how it can be made more open, connected, and impactful.
References
Pretolesi, Daniele, et al. “Geometric deep learning strategies for the characterization of academic collaboration networks.” IEEE Transactions on Emerging Topics in Computing 12.3 (2023): 840-851.
Pretolesi, Daniele, et al. “Artificial intelligence and network science as tools to illustrate academic research evolution in interdisciplinary fields: The case of Italian design.” Plos One 20.1 (2025): e0315216.Collaboration with
Daniele Pretolesi, Andrea Vian
Tennis Evolution, as part of MLES, uses machine learning to analyze the evolution of tennis over time. By studying performance metrics, player data, match statistics, and even tactical shifts, Tennis Evolution uncovers patterns in the game’s progression. This subfield explores how player strategies, training methods, and even changes in technology have influenced the way the game is played and perceived. Machine learning techniques help identify emerging trends, predict future developments, and understand how different factors have shaped the sport’s evolution. Whether it’s player performance or match dynamics, Tennis Evolution provides valuable insights into the game’s continuous transformation
References
Bayram, Firas, Davide Garbarino, and Annalisa Barla. “Predicting tennis match outcomes with network analysis and machine learning.” International Conference on Current Trends in Theory and Practice of Informatics. Cham: Springer International Publishing, 2021.
Collaboration with
Firas Bayram
Collaboration with
Andrea Vian, Daniele Pretolesi