Modeling the socio-semantic dynamics of scientific communities

How are the social and semantic structures of a scientific community driving future research dynamics? In this thesis we combine natural language processing techniques and network theory methods to analyze a very large dataset of scientific publications in the field of computational linguistics, i.e. the ACL Anthology. Ultimately, our goal is to understand the role of collaborations among researchers in building and shaping the landscape of scientific knowledge, and, symmetrically, to understand how the configuration of this landscape influences individual trajectories of researchers and their interactions. We use natural language processing tools to extract the terms corresponding to scientific concepts from the texts of the publications. Then we reconstruct a socio-semantic network connecting researchers and scientific concepts, and model the dynamics of its evolution at different scales. To achieve this, we first build a statistical model, based on multivariate logistic regression, that quantifies the role that social and semantic features play in the evolution of the socio-semantic network, namely in the emergence of new links. Then, we reconstruct the evolution of the field through different visualizations of the knowledge produced therein, and of the flow of researchers across the different subfields of the domain. To summarize, we have shown through our work that the combination of natural language processing techniques with complex network analysis makes it possible to investigate in a novel way the evolution of scientific fields.