Topic: Words and Networks: Considering the Substance of Information for Constructing and Analyzing Socio-Technical Networks

Speaker: Jana Diesner [Website]

Affiliation: Carnegie Mellon University

 

Abstract:

 

Socio-technical networks represent the interactions between social agents and infrastructures in systems such as companies, countries and communities of practice. The functioning and evolution of these networks involves the frequent production and processing of information. This information often occurs in the form of natural language text data, and can serve as a source of information about networks. I present two approaches that we have developed for integrating theories and models from the social sciences with methods from natural language processing and machine learning in order to facilitate the consideration of the content of text data for network analysis.

 

First, I describe a computational solution for extracting the structure of socio-technical networks from large-scale text corpora. This approach is particularly valuable when text data are the only source of information about a network. To put this approach into an application context, I show how we extracted network data about the Sudan from news wire data. I report on our findings from using the extracted data to answer the following substantial questions: Who are the key players in this country, and how does their importance change over time? What themes connect ethnic groups? What resources are involved if ethnic groups are associated with conflicts or war?   

 

Second, I present a methodology that uses a theory from socio-linguistics about innovation diffusion to inform the detection of people who are likely to induce or resist changes of affairs in networks, and contrasts the language use of people in these roles. I discuss our findings from applying this methodology to collaboration networks constructed from project proposals funded under the European Union’s Framework Programmes.

 

About Jana Diesner:

 

Jana Diesner is a PhD candidate at Carnegie Mellon University, School of Computer Science. Her research mission is to contribute to the computational analysis and better understanding of the interplay and co-evolution of information and the functioning of socio-technical networks. The goal with her interdisciplinary work is to advance methods and technologies that help people to collect rich network data, and to support meaningful reasoning about the underlying systems. In her empirical work, Jana studies networks from the business, science and geopolitical domain. She is particularly interested in factors that impede the sustainable development of networks and their wider context, and in covert information and networks.