Cihan Duran, and coauthors Bruce Desmarais and Yubai Yuan, has been selected as the winner of this year’s John Sprague Award for their paper “Graph Neural Networks for Inferential Network Analysis.” The John Sprague Award is given annually to the best paper on political networks presented by a graduate student at a political science conference in the previous year.
The paper introduces a new framework that allows social scientists to use deep learning techniques to study complex social networks without losing the ability to explain exactly why connections form. By applying this method to the network of U.S. state legislators on X, formerly Twitter, they discovered a “State Filter” effect, revealing that rigid state boundaries, rather than just partisan loyalty, act as the ultimate barrier to how politicians interact with one another online. Ultimately, this approach opens the door for researchers to apply deep learning tools across the social sciences.
