Esin Durmus is a fifth year PhD candidate in Cornell University. She is part of the Cornell Natural Language Processing group and is advised by Claire Cardie. Before starting her PhD she received a BSc in Industrial Engineering and Computer Engineering from Koc University. She is interested in Natural Language Processing, Social Media Analysis and Computational Social Science.
Argumentation is a discussion in which reasons are provided for or against some proposition or proposal. Effective argumentation is crucial to encourage critical thinking and motivates people to make more fair and informed decisions since it enforces people to be exposed to diverse set of perspectives and to communicate with people from different backgrounds. Previous work in Social Sciences and Psychology has shown that there are three main important components of opinion formation and persuasion in argumentation independently: the argument itself, the source of the argument and the audience to whom the argument is conveyed. My research focuses on understanding relative effect of these factors computationally taking advantage of the data on social media and online argumentative platforms. To achieve this goal, I have proposed new datasets and developed new techniques to automatically assess the interplay between these factors of opinion formation. Moreover, I have contributed to the field of automated argument generation and argument quality assessment with the motivation that automated argument generation models and debating agents could be used to help people gain different perspectives on the controversial issues if they are able to generate effective personalized arguments for a particular group of audience.