Part of the 2014–2015 Fellows' Presentation Series
Lecture by Francesca Rossi RI '15
Preferences are ubiquitous in our lives, and we make decisions based on them—at times by compromising with individuals to whom we are connected on social media such as Twitter or Facebook. Knowing how to extract, model, and combine the preferences of several individuals in such online social settings therefore has great potential. The main goal of Francesca Rossi’s project is to study and develop techniques for preference extraction, modeling, and aggregation for collective decision making. Rossi aims to generalize the scope and goals of sentiment analysis in order to understand preferences over several correlated items. She proposes to do so by exploiting and adapting notions and techniques from AI (such as knowledge representation, natural language processing, reasoning with uncertainty, and machine learning) and voting theory.