The Next Level of Modeling Social Interaction: How to Detect, Quantify and Utilize Emotional Influence

Frank Schweitzer 

Wednesdays@NICO Seminar, Noon, November 16 2011, Chambers Hall, Lower Level

Prof. Frank Schweitzer, ETH Zurich

Abstract

Models of (bounded) rational agents failed to predict, or even capture, recent collective phenomena in social and economic systems. Ranging from the current financial crisis to the Arab spring, social "ingredients" such as herding, (dis)trust, empathy, aggression, or other forms of positive or negative emotions seem to play the major role in amplifying a critical situation. Do we have tools to detect and to quantify such emotions? Online datasets (written text from fora, microblogs, or reviews) can be used to apply sentiment analysis algorithms and to map the writer's emotions along the dimensions of valence and arousal. But what do we learn from that analysis, beyond the nation's mood in the morning? How do expressed emotions affect the response of other users in the cyberspace? Can we develop an interaction model of emotional agents to reproduce the stylized facts observed? Could we even manipulate cyber emotions?

Biography

Frank Schweitzer is Professor and Chair of Systems Design at ETH Zurich, one of the leading research universities in the world. Starting out as a theoretical physicist, he received a second Ph.D in philosophy of science, before turning his scientific interests on social and economic systems. Currently, he is visiting professor at MIT. http://www.sg.ethz.ch