Peter Krafft

PhD Student, MIT

I am a PhD student at MIT co-advised by Sandy Pentland and Josh Tenenbaum. I am academically located in EECS and CSAIL, but I spend my time in the Brain and Cognitive Sciences and the Media Lab. I study computational social science and collective intelligence, often from the perspective of cognitive science.

Selected Ongoing Projects

Modeling Collective Decision-Making

The most popular account of collective intelligence, the Wisdom of Crowds in Galton’s famous ox weighing, seems to suggest that independence of decision-making is critical for collective intelligence; yet depending on others for information or inspiration is clearly a central aspect of human life and society. In this work, I examine the trade-offs between independence and dependence in collective decision-making. Using a unique data set of millions of individual decisions in a social decision-making context, I then analyze how humans balance these trade-offs.

Herding in Financial Markets

Could financial bubbles be driven by the aggregate decisions of many small individual players? Herding behavior in humans has been studied extensively in laboratory settings and certain real-world settings, but the extent to which herding occurs in actual financial markets is poorly understood. In this work, we pursue a field experiment to test the extent of herding in a real asset market consisting of hundreds of thousands of traders.