Introduction to CDL


Spending less time stuck in traffic, living in cities with little or no pollution, or leading healthy lives are outcomes that most of us would like. Achieving these outcomes, requires that people cooperate on their actions. For example, If everyone decides to use the same major expressway at the same time to get to work, they would experience congestion, the opposite of their individual goal. Similarly, if fewer people paid attention to the food that they ate and didn’t exercise, we would see more unhealthy people in the long run, increasing health costs for all.

How do we reconcile an individual’s desire for agency and choice in what they do, with what everyone wants for society as a whole?

While we don’t make often make the best decisions, working at scale and using readily available technology can help. Often, we make decisions with incomplete information; let our emotions influence our decisions; discount the future, or take decisions under scarcity. As a start, we need to provide information presented in an engaging format, about choices available to individuals when they are making the decision. Harmonizing individual goals and group goals is easier when more people participate. While the computational infrastructure to understand behavior and to better inform decision making at scale is challenging to develop, scale allows use to focus on proportions of different behavior in the group as a whole, instead of any individual’s behavior. Finally, scale is easier to achieve today, since most people walk around with a powerful computer in their pocket—their smartphone. The smartphone allows us to reach individuals no matter where they are and enables them to make informed choices.

We are increasingly actors in a play about information asymmetry and power. While human beings make well known judgement errors, our biases and judgement are systematically tracked by the different entities with whom we interact—including our grocery store (e.g. frequent shopper cards), social networks and search engines—and analyzed with powerful machine learning algorithms. More often than not, we have little understanding about the information collected, its use, and the decisions made from that information. Enabling individuals to take better decisions also implies we give them a more detailed understanding about what the entities that track their behavior know about them. Sometimes, better decision making includes the choice to hide.

Our goal at the Crowd Dynamics Lab is to empower individuals to take better decisions to benefit them, and to benefit society in the long run. We would like to do so at large scale. We develop algorithms for empirical analysis of behavior at scale (what can we infer? how well?), build systems (how to empower better decisions? provide choices in real-time where a person makes decisions—at the grocery store; on the smartphone while walking down the street) and develop theory (how much benefit could a particular framework provide? why do we observe the behaviors that we do?). You can read about work on information markets in out first post.