About

For over a century, individuals have been studied to understand social phenomena like how communities are built, or whom we trade with. From sociology and sociometry to social network analysis and machine learning, different approaches have been proposed to not only understand and explain the underlying mechanisms of link formation, but also to predict, for example, whom should we connect with to gain more visibility in a network, or who should be vaccinated first to decelerate the spread of an infectious disease. Today we know that the way we connect to others is to some extent explained by laws such as homophily and preferential attachment. However, social networks are complex systems and many other mechanisms may co-exists among all or certain groups of people. The challenge is then to understand how these mechanisms, that shape the structure of our networks, affect our daily lives, our decisions, omissions, and vice versa.

Our society is far from perfect. Wealth inequality, racial discrimination, urban segregation and gender glass ceiling are just a few examples of the problems that our society stills faces and fights on a daily basis. The social sciences have long studied these issues using surveys and census data. However, with the introduction of the Web, mobile applications and social media platforms more information about ourselves can be inferred thanks to Machine Learning. While these new approaches are good at explaining what we see in the data, they are not calibrated to mitigate these biases and instead they end up amplifying them when used for inference. A further problem is that these predictions are often used for decision making. This creates a feedback loop making it difficult to enforce fairness in a sometimes unfair society.

This satellite aims to open up a wide range of questions, discussions and possible new directions on explaining and solving societal issues using network research. Furthermore, these discussions will allow us to reflect on the challenges and opportunities to mitigate societal issues with (or without) algorithmic interventions.