MSR NYC Data Science Seminar Series

Опубликовано 27 июня 2016, 18:48

Duncan Watts - Opening Remarks Yann LeCun - Demo Mor Naaman - Data and People in Connective Media In five minutes or less, I will talk about how we use methods from social science, people-centered design, data science and machine learning to understand social media data large and small, and build new applications that help us make sense of the city from (public) social media data. I'll also say a word about Cornell Tech and our Connective Media hub. OK, six minutes may be needed to squeeze it all in. Tony Jebara - Learning From Network Connectivity and Mobile Phone Data Many real-world networks are described by both connectivity information as well as features for every node. While most network growth models are based on link analysis, we explore how an individual's data profile without any connectivity information can be used to infer their connectivity with other users. For example, in a class of incoming freshmen students with no known friendship connections, can we predict which pairs will become friends at the end of the year using only their profile information? Similarly, can we using co-location to predict communication? In other words, by observing only the mobile location data from users, can we predict what pairs of users are likely to communicate? To learn how to reconstruct these networks, we present structure-preserving metric learning and apply it to FaceBook data, Wikipedia data, FourSquare data and mobile phone call detail records.