Collective Inference of Large-Scale Genealogical Networks

Genealogical networks, also known as family trees or population pedigrees, are commonly studied by genealogists wanting to discover their ancestors, but they also provide a valuable resource for studies in social science, genetics, historical demography, etc. These networks are typically constructed by hand which is a very time-consuming process since it requires comparing a large number of historical records manually. I present two approaches for inferring genealogical networks computationally: (1) linking historical vital records, and (2) aligning existing genealogical networks. In my PhD, I have developed novel probabilistic, combinatorial optimization, and active learning methods for these two approaches. An experimental evaluation of the proposed methods shows that they allow accurately inferring genealogical networks consisting of up to millions of individuals. Finally, I discuss the opportunities these networks offer to the new field of computational social science.

Eric Malmi is a doctoral student at Aalto University. His research interests include computational social science, machine learning, data mining, and natural language processing. Previously, he has done internships at Google Research (Zurich), Qatar Computing Research Institute, Idiap Research Institute, and CERN. His works have been featured widely in the media and they have received one best paper award nomination and one best paper award.