Abstract:   Human behavior, either at an individual or collective level, is complex. This complexity warrants a myriad of disciplines dedicated to the study of human behavior, each with a unique perspective.  Investigators from disparate fields find themselves asking overlapping questions of similar data—for example, historians, political scientists, and journalists commonly read large quantities of text to construct an accurate picture of when and where an event happened, who was involved, and in what ways.  Computational methods assist investigators in the analysis of massive unstructured data; these methods are sufficiently generic to expose patterns in the data that are of interest across disciplines.  In this talk, I will present a Bayesian model for analyzing documents to detect and characterize events of potential significance. Specifically, significant events are characterized by interactions between entities (such as countries, organizations, or individuals) that deviate from typical interaction patterns; our model builds on topic modeling and distinguishes between topics that describe “business as usual” and topics that deviate from these patterns. To demonstrate this model, we will explore results on a corpus of over two million U.S. State Department cables from the 1970s.

Bio:    Dr. Allison Chaney is a Postdoctoral Research Associate at Princeton University, currently working with Barbara Engelhardt and Brandon Stewart.  She also received her Ph.D. in Computer Science at Princeton, under the advisement of David Blei, and holds a B.A. in Computer Science and a B.S. in Engineering from Swarthmore College.  Her research focuses on developing scalable and interpretable machine learning methods to identify influences on human behavior.