Events Calendar

AI Seminar-The Structure of Sequences Mining and Interpreting Networks from Event Log Data

Speaker: Brian Keegan, Harvard University

Talk Title: The Structure of Sequences Mining and Interpreting Networks from Event Log Data

Series: Artificial Intelligence Seminar

Abstract: Network science provides a rich set of theories and methods to understand the structure and dynamics of complex social, information, and biological systems. These approaches traditionally demand data with explicitly declared dyadic relationships or interactions such as friendship or affiliation. However, socio-technical systems like Wikipedia, Github, or Twitter often encode latent relationships within event logs and other databases. Temporal adjacencies in these event logs reveal sequences of actions that have complex and non-random properties that illuminate hidden structures within peer production systems. Using several case studies, I describe how complex networks can be extracted from event logs to understand the behavior of both users and artifacts within these systems. These networks encode a variety of rich structural and dynamic data distinct from traditional network approaches and illustrate user social roles within distributed collaboration as well as context and shifting interests of users based on their contributions. This approach has rich implications for mixed-methods research as it allows researchers to collapse large-scale event log data into more parsimonious network representations that can motivate qualitative analysis, visualization, and statistical modeling of complex behavior in socio-technical systems.

Biography: Brian Keegan is a research associate and data scientist for the Harvard Business School Hbx online learning platform. He received his PhD from the Northwestern University School of Communication in 2012 and was a post-doctoral research fellow in network and computational social science at Northeastern University until 2014. His research analyzes the structure and dynamics of online knowledge collaborations such as Wikipedia, Twitter, and online education under high-tempo and bursty conditions.

Host: Emilio Ferrara

Webcast: http://webcasterms1.isi.edu/mediasite/Viewer/?peid=0b39bdb4046d4835af24d94a23ddf6061d

Location: Information Science Institute (ISI) - 11th Flr Conf Rm # 1135, Marina Del Rey

WebCast Link: http://webcasterms1.isi.edu/mediasite/Viewer/?peid=0b39bdb4046d4835af24d94a23ddf6061d

Audiences: Everyone Is Invited

Contact: Peter Zamar