Dissertation Defense

Social Reference Processing with Collaborative Human-AI Systems

Jordan HuffakerPh.D. Candidate
3941 Beyster BuildingMap

Hybrid Event: Zoom  Passcode: 80364

Abstract: I introduce several human-AI systems that can understand and process social references — language that invokes connotations by overlaying parts of social and cultural contexts. I argue that developing a rich understanding of social references is a key challenge that must be overcome in order to accomplish numerous communication tasks, especially in online environments where language rapidly develops. In particular, I demonstrate the importance of understanding social references in order to distinguish emotionally manipulative language from content, to identify rhetoric that others marginalized people groups, and to ensure high-quality language translations for complex terms. In each of these problem domains, I examine the challenges that limit existing systems, including systems that rely on the capabilities of crowds and machines on their own. Instead, I propose a series of human-AI approaches that make it possible to coordinate groups of people and machines to accomplish each of these tasks more effectively.


CSE Graduate Programs Office

Faculty Host

Prof. Mark Ackerman