Human-Centered Machine Learning Approaches to Dangerous and Complex Behaviors Online

February 10, 2020 - 11:15am to 12:15pm
Stevie Chancellor
3-180 Keller Hall
Lana Yarosh

Abstract: Research and industry use machine learning to identify and intervene in physically dangerous and high-risk behaviors discussed on social media, such as advocating for self-injury or violence. There is an urgent need to innovate in data-driven systems to handle the volume and risk of this content in social networks and its propagation to others in the community. However, traditional approaches to prediction have had mixed success, in part because technical solutions oversimplify complex behavior and the unique interactions of dangerous communities with both individuals and platforms. The difficulties in computationally handling these circumstances threatens the applications of these techniques to pressing social problems.

In this talk, I will describe my work in human-centered machine learning, an approach that refocuses technological innovation on the needs of humans, communities, and stakeholders. I study this through the domain of high-risk and dangerous mental illness behaviors in online communities, like opioid abuse, suicidal ideation, and promoting eating disorders. First, I will talk about my work in building novel prediction systems on large-scale social media datasets. These systems handle millions of data points and predict robust and accurate prediction of mental illness signals across several conditions. Then, I will discuss recent research on study design practices in applied machine learning, where I find alarming gaps in construct validity and rigor that jeopardize the state-of-the-art. Together, these inform an agenda for human-centered machine learning that is scientifically rigorous and more considerate of social contexts in data, providing a pathway for more impactful and ethical problem solving.

Bio: Dr. Stevie Chancellor is the CS + X Postdoctoral Fellow in Computer Science at Northwestern University. Her research combines approaches from HCI and machine learning to build and critically evaluate human-centered machine learning in online communities, focusing currently on high-risk health behaviors. Her research agenda has produced 13 publications in premier venues and has received four Honorable Mention awards at CHI and CSCW. Additionally, her work has been featured in national publications such as Wired and Gizmodo. She recently received her doctorate in Human-Centered Computing from Georgia Tech, supported by both the Snap Inc. Research Fellowship and the Georgia Tech Foley Scholars Award.

*CSE Faculty may sign up for an appointment to meet with Stevie Chancellor at the following link: Appointment Sign Up Sheet

All other scheduling inquiries may be sent to < >.