Stay Informed:
Baskin Engineering COVID-19 Information and Resources
Campus Roadmap to Recovery
Zoom Links: Zoom Help | Teaching with Zoom | Zoom Quick Guide

Machine Learning and Social Science for Business and Policy Innovation

Speaker Name: 
Kristen Altenburger
Start Time: 
Wednesday, March 10, 2021 - 11:00am
End Time: 
Wednesday, March 10, 2021 - 12:15pm
Via Zoom Link
Alvaro Cardenas



Following recent technological innovations, modern organizations frequently develop digital, social platforms that apply machine learning algorithms to user data. By efficiently learning from data, these algorithms may enable organizations to make better business and policy decisions. However, organizations face several challenges to ensure these algorithms are both fair and protect user privacy. This talk will first discuss my work to promote fair and equitable digital systems. One project examines the novel use of social network data in regulatory enforcement decisions (e.g., targeted food safety inspections), and demonstrates that the use of this data can import consumer bias into these decisions. Next, the talk discusses my joint work as part of LinkedIn’s first Economic Graph Challenge. In this work, we examined the “feminine modesty effect,” the notion that women are more modest than men in expressing accomplishments, in an online labor market. We find that in a sample of recent MBA graduates, women are less likely than men to complete informative text-based fields such as the Summary field. This talk will then discuss my work that develops statistical methods for characterizing social structure in attribute-rich networks. In settings where a user does not disclose their own information, homophily (i.e. “birds of a feather flock together”) becomes an important predictive signal for inferring that private user’s information. While work on homophily focuses on using information from one’s friends, our work demonstrates that additional information is contained in one’s “friends-of-friends.” This finding offers an alternative perspective on prediction based upon network structure, complicating the already difficult task of protecting privacy on social networks. In sum, these projects highlight the importance of understanding how new technologies intersect with existing social networks and demonstrate the unique role of computational social science in innovation. As society’s digital platforms continue to promote instantaneous connectivity, my work aims to develop analytical tools for organizations to facilitate secure and equal-opportunity platforms. 




Kristen M. Altenburger (she/her) is a Research Scientist on the Networks & Behavior group within Facebook’s Core Data Science team and is a Non-Resident Fellow with the RegLab at Stanford Law School. Her research focuses on developing statistical methods for characterizing social structures in networks and focuses on promoting equitable digital systems that feature complex cultural and political considerations. At Facebook, she is also involved in the co-teaching program with Georgia Tech which is aimed at increasing pathways into AI. She received her PhD (January 2020) in Computational Social Science in the Management Science & Engineering Department at Stanford University advised by Johan Ugander. Her graduate work was supported in part by a National Defense Science and Engineering Graduate Fellowship. She received her BS in Mathematics from Ohio University in 2012 where she was also a Barry M. Goldwater Scholar, completed a research fellowship at Stanford Law School in 2012-2014, and received her AM in Statistics from Harvard University in 2015. She was previously a Member of Technical Staff in the Data Science and Cyber Analytics Department at Sandia National Laboratories and was a 2016 SPOT Award recipient based on her research. During the summer of 2017, she was the first intern for the Social Science & Algorithm team at Netflix. 

As a gentle reminder, please respect the privacy of faculty recruitment by not sharing the candidate status of our guests with others outside of our organization.

Zoom Link: 

Event Type: