In sociology, social networks are the networks created by personal relationships between people. Hannah Beilinson’s thesis examined how the spread of information in social networks advantages some individuals over others.
“In theory, this means, for example, that if I analyzed the social network of Haverford, I could identify groups of individuals that are very likely to receive important information and groups of people who are less likely to receive this information,” said Beilinson, a computer science major and statistics minor who graduated last year.
The method her thesis proposes could identify individuals in real-world social networks with less access to information so that gaps could be bridged.
“My dream would be for this work to provide a structure for decreasing the inequity in social networks so that more groups can have access to important information quickly,” said Beilinson.
Though she moved to New York to work as a software engineer at Accenture after graduation, Beilinson has continued working with her advisor, Associate Professor Sorelle Friedler, and collaborators since submitting her thesis last spring. Together they have focused on applying the code developed in Beilinson’s thesis to other examples of social networks.
“To me, my thesis was really about communities and the importance of connection and social structures. It showed me that those things aren’t just something I want, they’re something I’m passionate about,” she said. “I would love to refocus my future around advocating for communities and justice within them.”
What inspired your thesis work?
I started working with my advisor, Sorelle Friedler, in my sophomore year at Haverford. I participated in a reading group with her and a few other students on various approaches to analyzing the fairness of algorithms. At the start of my junior year, Sorelle was starting to think about how to extend this line of thinking to algorithms that operate on social networks.
[…] I’m fascinated by the phenomenon of friends-of-friends and the way in which everyone in small communities like Haverford is connected, whether or not they’ve ever spoken. I also really like the mathematical representation of these networks through nodes and edges (where nodes represent people, and edges represent the connections between them). The visual representations of mathematically defined social networks really pinpoint the intersection of humanities, social science, and math in a way I find deeply satisfying. Therefore, it was a no-brainer to get involved in Sorelle’s research on fairness in social networks.
How did Sorelle help you with your thesis?
She helped me in my thesis process in almost every way. As I mentioned in my previous answer, I was only even aware of algorithmic fairness as a field because of independent work I did with her throughout college. The idea to extend fairness principles to apply to social networks was hers as well. Sorelle helped me develop my ideas and guided my process of formalizing and testing them. […] And now that I’ve graduated, Sorelle is still pushing to integrate my thesis work into a published paper. She works tirelessly to make sure her advisees succeed and grow and orient themselves towards improving the field they’re in. I cannot thank her enough.
Though the Haverblog is now sharing where the Class of 2021 is headed and what they learned during their thesis process this year, Associate Professor Sorelle Friedler asked that we include Beilinson’s outstanding research since these blog series were preempted by COVID last spring.
“What They Learned”is a blog series exploring the thesis work of recent graduates.