- March 5, 2020
11:00 am - 12:00 pm
Please join NYU CUSP for our new seminar series, featuring leading voices in the growing field of Urban Informatics.
Our next seminar will feature Mark Díaz, PhD Candidate in the Technology & Social Behavior program at Northwestern University, for a discussion on “Characterizing Social Bias in Algorithms.”
Thursday, March 5th, 2020
11:00am to 12:00pm
Room #1325 (Director’s Conference Room)
370 Jay Street
Brooklyn, NY 11201
The event is open to the public. Please RSVP below.
Characterizing Social Bias in Algorithms
This talk explores the use of algorithmic tools to analyze human behavior and highlights the value of mixed methods in evaluating algorithmic systems. Mark’s research unpacks how algorithmic tools measure and quantify human behavior, giving heed to the potential impacts of these systems on underrepresented communities. Because algorithmic tools are often created using data processed similar ways (e.g., using common, publicly available datasets, or collecting data exclusively through crowd work platforms), these tools can fail to capture data and patterns that reflect underrepresented groups. The result can be unintended social bias. The research questions underpinning this work are namely, How can researchers evaluate algorithmic tools’ fitness for their research contexts? How do algorithmic tools characterize attitudes and experiences—particularly those of underrepresented populations? How can researchers better incorporate expertise from from members of underrepresented communities to mitigate unwanted bias in algorithmic models?
Mark Díaz is a PhD Candidate in the Technology & Social Behavior program at Northwestern University where he is advised by Darren Gergle. In addition to his work on social bias in algorithms he has conducted research assessing low-income Chicago residents’ perceptions of city technology policy and workarounds to daily technological challenges. He also worked as a graduate fellow with the Chicago City Tech Collaborative to help launch the Chicago Data Collaborative– a cooperative effort by newsrooms, academics, and nonprofit researchers to assemble criminal justice data to support research and advocacy work. Before beginning his doctoral work, he worked in Accenture Labs developing virtual reality mobile applications and worked as a research assistant and programmer in the Virtual Human Interaction Lab with Jeremy Bailenson at Stanford University.