Nehal Naqvi is a senior Political Science, and Applied Statistics major.
He is a member of the Honors College and a Sondheim Scholar.
Title of your research project: Interdisciplinary educational training to improve students' audio deepfake discernment
Describe your project: We evaluated whether a deepfake audio training created by sociolinguists at UMBC would be effective in improving students’ discernment of audio deepfake files.
Who are your mentors for your project?
Dr. Christine Mallinson (LLC), Dr. Vandana Janeja (IS), and Dr. Elizabeth Stanwyck (Math/Stat). I found my mentors (Dr. Mallinson and Dr. Janeja) through the Data Science Scholars program at UMBC, which is an interdisciplinary program that allows students to apply their technical skills towards faculty research projects. Currently, my research has transitioned from the Data Science scholars’ program to UMBC CISAAD (Community Infrastructure to Strengthen AI for Audio Deepfake analysis) where students and faculty are working to tackle deepfake recognition through various academic disciplines (AI, linguistics, educational training).
Dr. Stanwyck was my faculty advisor for statistics and was my instructor for STAT 454 Applied Statistics. As I applied paired hypothesis testing method which were covered in Applied Statistics, I often would go back to Dr. Stanwyck to make sure I was fulfilling the assumptions for each method.
How did you become interested in this project?
I became interested in this project because of the impact of deepfakes in politics, especially within U.S. elections. I saw this research experience as an opportunity to apply my skills in statistics in an interdisciplinary context. I think researchers should be open to applying their skills in unfamiliar disciplines, as it can expose oneself to new perspectives and methodologies.
What has been the hardest part about your research/what was the most unexpected thing about being a researcher?
The hardest part of my research was explaining the statistical methodology to my PIs, as well as PhD and graduate students on team who came from a linguistics or AI background. I found it to be really important to explain the methods that one plans to use in an approachable and simple manner, so that individuals outside of the discipline can understand why the problem is being approached in a certain way.
To be a Researcher of the Week, email: aprilh@umbc.edu
