YASS, QUEEN!: A Proposal of an AAVE to combat racial bias in Sentiment Analysis Research
Jazmia Henry seeks to reduce racial and gender bias in sentiment and textual analysis systems through the creation of an African American Vernacular English corpus. She received a Bachelor's Degree from Tulane University in 2016 and a Masters Degree at Columbia University in 2018. She has a background in statistical analysis, Natural Language Processing, Machine Learning, Econometrics, and Politics. She currently works as a Machine Learning Specialist at Motley Fool and has formerly worked for Morgan Stanley and the Hillary Clinton campaign.
What is your research focus?
Natural Language Processing: Reducing Bias in Sentiment Analysis Systems.
How do you plan to change the world?
I seek to quantify the historically unquantifiable through machine learning. Our technological and mathematical advancements have given us the ability to quantify the experiences of marginalized groups. By creating an AAVE corpus for public use, we can improve large-scale sentiment systems that have traditionally shown racial and gender bias.
What is an interesting fact about yourself?
I am the living legacy of my grandmother. She was born in Santo Domingo and came to Frederiksted, VI as a young teenager, taught herself the language, and became active in her community. She made sure her children and grandchildren valued education and took advantage of the opportunity to learn even though she never graduated from high school. She never got to see her granddaughter become the first in the family to attend prestigious colleges and first girl to get a graduate degree, but to know me and to benefit from any work that I do is to know and benefit from the work of my grandmother.
What is music/film/art that represents who you are?
Music: Erykah Badu, Teyana Taylor, Tank and the Bangas
Film: Spike Lee films made before 1992.