Denisse Ventura studies one-shot learning algorithms
By her senior year in high school, Florida native Denisse Ventura had not yet considered a degree in STEM.
She says her parents, who had immigrated to the United States from Guatemala just before she was born, were not familiar with the U.S. educational system. Few in her family had ever been to college. Some, back in Guatemala, had no formal education whatsoever. Because of this, she had no mentors growing up who encouraged her to enter science-related fields.
But during her senior year, she had a computer science teacher who encouraged her to pursue the field and enter the University of Florida. And over the summer of 2020, Ventura took part in SURF — the Stanford Undergraduate Research Fellowship.
She worked with Priyanka Raina, a professor of electrical engineering, and doctoral candidate Haitong Li, studying one-shot learning algorithms, where machine learning models can learn using only a few examples of data. Ventura’s work focused on evaluating how to develop one-shot algorithms by using hashing techniques, which allow large and complex data sets, such as those produced by machine learning systems, to be searched and accessed more efficiently.
In the future Ventura hopes to conduct research in natural language processing in the areas of algorithmic fairness and low-resource languages. Motivated by her upbringing in a Spanish-speaking household and her experiences translating for her parents on multiple occasions, Ventura wants to create fair natural language processing systems that can be used in multilingual settings to better serve people of various cultural backgrounds.
In addition to the practical experiences gained in a world-class research setting, Ventura credits SURF with providing valuable insight into the graduate school application process.
“I’d like to pay this kindness forward and mentor other students, especially underrepresented minorities in STEM, like myself,” Ventura says.