Undergraduate Computing Student Presents Paper as Lead Author at ACM CKIM 2023
Vaisnavi Nemala, a senior data science student in the Ying Wu College of Computing and Albert Dorman Honor’s College, presented a paper on which she served as lead author at the 32nd ACM International Conference on Information and Knowledge Management (ACM CIKM) 2023 in Birmingham, England.
The achievement is especially noteworthy, as she is one of very few undergraduate students to present as a lead principal investigator at a conference considered among the most prestigious in the field of technology. Nemala, furthermore, introduced herself to the project during the very first week of her freshman year at NJIT.
Her research for the paper titled Differential Privacy in HyperNetworks for Personalized Learning, conducted under a team lead by Associate Professor Hai Phan, combines the concepts of preserving differential privacy in user member datasets with federated learning (FL), a framework for collaborative learning through a coordinated server. The team’s proposed algorithm, called UDP-Alg, personalizes machine learning models in a federated mechanism for better privacy protection. Results confirmed that UDP-Alg significantly improved privacy protection at a modest cost in utility.
According to Phan, “It has a dizzying array of applications, from human sensing to healthcare, banking and mobile computing.”
Nemala’s interest in artificial intelligence (AI) research inspired her to attend an annual fall research meeting during her first week at NJIT, where faculty provide an overview of research they are conducting in an effort to recruit student collaborators. With little prior programming knowledge, she entered Phan’s lab under his mentorship.
“I learned machine learning from scratch on my own at the same time I was taking CS 100 (an introductory first year course). I continued to brush up during the following summer so I could continue to contribute meaningfully to the project,” she said. This allowed her to take a “giant leap” forward toward creating an algorithm later on.
The project was done over a two-year period, which began during the second quarter of Nemala’s second semester. She spent the first semester focusing on necessary learning, with a pause during the following semester to take part in a co-op opportunity with Optum, Inc., a healthcare services provider. She then returned to the project as a principal investigator and lead author until its completion.
Funding to attend the international ACM CIKM conference was provided through a Dean’s Fund for Student Development Grant in the Albert Dorman Honor’s College.
As a first-time presenter for an audience of over 100 people during the course of the day, Nemala, who attended the conference as NJIT’s sole representative, found the experience slightly overwhelming but highly rewarding nonetheless.
She observed, “I was the only undergraduate among mostly Ph.D. and master’s students. It was thrilling to be able to converse with them at a technical level and understand the nuances.”
Nemala will spend her last semester continuing to work with Phan on another project related to Explainable AI. After graduation, she will join Mitre, a non-profit operating federally funding R&D centers in support of government agencies, as an associate data scientist.
Phan remarked, “[Vaisnavi’s] dedication and persistence in solving problems independently and collaboratively are key attributes that only 10% of undergraduates tend to possess. I can see a bright future for her. She will be a leader in whatever she is passionate about.”