Fall 2024 alumni gold medal winners, Amy Tai and Nils Lukas
Friday, October 25, 2024

Graduate students win prestigious Alumni Gold Medal

Since 1970, Waterloo has honoured top graduating students with the Alumni Gold Medal. The Alumni Gold Medal symbolizes not only the students' accomplishments but also the proud tradition of exploration and leadership to which all Waterloo alumni belong.

Each year at fall convocation ceremonies, medals are presented to one master's and one doctoral student. This year, Amy Tai and Nils Lukas are the winners of the prestigious Alumni Gold Medal.

Master's winner

Advancing breast cancer research

Amy Tai

Amy Tai, a recent MASc graduate from Systems Design Engineering, focused her research on breast cancer clinical support. Breast cancer is the second most common cancer in women in Canada and the U.S., making up over 25% of new cases in women. Recent MRI technologies focused on other types of cancers and hadn't proved their effectiveness for breast cancer, but Amy's research sought to change that.

Amy's research designs new deep-learning models to improve two critical tasks in breast cancer treatment: predicting if cancer will fully respond to treatment, and accurately determining its grade. Amy's scientific contributions have been published in 25 scientific articles, have been presented at conferences across the globe, and have gained worldwide media coverage. 

Read more about Amy's research.

Doctoral winner

Investigating personal privacy concerns in generative AI 

Nils Lukas

Nils Lukas is a recent doctoral graduate from Computer Science and focuses his research on generative AI. The rapid advancement of generative AI models in recent years has the potential to transform business and society, but they also post novel trust, security, and privacy challenges. Dr. Lukas's research addresses these concerns in the context of machine learning models.

In a recently published scientific paper, Dr. Lukas and his team found that while a common privacy approach known as sentence-level differential privacy reduces the risk of personal data leaks, it still allows a small percentage of personal information to slip through. This research is one of the first investigations into the risk of personal data "memorization" in language models, highlighting that sentence-level privacy methods alone aren't enough to completely protect personal information. Dr. Lukas has made his code public to encourage more research on this important topic.

Read more about Nils Lukas's research.